June 2009

Positive Report by Access Economics Pty Limited for

The Australian Government Department of Families, Housing, Community Services and Indigenous Affairs

Family Functioning
Scoping Study

Contents

Executive summary.i

1.        Meaning and measurement1

1.1     Definitions.1

1.2     Methods of measuring PFF.5

1.3     Incidence and prevalence approaches.9

2.        Measuring outcomes.11

2.1     Literature findings.12

2.1.1       Health outcomes.12

2.1.2       Productivity outcomes.20

2.1.3       Social/criminality outcomes.22

2.1.4       Summary of literature based findings.25

2.2     Longitudinal child and youth studies.26

2.2.1       Longitudinal Study of Australian Children (LSAC)29

2.2.2       Longitudinal Survey of Australian Youth (LSAY)32

2.2.3       Australian Temperament Project (ATP)32

2.2.4       Household, Income and Labour Dynamics in Australia (HILDA)35

2.2.5       Summary of Australian longitudinal data sources.35

2.2.6       Proposed method of determining outcomes.36

2.3     Cost components and considerations.37

2.3.1       Health system expenditure.39

2.3.2       Productivity and carer costs.40

2.3.3       Criminal justice system costs.46

2.3.4       Taxation revenue forgone and associated DWLs.47

2.3.5       Other indirect financial impacts.48

2.3.6       Burden of disease.49

2.3.7       Net present value and the discount rate.52

3.        Efficiency of current Australian and International interventions.54

3.1     Family assistance programs and income support payments.56

3.2     Family relationship services.60

3.3     Stronger Families and Communities Strategy 2004-2009.66

3.3.1       Communities for Children.67

3.3.2       Invest To Grow..70

3.4     Triple-P (positive parenting program)72

3.5     Early Childhood education.77

3.6     Peel Child Health Survey.84

3.7     Responding Early Assisting Children (REACh)84

3.8     Reconnect85

3.9     Youthlinx.88

3.10  Transition to independent living allowance (TILA)90

3.11  SureStart91

3.12  Parents under pressure (PUP)96

3.13  Methodology for CBA and/or CEA.102

4.        Conclusion.105

References.106

figures

Figure 1‑1:  Prevalence and incidence approaches to cost measurement10

Figure 2‑1:  Influence of family time on smoking (left) and drinking (right) attempts 13

Figure 2‑2:  Pathways predicting ‘happiness’ using different aspects of mediation 17

Figure 2‑3:  Optimal investments in human capital, by lifecycle 20

Figure 2‑4:  Parental inputs (age three) by household income (age nine months)22

Figure 2‑5:  Comparison of offending trajectory groups on measures of FF 25

Figure 2‑6:  Longitudinal pathways to adolescent depression (15-16 years), example only 34

Figure 2‑7:  Longitudinal pathways to adolescent antisocial behaviour (15-16 years), example only 34

Figure 2‑8:  Approximate age of study cohorts and bridging the current information gap 36

Figure 2‑9:  Conceptual map for valuing costs of NFF, example only 38

Figure 2‑10:  Productivity losses 42

Figure 2‑11:  DWL of taxation 47

Figure 2‑12:  Ten-year government bond yields (daily)52

Figure 3‑1:  Prevention continuum  54

Figure 3‑2:  Diagram of Triple P system of parenting and family support73

Figure 3‑3:  Prevalence of factors across Reconnect cases 87

Figure 3‑4:  Cost effectiveness charts 104

Tables

Table 1‑2:  Theoretical models of general FF 7

Table 2‑1:  Maternal and paternal influences on alcohol consumption 13

Table 2‑2:  Adolescent smoking determinants in New Zealand 14

Table 2‑3:  Rate ratios for comparing the presence or absence of parental and peer factors influencing smoking status 15

Table 2‑4:  Multivariate regression results predicting substance abuse/dependence symptoms 16

Table 2‑5:  Significant influences of independent variables on time-sampled peer interaction variables, using linear regression 18

Table 2‑6:  Significant influences of independent variables on event-sampled peer interaction variables, using logistic regression 19

Table 2‑7:  Family risk factors associated with child mental health problems 20

Table 2‑8:  Child and youth related longitudinal studies 27

Table 2‑9:  Average employment rate and AWE in the Australian population 44

Table 2‑10:  Expected retirement age and remaining lifetime earnings ($2005)45

Table 3‑1:  Family Tax Benefit Part A 57

Table 3‑2:  Family Tax Benefit Part B  58

Table 3‑3:  Double Orphan Pension 58

Table 3‑4:  Maternity Immunisation Allowance 59

Table 3‑5:  Baby bonus 59

Table 3‑6:  Family dispute resolution, means and SDs of parent measures 64

Table 3‑7:  Family dispute resolution, other parent measures 65

Table 3‑8:  Family dispute resolution, means and SDs of children measures 65

Table 3‑10:  Estimated effect sizes for different intensities and formats of Triple P  74

Table 3‑11:  Results of meta-analyses of Triple P across modalities 75

Table 3‑12:  Costs of conduct disorders (2002-03 A$)76

Table 3‑13:  Threshold analysis of Triple P (cohort)76

Table 3‑14:  Summary of evidence on ECE outcomes 80

Table 3‑15:  BCR of ECE programs 82

Table 3‑16:  CBA of ECE programs 83

Table 3‑17:  Child abuse prevention (including REACh)85

Table 3‑18:  Reconnectperformance indicators 88

Table 3‑19:  Youthlinx performance indicators 90

Table 3‑20:  Outcome variables by research question in a study of SSLPs 93

Table 3‑21:  SSLPs and comparison groups, significant mean score differences 94

Table 3‑22:  Measures that show significant differences as a function of SSLP lead agencies 95

Table 3‑23:  PUP modules 97

Table 3‑24:  Means and SDs for outcome measures - PUP  99

Table 3‑25:  Changes in risk status and clinically significant changes (Reliable Change Index) from pre-treatment to six months follow-up 100

Table 3‑26:  Means and SDs for outcome measures – Women Prisoners Study 101

Table 3‑27:  Means, SDs and T-test statistics of key outcomes – child abuse study 101

Acknowledgements

Access Economics would like to acknowledge with gratitude the expert knowledge and inputs provided by members of the Expert Reference Group for this project.

Brian Babington

Chief Executive Officer, Families Australia

Lance Emerson

Australian Research Alliance for Children and Youth (ARACY)

Jan Nicholson

National Family Functioning Team, Murdoch Children’s Research Institute

Frank Oberklaid

Director, Centre for Community Child Health, Royal Children’s Hospital

Simon Schrapel

Anglicare South Australia

Dorothy Scott

Director, Centre for Child Protection, University of South Australia

Stephen Zubrick

Co-Director for Developmental Health, Curtin University of Technology

Acronyms

Executive summary

Access Economics was commissioned by the Australian Government Department of Families, Housing, Community Services and Indigenous Affairs (FaHCSIA) to conduct a scoping study estimating the feasibility of, and the methodology for, undertaking a research project that quantifies, in economic terms, the value of ‘goods and services’ provided by positive family functioning (PFF). 

As the initial Statement of Requirement acknowledged, there are a number of definitional and confounding issues associated with the task, not least issues of aggregation of the impacts of many factors occurring together, time lags between positive or negative family functioning (NFF) and their impacts, and data challenges attributing effects from different types of family functioning (FF).

The counterfactual of PFF (namely NFF) is used in the quantification and costing process.  This is because the benefits of PFF are the same as avoiding the costs of NFF.  In addition to this, the process of evaluating consequences of NFF has a simpler methodology.  While the key outcomes discussed in this scoping study may seem to be a simplification of the real consequences of NFF, they are aspects that can be quantitatively measured.  Although other aspects cannot (as of yet) be accurately quantified, they could be discussed in a qualitative manner.

In addition to proposing a methodology for quantifying the economic value of PFF, a second goal of this scoping project was to assess methods for conducting cost benefit analysis (CBA) and/or cost effectiveness analysis (CEA) to establish the returns to government and society for investments made in supporting FF.  A final task was to scope the likely costs, timeframes and outputs from a full study applying these methods to estimate the value of PFF and returns on investments.  To achieve these goals, a four-step process was employed.

q  Step 1 :  Establishing an expert reference group (ERG) in conjunction with FaHCSIA, to determine a definition of PFF as well as to advise on particular data sources, technical aspects and to provide their professional opinion.

q  Step 2 :  Literature and data investigations.

q  Step 3 :  Methodological review and discussion for estimating the value of PFF.

q  Step 4 :  Review of CBA and CEA for various FF interventions.

In Step 1, the ERG was established with members participating on a voluntary basis.  Two structured meetings were held for members to provide their inputs and comments.  These were a project inception meeting (PIM) – where a definitional framework was developed – and a second meeting to provide further input focusing on the impacts of PFF and on interventions. 

A simple definition of PFF could not be established based on a review of the literature.  Instead, consistent themes (or ‘domains’ of FF) that recurred through the literature were identified, developed and agreed, in consultation with the ERG and FaHCSIA.  These domains provide an overarching definition of the FF environment, as depicted in the next table.  A summary is provided below.

Family functioning – positive and negative – is defined through a variety of emotional attributes, family governance frameworks, cognitive engagement and development characteristics, physical health habits, intra-familial relationships and social connectedness.  PFF is characterised by emotional closeness, warmth, support and security; well-communicated and consistently applied age-appropriate expectations; stimulating and educational interactions; the cultivation and modelling of physical health promotion strategies; high quality relationships between all family members; and involvement of family members in community activities.

Characteristics of FF domains

Source:  Access Economics in consultation with the ERG and FaHCSIA.

These domains of FF should not be interpreted as being independent of each other. Rather, they are characteristics that interact, complement and co-exist.  Historically there are a number of theoretical models that have been used to describe FF.  These include (see Table 1‑2 for further details):

q  Beaver System Model – describes FF in a two dimensional domain of competence and style;

q  McMaster Model of Family Functioning – asserts that the primary function of family is the development and maintenance of family members on social, psychological and biological levels;

q  Circumplex Model of Marital and Family Functioning – describes FF in terms of cohesion and flexibility;

q  Process Model of Family Functioning – discusses the goal of the family as the successful achievement of basic development in critical tasks;

q  Family Environment Scale – looks at three factors of social climate (interpersonal relationships, personal growth and family structure) that underscore FF;

q  Iowa Family Interaction Rating Scales – comprises a collection of scales that are underlined by social interactional, behavioural and social contextual theories; and

q  The System for Coding Interactions and Family Functioning – is based on a number of theories, including systemic family theory, structural family theory and social learning theory; functioning is assessed in a number of ways, including the family as a whole, each parent-child dyad as well as for each individual family member.

The theoretical models that best align with the family characteristics outlined above are the McMaster Model of Family Functioning (MMFF), Iowa Family Interaction Rating Scales (IFIRS) and The System for Coding Interactions and Family Functioning.  Of these, the McMaster Model has most frequently been used in the Australian setting.

There is a focus on outcomes associated with children in the literature and analysis of the impacts of FF. This is the result of the child’s inability to explicitly control their family environment and children are often viewed as the main victims of NFF.  The definition of family is as per the Australian Bureau of Statistics (ABS), but due to the emphasis of this study on child outcomes, there is no focus on couple-only families, couple families that are expecting children (in-utero) or couples/individuals caring for older relatives.

Literature and data investigation

Impacts from FF occur with lags.  Health, economic and social/relational outcomes do not necessarily occur straight away but can take many years of exposure before they manifest in later years, and then can extend for varying periods – sometimes for life.  The literature review process revealed three broad areas of outcomes.

q  Health outcomes were mostly observed through the occurrence of mental illnesses such as anxiety and depression later in life, but also included eating disorders, health behaviours (e.g. unsafe sex, physical inactivity, overweight and obesity) and substance abuse (e.g. smoking, alcohol and drug abuse), with the consequent physical impacts of these risk factors on morbidity and mortality outcomes.

¬  Health outcomes were most dependent on issues such as family attachment, joint family time and role modelling. 

q  Productivity outcomes were reflected in rates of labour force participation, employment and hourly wage rates, with a number of intermediate measures reported in the literature, such as reduced levels of literacy and numeracy and other measures of educational achievement.

¬  Productivity outcomes were most frequently associated with socioeconomic status of the family (parental income and asset base) in turn linked to parental involvement at school and parents engaging with their children more.

q  Social outcomes were measured primarily through their negative manifestations –involvement in antisocial behaviour such as delinquency, and the probability of criminal behaviour during youth and later in life.  In contrast it was harder to quantitatively associate PFF with positive manifestations such as the quality of inter-personal relationships and future community contributions.

¬  Social/criminality outcomes were linked most to the quality of the parents’ relationship, bonding between parents and children, length of time the biological father remained in the family home, number of changes of parents, parental conflict, and poor social behaviours by parents (criminality, substance abuse).

The literature review provided a valuable insight into the outcomes associated with FF, while also highlighting a number of methodological problems in potentially valuing PFF. 

q  First, many studies reported correlation between FF aspects and outcomes, but it was sometimes difficult to ascribe this to causation, particularly for cross-sectional analyses.

q  Second, some outcomes cannot be easily converted into economic costs, especially when outcomes are interim (e.g. child literacy and numeracy standards) rather than completion of Year 12 or adult employment, labour force participation or earnings.

q  Finally, different and inconsistent measures of FF and of outcomes have been applied in past studies, presenting a challenge for quantitative interpretation.

A number of current Australian longitudinal studies provide a potentially valuable source of information to determine the causal effects of FF in Australia.  Four potential longitudinal studies identified are summarised below.

q  Longitudinal Study of Australian Children (LSAC):  funded by the Australian government, managed by FaHCSIA and initiated in 2003, the study recruited 5,000 children aged less than 12 months as well as 5,000 children aged 4‑5 years.  Data are being collected on the child (health, temperament, literacy experiences), parents (socioeconomic status, parenting style, health) as well as the broader family, child care, school and community environments.

q  Longitudinal Survey of Australian Youth (LSAY):  jointly managed by the Australian Council for Educational Research and the Australian Government Department of Education, Employment and Workplace Relations, the study examines the transitions of young Australians through school, education after secondary school and/or work.  Ages of participants vary from those born in 1961 to those that were in Year 9 of school in 1998.  Specific issues examined are achievement and completion of school, participation in vocational and university education as well as employment, although FF was not a measure recorded in this survey.

q  Australian Temperament Project (ATP):  initiated in 1983 with a representative sample of more than 2,000 infants and families from urban and rural Victoria, the project examines psychosocial adjustments through childhood and adolescence as well as influences of personal, family and environmental factors.

q  Household, Income and Labour Dynamics in Australia (HILDA):  initiated in 2001, commissioned as well as funded through FaHCSIA, and managed through the Melbourne Institute at the University of Melbourne.  The study contains a representative sample of Australia’s population with approximately 14,000 respondents in Wave 1 of the study.  Information is collected annually with four key areas (‘households and family life’, ‘incomes and economic wellbeing’, ‘labour market outcomes’ and ‘life satisfaction, health and wellbeing’). 

Two of these studies (LSAC and ATP) have the potential to be useful for the determination of causal impacts of FF.  As LSAY did not collect information on FF variables, this longitudinal survey is unfortunately of little use here and HILDA would be limited to examining relationships between the family environment and parent’s participation in the labour force.  LSAC has the advantage of containing a breadth of data, valuable for controlling purposes.  However, a disadvantage is the relatively short timeframe of data collection.  LSAC would only be useful in providing information for causal impacts in the short to medium term, as the eldest participants from the child cohorts are currently 10‑11 years of age.

ATP is currently the only study in Australia that could be used to determine long term impacts of FF on health, economic and social outcomes.  Participants in this study are currently 25 years of age and would provide sufficient data to determine long term outcomes.  Indeed, preliminary works have been published which map the inter-relationships of a number of variables that interact over time and result in outcomes such as anxiety, depression, substance abuse, antisocial behaviour, academic achievement, risky driving and measures of social capital.  The main limitation of ATP is the length of time that measures of FF and parenting have been recorded.  These measures have only been taken since the participants were in their early teens, with no measures during infancy or early childhood.

Proposed method of determining outcomes

Outcomes would be mapped from the ATP and LSAC databases using a similar methodology to that developed by Prior et al (2005) (see Figure 2‑6 and Figure 2‑7).  Outcomes would be mapped for each wave of the study so that the likelihood of one of the events of interest occurring could be established across different age groups.  The ability to join information from LSAC and ATP is limited by slightly different methods of measuring FF.  There are, however, some consistencies, although one or two more waves of information would ideally be required to successfully bridge the gap in ages between the two studies, by mapping outcomes for children aged 10-11 years also to those aged 11 to 13 years. This approach is considered quite acceptable, as opposed to waiting some years for additional study waves to be completed.

Approximate age of study cohorts and bridging the current information gap

In order to complete the study in this way, a partnership with the Australian Institute of Family Studies (AIFS) and in particular the ATP study group would be required.  Such a partnership would offer quality benefits in terms of peer review insights although risks may be incurred in terms of timeframes and additional resource costs to ensure consistency of approach. Alternatives to such a partnership are limited (eg, waiting an additional 15 years until further LSAC data are available or using international sources if they exist, which would be less desirable).

Methodology review and discussion

There are two main approaches to measuring the value of PFF, which is equivalent to estimating the cost of NFF.

q  The incidence approach measures the number of new cases of NFF in a base period (say, 2009) and attributes costs due to that NFF, both financial and non-financial, over the affected persons’ lifetimes.  The total cost of NFF represents the net present value (NPV) of current and future costs incurred due to all new cases in that year.

q  The prevalence approach measures the total number of people ever affected by NFF in a base period and costs associated with those cases, both financial and non‑financial costs, in that year. 

These methodologies present information in slightly different ways and have different applications.  The prevalence approach is a ‘snapshot’ as it reports impacts in one year and, in general, prevalence estimates can be easier to understand by policy makers as they relate to a single point in time.  Also, the financial value of impacts can be compared to measures such as gross domestic product (GDP).  The incidence approach reports the lifetime impacts associated with a condition; as policy interventions are focused at preventing NFF, this methodology is most relevant for its application in CBA or CEA studies.

Aetiology, data sources and the study design used to track outcomes are key deciding factors determining which approach is best.  In this instance an incidence approach is considered most appropriate, due to these factors and the lag times between exposure to the family environment and associated outcomes. 

The valuation exercise would incorporate a number of cost components related to the three broad areas of outcomes (health, productivity and social/criminality) previously discussed.  A number of these cost components will be attributable to more than one of the impacts discussed above (for example, health system expenditure can be attributable to both health specific outcomes and criminal behaviour).  These components are:

q  health system expenditure;

q  productivity and carer costs/gains;

q  criminality costs;

q  taxation revenue forgone and associated deadweight losses (DWLs);

q  other financial impacts; and

q  burden of disease (measured in disability adjusted life years or DALYs).

Cost components map to FF as shown in the diagram below.

Conceptual map for valuing costs of NFF, example only

New interventions could be evaluated via their impact on factors causing key outcomes.  The resulting reductions in the incidence of negative outcomes (such as anxiety, depression or criminality) have economic benefits attached.  The difference between the economic benefits and costs under a scenario with the intervention and one without would be estimated.  Net economic gains between the scenario and the ‘base case’ (no change) are evaluated to determine the intervention’s overall cost effectiveness or return on investment.  These analyses can be broken down further to incorporate who bears the costs or benefits of the program and hence the net benefits for the individual, Government and so on.

CBA and CEA literature review of FF interventions

Literature was sourced on the effectiveness of a broad range of interventions targeting FF.  These interventions were chosen, in close consultation with the ERG and FaHCSIA, as a representative sample selected on the basis of achieving a balance of:

q  Australian Government, state/territory and international;

q  focus on parent, child, family and community/environment;

q  different life cycle stages (pre-birth, infancy, later); and

q  existing evidence compared to future potential evidence.

General families

q  Family assistance programs and income support payments.

q  Family Relationship Services (FRS) and FaHCSIA’s Children and Parenting programs (soon to be combined as the FSP), focusing on a couple of individual services e.g. Family Education and Skills Training around parenting, Adolescent Mediation for young people.

Young children

q  Stronger families and communities strategy 2004-2009

o  Communities for children

o  Invest and grow (a mostly early years’ intervention, targeted at socially disadvantaged areas)

q  positive parenting programs (Triple-P).

q  early education programs

q  Peel Child Health Study (Our Children, Our Family, Our Place): looks at biological factors interacting with community (school, daycare, recreation) and how it influences health and development.

q  REACh, a program intended to improve the capacity of families and caregivers to respond appropriately to children's needs for care, development and safety through timely access to community resources that can support them in their parenting role.

Youth

q  Reconnect: The Reconnect program uses community-based early intervention services to assist young people aged 12 to 18 years who are homeless, or at risk of homelessness, and their families. Reconnect assists young people stabilise their living situation and improve their level of engagement with family, work, education, training and their local community.

q  YouthLinx - is prevention and early intervention program that supports young people aged 11 to 16 and their families build self reliance, strengthen their family relationships and encourage community involvement.

q  TILA - helps young people exiting formal/informal care to transition to independent living through enhanced access to accommodation, employment/education and the development of life skills/knowledge.

Other

q  SURE start (from the UK).

q  Parents Under Pressure (PUPs) – in relation to parental substance abuse.

The quality of evidence on efficacy of these programs was varied, with a number only reporting qualitative findings.  Some programs that had a great deal of information on efficacy suffered other limitations, such as poor take-up in the community or limited applicability across all socioeconomic groups.  However, a short list of interventions could form the basis of more detailed evaluation using standard CEA and CUA techniques.

Conclusion

The method proposed for estimating the goods and services produced from PFF and evaluating FF interventions is summarised below.

q  The first step is measuring the incidence of NFF in Australia in a given year and mapping NFF to health, productivity and social/criminality outcomes or to interim outcomes that other literature evidence can link to these end outcomes – using LSAC and ATP datasets.  Since parallels and linkages between measures in these studies are limited, long term associations between FF in childhood could not be measured out into adulthood without some caution over the results, which would need to be addressed through sensitivity analysis.  In order to achieve these measurements, a partnership with ATP/AIFS would need to be formed.

q  Costs would be estimated for each outcome using: AIHW data on health system expenditures for the attributed health conditions; largely ABS data for productivity costs for the individual and for carers; largely SCRGSP data for criminal justice system and other government program costs such as educational assistance or child protection services responding to abuse notifications; macroeconomic data for estimating taxation revenue forgone and deadweight losses; Centrelink data for estimating welfare transfers; SDAC data for estimating aids and home modifications; other data for other minor financial costs; and AIHW data for estimating the burden of disease together with Department of Finance and Deregulation estimates of the VSLY.

q  CBA and CEA techniques could evaluate interventions to enhance FF and increase the goods and services available in the Australian economy, using information on a short list of the interventions reviewed in this scoping paper, along with their cost and efficacy. A first step in the CBA/CEA would be mapping in greater detail each intervention in this short list to the outcomes and costs as per the costing exercise.  Markov modelling or Monte Carlo simulation could then assess selected interventions relative to comparators (generally the base case could be expected to be continuation of the status quo). CUA and CBA metrics could then be compared to benchmarks such as WHO $/DALY guidelines for cost effectiveness of the breakeven point in CBA – net benefit, BCR>1 or return on investment >0.

Access Economics estimates that such an exercise would take some 6-9 months to complete with a ballpark cost of $150,000-$190,000 (plus data costs and GST, and allowing for the evaluation of up to three interventions).  Risks to timeframes and data lie primarily in access to data.  Data from LSAC can be purchased but costs could only be determined once a specific data request was submitted and quoted for.  Collaboration with AIFS would be required for access to the ATP database, which would involve additional inputs, expertise, opinions and analysis from AIFS researchers.  Estimates of timeframe or costs involved in this type of analysis could not be obtained from the researchers at AIFS at this preliminary stage, as a precise quote would depend on the scope and depth of analysis required.  It is likely however that these costs would be substantial ($10,000-$20,000 as a guesstimate).  A more precise quotation could be provided depending on consideration of the findings of this methodological scoping study.

Access Economics

June 2009

1.          Meaning and measurement

Working definitions and metrics undergird quantitative estimation of goods and services affected by positive family functioning (PFF) and negative family functioning (NFF).  Definitions from the literature relating to family functioning (FF) in general only measure one particular element of FF.  Thus definitions and measurements can vary quite significantly and are discussed further in Section 1.1 and Section 1.2.

The counterfactual of PFF (namely NFF) is used in the quantification and costing process.  This is because the benefits of PFF are the same as avoiding the costs of NFF.  In addition to this, the process of evaluating consequences of NFF has a simpler methodology.  While the key outcomes discussed in this scoping study may seem to be a simplification of the real consequences of NFF, they are aspects that can be quantitatively measured.  Although other aspects cannot (as of yet) be accurately quantified, they could be discussed in a qualitative manner.

1.1                        Definitions

The initial step of determining a definition for FF is to establish what is meant by the term ‘family’ for the purposes of this scoping study.  Edgar (1992) noted that the term ‘family’ itself carries degrees of ‘emotional baggage’ based on personal experiences as well as ideology.

It is also important to note that families are dynamic in nature and can follow an unstructured life cycle.  In general, families begin with the union of two people (not necessarily through marriage) and may move to a subsequent stage where infants or children are introduced.  The relationships formed between parents and children in the family can have origins that are either genetic or social (for example, in the case of adoption in contrast to biological parents).  The structures of these families are subject to change through separation, divorce, repartnering, remarrying and death.  Further, relationships within the family change as children age, leave the family home and commence families of their own.  Generally accepted definitions of family are broad, for example the Australian Bureau of Statistics (ABS) states that a family comprises (ABS, 2008);

q  couples with or without co-resident children of any age;

q  lone parents with co-resident children of any age; or

q  other families of related adults, such as brothers or sisters living together, where no couple or parent-child relationship exists.

The bulk of literature indicates that the major impacts of FF are borne by the children within the family unit.  Further to this, research has indicated that positive experiences in the early stages of the life cycle (where the role of family is greatest) increase the probability of positive future experiences and external social influences become dominant later in life (Sanson et al, 2002).  With this in mind, a greater emphasis is given to young families (where children have not reached adulthood) in this scoping study, with a particular focus on the long term outcomes for children as a result of their family environment. 

Children are unable to explicitly control their family environment and are often considered to be the main victims of NFF.  The implicit focus on outcomes associated with children as a result of NFF excludes some families from the evaluation.  In particular there is no focus on couple-only families, couple families that are expecting children (in-utero) or couples/individuals caring for older relatives.

FF is considered in this scoping study independent of these many forms of family structure.  Lippman et al (2007) discussed the importance of ‘father involvement’ as a family wellbeing indicator, while noting that the quality of the relationships in the mother-child-father triad (Fletcher, 2007) is also important. 

Defining PFF is a difficult task.  A review of the literature did not result in a clear, concise or consistent definition of FF.  Similar difficulties have been reported by other researchers.  For example, Fletcher (2007) noted that no developed or accepted model of ‘family wellbeing’ existed for the development of a family wellbeing framework in New Zealand.  In a review of family measures for paediatric psychology, Alderfer et al (2008) described the common themes relating to the functioning of the family system as follows.

‘At a basic level, families are conceptualised as organised systems that aim to keep balance and order through communication and assigned roles (Fiese et al, in press; Kazak et al, 2003).  When families are functioning well, roles are clear, communication is open and straightforward, and affect is well regulated.  In contrast, when a family is functioning poorly, it may, for example, respond to stress by becoming disorganised and chaotic with unfocused communication patterns and emotional dysregulation.  Alternatively, it may become over-controlled, with increased rigidity of roles, inadequate communication, and a restricted range of affect’ (Alderfer et al, 2008).

In a presentation that discussed the conceptual framework for the development of family wellbeing indicators, Lippman et al (2007) introduced three broad child and adolescent outcomes.  These categories were health and safety[1], education and learning[2] as well as social and emotional wellbeing[3].  Lippman et al (2007) placed a stronger emphasis on the domains of intra-familial relationships and external family relationships.  In particular, the conceptual framework of family wellbeing included aspects of family strengths, healthy couple relationships, positive parent‑child relationships, father involvement (especially for non-residential fathers) as well as connections between the child, family and community.

Positive and negative family influences and behaviours were discussed in a paper by Aufseeser et al (2006).  Four specific behaviours were discussed that are associated with direct influences by parents on children.

q  Parent-adolescent closeness and communication :  through close relationships, healthy open communications and perceived parental support.

q  Parental monitoring :  that included knowing the child’s whereabouts after school as well as their activities and friends.

q  Eating meals together :  as family mealtime was stated to be an important time for children to communicate and spend time with their parents.

q  Parental health behaviours :  through the provision of role modelling of positive as well as negative behaviours.

Fletcher (2007) presented the New Zealand Families Commission approach to defining family wellbeing, outlining four key social domains.  These domains were desired outcomes for policy interventions to address.

q  Positive family functioning would be bolstered by promoting parenting knowledge and skills, and would be further assisted by the promotion of positive family relationships.

q  Physical and social environments that promote family wellbeing would be addressed through the key areas of family violence prevention programs as well as the promotion of family-friendly environments.

q  Economic circumstances of families includes policy issues such as work/life balance as well as overall family income, and can be addressed through the promotion of flexible work conditions and environments as well as the provision of childcare services.

q  Family knowledge base could be increased through further research funding that examines family issues and effective interventions, community engagement and opportunities for community consultation.

Closely linked to concepts of family wellbeing are the greatly more established and studied concepts of child wellbeing.  Jespersen (2007) presented rankings of child wellbeing dimensions among OECD countries based on a UNICEF report.  Dimensions of child wellbeing were divided into six main areas.

q  Material wellbeing was measured through a number of metrics such as relative income poverty, households without an employed adult, reported deprivation (family affluence, education resources and books).

q  Health and safety canvassed three areas: first, infancy – through infant mortality rates and low birth weight; second, preventive health services – through immunisation rates; and, third, safety – through deaths from accidents and injuries.

q  Educational wellbeing took into account measures such as school achievement at 15 years of age, the continuation and further participation in education and training, as well as future aspirations for work.

q  Family and peer relationships were measured through influences such as the structure of the family unit, whether children participated in mealtimes and open conversations with parents, and whether their peers were perceived as ‘kind and helpful’.

q  Behaviours and risks comprised experiences and behaviours undertaken by children including eating and being healthy, risk of adverse behaviour such as sex, drugs and violence as well as the child’s experience of bullying.

q  Subjective wellbeing was measured as the child’s own views on their health status, whether they liked school and felt satisfied, and their own ranking of personal wellbeing.

These papers by Lippman et al (2007), Fletcher (2007), Aufseeser et al (2006) and Jespersen (2007) provide support for a broad definition of the characteristics associated with PFF.  Some authors indicated that caution was needed around some more conventional understandings of FF where cause and correlation may be potentially problematic.  In particular, Esping-Andersen (2008) stated that:

‘Conventional theory has emphasized monetary effects in general and poverty in particular.  This is without any doubt a major contributor to differential school success and, more generally, equal life chances.  But social scientists as well as policy makers have paid far less attention to non-economic factors in the inter-generational transmission of disadvantage.  Although research is on less than firm ground in this regard, there is a credible case to be made that non-economic mechanisms may be of equal if not greater importance than income.  To a degree, the two coincide: teen-age mothers, immigrants and low educated parents are also more likely to be income poor.  But we are almost certainly tapping two rather distinct dimensions, and this implies that a strategy based narrowly on income redistribution is unlikely to fully succeed.’

While the income and asset base of parents are key factors in the development of positive outcomes, FF itself is characterised by factors other than economic status.  A number of consistent themes arose in the literature describing the characteristics (or domains) which play a role in FF.  Consistent themes reoccurred in the literature around six main domains, which were:

q  emotional;

q  governance;

q  engagement and cognitive development;

q  physical health;

q  intra-familial relationships; and,

q  social connectivity.

The first four FF domains above focus on relationship elements between the parent(s) and the child.  The final two domains focus on relationships with other members of the immediate family, extended family and society.  Each of the FF domains is defined by a number of particular characteristics (Table 1‑1) that intrinsically overlap through the everyday activities and operations of the family unit. [4]  A summary is provided below.

FF – positive and negative – is defined through a variety of emotional attributes, family governance frameworks, cognitive engagement and development characteristics, physical health habits, intra-familial relationships and social connectedness.  PFF is characterised by emotional closeness, warmth, support and security; well-communicated and consistently applied age-appropriate expectations; stimulating and educational interactions; the cultivation and modelling of physical health promotion strategies; high quality relationships between all family members; and involvement of family members in community activities. 

Table 1 ‑ 1 :  Characteristics of FF domains

Source:  Access Economics in consultation with the ERG and FaHCSIA.

These domains should not be interpreted as being independent of each other.  Rather, they are characteristics that interact, complement and co-exist. 

This is particular evident when considering stressors of the family environment, such as maternal depression, which may result in a lack of emotional engagement between the mother and child, and may impart a number of other negative impacts on the family.  However, factors such as maternal depression are risk factors for poorer FF, rather than attributes that define it.

Another risk factor is the stress brought on by negative changes in economic conditions, such as the loss of employment by one or more of the parents in the family.  In this instance, increased conflict may build in the relationships between the parents, as well as reduction in social connectivity through a contraction of family activities.  A loss of motivation or self esteem through unemployment may impact the warmth of relationships between the parents and children, and so on.

It is important to distinguish, as per the above examples, between the internal or basic core functioning of the family and external influences on that functioning.  Much of the literature has examined the adverse impacts of external influences on the family, such as a significant change in health for one member of the family.

1.2                        Methods of measuring PFF

A number of Australian and international studies have attempted to measure a specific aspect of FF or the overall general functioning of the family.  It is the latter of these two metric types that is most relevant to this study, although specific aspects provide important insights into FF. For example, the Mealtime Interaction Coding System (MICS) measures interactions, roles and the structure of family meal time, but is not able to gauge the level of social connectivity within the family. Broadly, metrics of general FF take into account the following components (Alderfer et al, 2008):

q  family organisation (roles, leadership and alliance formation);

q  cohesion (involvement and closeness);

q  communication (clarity of expression and directness);

q  affective environment (expression of feelings and conflict), and;

q  problem solving ability (goal-directed negotiation and task accomplishment).

The main models of FF are outlined in Table 1‑2 with their respective measurement metrics.  These FF models have different constructs and approaches to conceptualising and measuring the influences and outcomes of FF.

Table 1 ‑ 2 :  Theoretical models of general FF

Source:  Alderfer et al (2008)

[Obs] – observational measure, [SR] – self reported measure

FF metrics are estimated in a number of different ways, such as through self-reported questionnaires, coded observed interactions and interviews, each providing a different window of information into FF (Alderfer et al, 2008).  In Table 1‑2 above, measures are generally observational or self-reported in nature.  A trained professional’s opinion on the functionality of a family can differ from a family member’s opinion of their environment.

Based on the descriptions in Table 1‑1, the theoretical models that best align with these characteristics are the McMaster Model of Family Functioning (MMFF), Iowa Family Interaction Rating Scale (IFIRS) and The System for Coding Interactions and Family Functioning.  Of these the McMaster Model is the most frequently used in Australia.  This model, particularly in the self-reported Family Assessment Device (FAD) format has been used in a number of Australian studies including:

q  the New South Wales Population Health Survey 2005-2006;

q  the New South Wales Child Health Survey 2001; and

q  the Western Australian Child Health Survey 1993.

The FAD metric generates a result between 1 (healthy FF) and 4 (unhealthy FF).  Parents and/or caregivers in the family are provided a list of 12 statements and asked whether they strongly agree, agree, disagree or strongly disagree with each one, within their own family environment.  The list of statements is as follows.

q  Planning family activities is difficult because we misunderstand each other.

q  In times of crisis we can turn to each other for support.

q  We cannot talk to each other about sadness we feel.

q  Individuals in the family are accepted for who they are.

q  We avoid discussing our fears and concerns.

q  We express feelings to each other.

q  There are lots of bad feelings in our family.

q  We feel accepted for who we are.

q  Making decisions is a problem in our family.

q  We are able to make decisions about how to solve problems.

q  We don’t get on well together.

q  We confide in each other.

The New South Wales Population Health Survey (2005-2006) reported that overall 94.8% of children aged between 0 to 15 years lived in households with healthy FF.  Variations between age groups, socioeconomic status as well as rural and urban areas did not show significant variations.

The New South Wales Child Health Survey (2001) reported a medium score of 1.65 for FF on the FAD measurement, reflecting overall high levels of healthy FF.  Groups that showed a higher level of healthy FF were those that had a mother with tertiary level education qualifications (1.58), where all children were over the age of five years (1.58) and where the family was in the least socially-disadvantaged quintile by SEIFA index of postcode.

The Western Australia Child Health Survey (1993) included the McMaster FAD scale in its series of questions, although the results on the overall level of FF from this survey could not be obtained.

Further to these Australian population based studies, the National Survey of Mental Health and Wellbeing (2007) collected generic measures of social interaction with family members.  Only frequency of interactions, as well as mode of contact was collected (for example, through visits, phone calls, emails, etc) with only the recorded ability to confide in these members a reflection of the quality of the relationship. 

1.3                        Incidence and prevalence approaches

There are two main approaches to measuring the value of PFF, which is equivalent to estimating the cost of NFF.

q  The incidence approach measures the number of new cases of NFF in a base period (say, 2009) and attributes costs due to that NFF, both financial and non-financial, over the affected persons’ lifetimes.  The total cost of NFF represents the net present value (NPV) of current and future costs incurred due to all new cases in that year.

q  The prevalence approach measures the total number of people ever affected by NFF in a base period and costs associated with those cases, both financial and non‑financial costs, in that year. 

Figure 1‑1 depicts the conceptual differences between these two approaches with 2009 depicted as the base year.

Figure 1 ‑ 1 :  Prevalence and incidence approaches to cost measurement

These methodologies present information in slightly different ways and have different applications.  The prevalence approach is a ‘snapshot’ as it reports impacts in one year and, in general, prevalence estimates can be easier to understand by policy makers as they relate to a single point in time.  Also, the financial value of impacts can be compared to measures such as gross domestic product (GDP).  The incidence approach reports the lifetime impacts associated with a condition; as policy interventions are focused at preventing NFF, this methodology is most relevant for its application in CBA or CEA studies.

Aetiology, data sources and the study design used to track outcomes are key deciding factors determining which approach is best.  In this instance an incidence approach is considered most appropriate, due to these factors and the lag times between exposure to the family environment and associated outcomes. 

A potential difficulty with the incidence approach is conceptually defining a ‘new case’ of NFF.  It is easier to define the incidence of a disease where an initiation of a change in a person’s health state is clear, in this case NFF exposure to this environment (or risk factor) occurs over a potentially long period of time, with some fluctuations in the family environment resulting from external factors (as the result of changes in health, social or economic circumstances).  Ideally, to alleviate this problem there would be a survey question that asked respondents when NFF first began, or when they first became aware of it.  In the absence of such data, incidence could be proxied from a snapshot of children exposed to NFF in a base year (one year current prevalence), together with estimates of the average duration of exposure to NFF.

2.   Measuring outcomes

Impacts from FF occur with lags.  Health, economic and social/relational outcomes do not necessarily occur straight away but can take many years of exposure before they manifest in later years, and then can extend for varying periods – sometimes for life.  The literature review process revealed three broad areas of outcomes.

q  Health outcomes were mostly observed through the occurrence of mental illnesses such as anxiety and depression later in life, but also included eating disorders, health behaviours (e.g. unsafe sex, physical inactivity, overweight and obesity) and substance abuse (e.g. smoking, alcohol and drug abuse), with the consequent physical impacts of these risk factors on morbidity and mortality outcomes.

q  Productivity outcomes were reflected in rates of labour force participation, employment and hourly wage rates, with a number of intermediate measures reported in the literature, such as reduced levels of literacy and numeracy and other measures of educational achievement.

q  Social outcomes were measured primarily through their negative manifestations –involvement in antisocial behaviour such as delinquency, and the probability of criminal behaviour during youth and later in life.  In contrast it was harder to quantitatively associate PFF with positive manifestations such as the quality of inter-personal relationships and future community contributions.

Section 2.1 reviews the current literature quantifying causal links between NFF and these outcomes.  Multiple interactions between the characteristic domains outlined in Table 1‑1 influence these outcomes, obscuring causal pathways.  Section 2.2 examines the appropriateness of Australian data sources for quantification of causal linkages.  Section 2.3 discusses theories behind the units of measurement for the costing exercise.

Information from Australian and international studies was included in the literature review.  Specific Australian sub‑population groups such as Indigenous Australians were not examined.  Studies specifically looking at FF within the Aboriginal and Torres Strait Islander (ATSI) context have only recently been initiated.  One such example is the Western Australian Aboriginal Child Health Survey (WAACHS), which was conducted between May 2000 and June 2002, providing a snapshot of the ATSI population.  This study noted several factors influencing health and wellbeing in the ATSI population that were not present in an earlier Western Australian Child Health Survey.  According to Blair et al (2005) these were:

q  remoteness – 10% of all ATSI children lived in areas of extreme isolation, 44% lived in areas with some level of isolation and 34% lived with no isolation;

q  forced separation (or forced relocation) – just over 40% of children were in households affected by forced separation or forced relocation of at least one primary or secondary carer or grandparent;

q  major life stressors – a high proportion of ATSI children were affected by death, separation and divorce; and

q  financial strain – a high proportion of ATSI families suffering from financial strain.

A lack of sufficient information limits the extent to which the current study can be applied to the ATSI population.  FaHCSIA has initiated a longitudinal study Footprints in Time – Longitudinal Study of Indigenous Children (LSIC) that will (in time) provide pivotal information for the analysis of FF issues as well as the causality of ATSI child health and wellbeing.

2.1                        Literature findings

2.1.1                   Health outcomes

NFF can naturally have short term impacts on health and these are relatively easy to estimate from Australian data sources.  For example, the health impacts of child abuse that results in immediate injuries has been estimated in Taylor et al (2008).  However, short term health impacts for children due to NFF may lead to long term impacts on their future productivity through reduced educational achievement as a result of cognitive problems and non-attendance at school (Stam 2005).

In addition, studies of the impacts of FF on medium to long term health outcomes of children show affects in a number of areas (which can be harder to estimate), including:

q  eating disorders;

q  substance abuse; and

q  anxiety, depression and psychological issues.

Different aspects of eating disorders have been examined in the literature.  Parental influence was reported to have a significant role in determining the preferences and choices for different types of food in their children (Hill 2002; Benton, 2004).  These preferences were hypothesised by Benton (2004) to have a significant role in determining the development of obesity.  Parents, peers and siblings act as role models encouraging the tasting of novel foods, although forcing a child to eat a certain type of food had a negative impact on their preference.  Hill (2002) discussed very similar qualitative findings in the determination of preferences, with some negative impacts of parental techniques observed in developing food preferences.  Further to this, tactics such as the use of fatty foods as treats after a healthy food has been consumed, only reinforced some foods being viewed in a negative way by children. 

Apart from tactics employed by parents that aim to instil food preferences in children, Ringer and Crittenden (2007) examined correlations between eating disorders and family attachment.  Qualitative findings from this study showed that children who were less attached to their parents had higher rates of eating disorders. 

Snoek et al (2007) agreed that parents influence their children’s emotional eating behaviour through role modelling processes.  This study showed that higher maternal support was associated with lower emotional eating in younger adolescents while higher psychological control was associated with higher emotional eating in younger and older adolescents.

Parental influences were also shown to have a significant impact on rates of substance abuse (alcohol, drug and cigarette smoking addictions).  Garmiene et al (2006) showed that joint family activity time deficit together with parents who smoke or drink excessively underlies the development of alcohol and smoking additions in children.  Maternal influence was a strong factor for the uptake of smoking by sons (OR = 2.50, 95% CI 1.11‑5.64).  Sons appeared to be influenced (to a greater extent) in their consumption of alcohol by their parents, particularly by their father (Table 2‑1).

Table 2 ‑ 1 :  Maternal and paternal influences on alcohol consumption

Source:  Garmiene et al (2006).

Overall family time was also a significant influence on smoking and alcohol use attempts (Figure 2‑1).  Children who were exposed to less family time had higher rates of alcohol usage as well as attempts of smoking.  Garminene et al (2006) hypothesised that harmonious child‑parent interaction, open communication and parent support are factors that underlie a child’s successful mental and physical health development.

Figure2 ‑ 1 :  Influence of family time on smoking (left) and drinking (right) attempts

Source:  Garmiene et al (2006). * p<0.05 in comparison with ‘good family time’.

Parental behaviour that included smoking by the parent and other factors such as the provision of pocket money and allowance of smoking in the home were shown to be determinants of adolescent smoking in New Zealand by Scragg and Laugesen (2007).  Parental influences appeared to be stronger in people with European backgrounds and overall were able to explain approximately two thirds of adolescent smoking in New Zealand.  Exposure to peers who smoked significantly increased the relative risk of adolescent smoking (Table 2‑2).

Table 2 ‑ 2 :  Adolescent smoking determinants in New Zealand

Source:  Scragg and Laugesen (2007).

Correlations between parental influence or behaviour and adolescent smoking were examined by Wen et al (2005).  A measure of tender loving care (TLC) [5] was used in this study as a measure of FF.  Adolescents who had parents who smoked and low levels of TLC in the household displayed the highest rates of smoking.  Those with non-smoking parents and high levels of TLC had the lowest smoking prevalence rates (Table 2‑3).

Table2 ‑ 3 :  Rate ratios for comparing the presence or absence of parental and peer factors influencing smoking status

Source:  Wen et al (2005).

Parental limit setting, family conflict and perceptions of family experience were examined by Wu et al (2004), with a view to establishing their impacts on the severity of alcohol and drug problems.  Wu et al (2004) completed an interview at a chemical dependency program that was part of a health maintenance organisation (HMO).  The study showed that family conflict, limit setting and PFF scales were correlated with substance dependence.  Gender differences showed that girls had more dependence symptoms, psychiatric diagnoses (depression, anxiety, conduct disorder) and substance abusing family members.

Table2 ‑ 4 :  Multivariate regression results predicting substance abuse/dependence symptoms

Source:  Wu et al (2004).

Psychological disorders have also been a study outcome associated with FF.  Bogels and Brenchman‑Toussaint (2005) discussed an observed high correlation between anxiety disorders in parents and elevated rates of anxiety disorders in their children compared to the rest of the population.  These outcomes may be the result of a number of determinants, including an insecure attachment of the child to its caregiver.  Likewise, if parents are insecurely attached to each other, or their children, the child is more likely to be anxious.  Further to this, Bogels and Brenchman‑Toussaint (2005) found that parental conflict as well as marital quality disrupts bonds with children and serves as a stressor to a child’s environment, threatening their sense of security.

Specific parenting styles and personality traits were examined for their impact on the child’s happiness by Furnham and Cheng (2000).  This study found that self-esteem and happiness were directly correlated and that reasonable discipline exercised by mothers towards their children (as well as parental authority) assists with enhancing their child’s self-esteem.

Figure 2 ‑ 2 :  Pathways predicting ‘happiness’ using different aspects of mediation

Source:  Furnham and Cheng (2000).

Conditions affecting the family environment such as maternal depression or separation of parents have been shown to influence outcomes for children.  Family structure is widely studied in the literature and separation can proxy a poor relationship between parents (noting that poor relationships can also occur without separation).  Hipwell et al (2004) analysed the effect of postnatal depression and parental conflict on the social functioning and future development of children.  Postnatal depression disturbs mother and infant engagement and interactional style, but is only one reason for parental conflict between each other or the child.  Hipwell et al (2004) examined the impacts on school aged children who have been exposed to maternal depression as well as marital conflict.  Impacts of maternal depression and martial conflict were assessed using linear regression (Table 2‑5) as well as logistic regression (Table 2‑6).  Hipwell et al (2004) concluded that children exposed to marital problems or postnatal depression are at risk of development problems through childhood, especially conflict in peer play among children.

Table2 ‑ 5 :  Significant influences of independent variables on time-sampled pee r interaction variables, using linear regression

Source:  Hipwell et al (2004).

Table 2 ‑ 6 :  Significant influences of independent variables on event-sampled peer interaction variables, using logistic regressi on

Source:  Hipwell et al (2004).

Silburn and Zubrick (1996) used information from the Western Australian Child Health Survey, to determine health impacts associated with different family environments, structures and disciplinary styles.  Reported mental health problems related to categories of disorders such as anxiety and depressive, conduct disorder, attention deficit-hyperactivity disorder (ADHD) and somatisation disorder, identified by the parent or teacher on the Achenbach Child Behaviour Checklist within a clinical range.  Silburn and Zubrick (1996) noted that data from these studies should be interpreted with caution as they represented a correlation not a causal linkage.

‘[The survey] merely provides a ‘snap-shot’ of the average mental health status of the children who are living in different family living arrangements at a particular point in time.  These findings tell us nothing about the longer term outcomes for individual children…..The mechanism whereby children in non-traditional family forms come to have higher risks for adverse mental health and poorer educational outcomes requires a consideration of other factors’.

Interestingly, statistical analysis of the Western Australian Child Health Survey data set only showed a statistically significant relationship between three family environment measures.  These related to the family type, parental disciplinary style as well as the level of family discord.  No statistically significant association was shown with family income in predicting mental health status.  Table 2‑7 shows the odds ratios from family environment associated with the mental health problems.

Table 2 ‑ 7 :  Family risk factors associated with child mental health problems

Source:  Silburn and Zubrick (1996).

Note:  Risks are calculated relative to the least adverse combination of family risk factor, i.e. encouraging parenting style, original family type and low family discord.

Silburn and Zubrick (1996) noted that ‘regardless of income, children in one parent and step/blended families have a similar increased risk of mental health problems in comparison with children in original families when similar conditions of family discord and parental disciplinary style apply.  Thus in predicting or preventing child mental health problems, what happens between people within families is just as important as the family structure the child happened to be living in’.

2.1.2                   Productivity outcomes

Productivity outcomes resulting from the family environment are rooted in human capital theories.  Carneiro and Heckman (2003) showed that interventions made early in the lifecycle provided the most optimal investments in terms of human capital outcomes (Figure 2‑3). 

Figure2 ‑ 3 :  Optimal investments in human capital, by lifecycle

Source:  Carneiro and Heckman (2003).

This shift in human capital investment strategy has been noted by a number of authors.  Notable, Esping-Anderson (2008), stated that:

‘Human capital investments have, over the past half century, been almost exclusively directed at formal education.  It is only quite recently that we have come to realise that the foundations of learning – as well as the chief mainsprings of inequalities – lie buried in the pre-school phase of childhood and that schools are generally ill-equipped to remedy a bad start’.

A framework was produced by Amuwo et al (2004) to examine the effectiveness of parental discipline in producing desirable child outcomes.  The overarching theory was that the links between discipline and performance is really the result of two linkages – first, from discipline and other parental inputs to child effort and, second, from child effort to child performance.  Other factors such as governance and communication aid in the facilitation of these processes. 

Outcomes are described as being near-term (that is, within a generation) and include better child-family relationships and performance at school.  Alternatively, outcomes can be long term (that is, intergenerational) and include higher incomes earned by children when they become adults.

Socioeconomic status, parental involvement at school as well as parental income and asset base have been found by a number of studies to be strongly associated with outcomes of better school performance.  These factors may be strongly correlated (and indeed may proxy) other characteristics of FF.

For example, Ermisch (2008) found that the parent’s socioeconomic status affects children’s intellectual, emotional and behavioural development, which in turn has implications in subsequent years as adults.  This study found that higher income parents had children aged three years old with better cognitive and behavioural assessments, while another significant finding was that parents on a low income spent less time reading and teaching their children.

Parental socioeconomic status (high incomes of parents) was associated with a greater probability of parents engaging with children more.  Figure 2‑4 shows the proportions of parents in different income brackets associated with:

q  taking them to the library;

q  partaking in other educational activities (Eda4);

q  positive parenting style (Pstyle); and,

q  foetal growth of the child before birth (Fetgr) – a useful indicator of the health and nutrition of baby and parent.

Figure 2 ‑ 4 :  Parental inputs (age three) by household income (age nine months)

Source:  Ermisch (2008).

Income brackets (refers to when the child was aged nine months) – ‘less than £10,400’, ‘£10,400 to £20,800’, ‘£28,800 to £31,200’ and ‘more than £31,200’.

Ermisch (2008) found a significant correlation between parents’ household income and favourable parenting practices, which in turn is linked to early cognitive development for the child and thus a higher chance of lifelong economic success.

Zhan (2005) reported that associations between the parent’s assets and the children’s education attainment had a stronger association than parental income and was also positively associated with their expectations and involvement in school activities.  It was hypothesised that the parental asset base provided a better estimate of financial security within the family home and possibly provided a proxy for other characteristics such as the provision of books and other resources.  Measures of parental expectations mediated the processes in which school achievements were made.

2.1.3                   Social/criminality outcomes

In Australia, there has been a shift in the perspective of the aetiology of criminology towards that of a ‘criminal career’.  This new outlook suggests that criminal offences start occurring at one point in time and progress to another, with the frequency and severity of these offences dependent on the perpetrator’s past experiences and current circumstances (Stewart et al, 2004).

There are a number of risk factors involved in the development of a criminal career as well as a number of protective factors in the life transitions and trajectories for children and young people (Stewart et al, 2004).  Moffitt (1993) suggested that there may be two types of offenders - the ‘life-course persistent type’ and the ‘adolescent limited type’.

Developmental crime prevention is aimed at providing (or promoting) positive interventions at transition points in a child’s life as well as reducing exposure to risk factors.  Buffers such as support, advice, safety or opportunities may prevent the child from traversing a negative pathway.

The family environment has been accredited to a higher probability of being involved in criminality, through involvement in delinquency during adolescence and the development of criminal careers later in life. 

A longitudinal study examined by Catalano et al (2007) looked at the direct and indirect relationships between protective factors in the family environment during early and later adolescence and self-reported crime in the first year post-high school.  The study showed that children who were more bonded to parents in early- and mid- adolescence and whose bonding to parents decreased less during early adolescence were less likely to report criminal behaviour in emerging adulthood.  Level and change in family management practices in both early and later adolescence predicted criminal behaviour in emerging adulthood.

Mediation models indicated that effects of early adolescent family environment on later criminal behaviour were mediated by late adolescent family environment.  In turn, effects of late adolescent family environment were largely indirect, through level of criminal behaviour at the end of high school.  In other words, children who exhibited criminal behaviour at the end of high school – which is affected by late adolescent family environment – were more likely to continue with that criminal behaviour into early adulthood.

A number of studies examined the linkages between family structure on the probability of teenage promiscuity, substance use and crime.  For example, Antecol and Bedard (2002) found that when paternal presence was measured as a continuous variable there are statistically significant reductions in youth participation in smoking, drinking drugs, sexual intercourse and conviction before the age of 15 years, associated with the length of time the biological father remains in the household.  Antecol and Bedard (2002) noted that there were exceptions to this, when the father remains in the house but has such a negative relationship with the mother and children that youth outcomes are worse.  These findings add weight to the notion that it is the quality of the relationship between the parents and child as well as between the parents that is of importance in determining childhood outcomes.

Antecol and Bedard (2002) found that an additional five years with the biological father decreases the probability of smoking by 5.3%, drinking by 1.2%, engaging in sexual intercourse by 3.4%, marijuana use by 2.2%, and conviction by 0.3%.  However, youth outcomes could also be worse if family stress due to tension and arguing is higher with the father in the house.

Fergusson and Horwood (2002) studied trajectories of criminology from a longitudinal study conducted in New Zealand which followed 896 children from birth to the age of 21 years.  The underlying theory was developed from Moffit (1993) and the identification of different types of offenders.  Fergusson and Harwood (2002) noted that in addition to the Moffit (1993) criminal career theory a number of other studies had suggested that:

‘...early onset delinquency is largely shaped by a series of family processes, including inept parental discipline, poor parental monitoring and impaired family problem solving, that creates a socialisation process in which children learn that coercive and antisocial behaviours have an adaptive value.  In contrast, those who show late (after 14) onset offending lack the marked family difficulties that are present for the early onset group but are nonetheless exposed to greater family difficulties than non-offenders.  The moderate risk status of this group is exacerbated by the formation of affiliations with delinquent peers who may act to encourage, reward, and sustain tendencies to antisocial behaviours’.

Fergusson and Horwood (2002) identified five distinct groups of offenders in their New Zealand study.

q  The low-risk trajectory group had very low levels of conduct problems (less than 2%) during mid-childhood (8‑10 years).  Offending was scarce during adolescence but increased slightly at 17 years, declining thereafter.  This group had a low rate of offending with a reported 1.6 offences over the period from 14 to 20 years.

q  The early onset trajectory group had relatively low rates of conduct problems during mid-childhood (approximately 8%).  The group displayed a moderate risk of offending in early adolescence peaking around the age of 13 years and declining to a lower risk around the age of 17 years onward.  This group reported a mean of 4.8 offences over the period from 14 to 20 years.

q  The intermediate onset trajectory group, mostly did not display any conduct problems.  This group had a similar offending profile to the early onset group, but started offending during mid-childhood.  This group had a very clear peak of offending between the years 14 to 17, with a peak risk of 91% occurring at the age of 17.  This group had a much greater rate of offending overall with a mean of 48.8 offences reported between the ages of 14 and 20.

q  The late onset trajectory group had a relatively low rate of conduct problems (12%) and had an offending profile with low risk prior to 17 years of age, that increased to approximately 85% at 17 years and declined thereafter.  This group had an offending rate that was less than the intermediate but greater than the first two, with a mean rate of 24.1 offences.

q  The chronic offender trajectory group had a majority of people with conduct disorder (58%).  High rates of offending were observed between 11 to 17 years with a small decline at 20 years of age.  This group had a dramatically higher offending rate, with members reporting an average of 141.1 offences between the age of 14 and 20 years (35.3 offences per annum).

Overall, Fergusson and Horwood (2002) found that increasing criminal tendencies were associated with male gender, parental criminality, parental conflict, tendencies to novelty-seeking behaviours, low intelligence and low self esteem.  Figure 2‑5 shows the associations between sub-types of adolescent-limited offenders to FF risk factors.  These increased with the rate of offending, except for late onset offenders, who were more likely to be influenced by their peers and/or low self esteem.

Figure 2 ‑ 5 :  Comparison of offending trajectory groups on measures of FF

Source:  Fergusson and Horwood (2002).

2.1.4                   Summary of literature based findings

The literature review provided a valuable insight into the outcomes associated with FF, while also highlighting a number of methodological problems in potential application to valuing PFF. 

q  First, while many studies reported correlation between aspects of FF and outcomes, it was sometimes difficult to ascribe this to causation.  This was more of a problem for studies set up as cross-sectional analyses, although even with longitudinal studies it was still difficult to adequately control for factors that influence both and can change over time, such as socioeconomic factors.

q  Second, some outcomes cannot be easily converted into economic costs.  For example, literature generally examined outcomes associated with reaching interim literacy and numeracy standards, rather than completion of Year 12 (which can in turn be linked to later income), or indeed to adult employment, labour force participation or earnings.

q  Finally, different and inconsistent measures of FF and of outcomes have been applied in past studies.  The application of findings would thus result in somewhat of a mosaic of varying and overlapping definitional interpretations.

Interestingly, high level trends are evident from the literature.

q  Health outcomes appear to be most dependent on issues such as family attachment, joint family time and role modelling.  These characteristics have been described most frequently in the literature as underlying substance abuse and psychological outcomes in children. 

q  Productivity outcomes appear to be most frequently described as associated with socioeconomic status of the family (parental income and asset base) and parental involvement at school.  As noted by Ermisch (2008), these factors are associated with a greater probability of parents engaging with their children more.

q  Social/criminality outcomes were most frequently described in the context of the quality of the relationship between the parents as well as the level of bonding between the parents and children.  Factors such as the length of time the biological father remained in the family home, number of changes of parents, parental conflict, and poor social behaviours by the parents (criminality, alcoholism and illicit drug use) were all shown to be associated with a greater likelihood of criminality by children.

2.2                        Longitudinal child and youth studies

A number of longitudinal Australian as well as international studies of children and youth have either been completed or are currently underway with available interim results (Table 2‑8).  Large longitudinal studies offer the potential to specifically measure the impacts of FF on a range of social, health and economic outcomes on the child as well as the family. 

‘Longitudinal data are essential to answer many of the questions facing policy makers and researchers today.  We know that the roots of many problems in adolescence and adulthood can be found in early childhood.  The study will provide data tracking children over time, to help researchers to understand how, why and when children embark on pathways to adaptive and maladaptive outcomes and where the opportunities are to help children move onto better pathways’ (Sanson et al, 2002).

Controlling potential confounding variables enables an analysis of causal impacts. However, not all child and youth related longitudinal studies collected data on FF.  Some studies collected FF information that corresponded to a number of aspects of the family environment, but did not include any specific measures of FF.

In addition to this are considerations about the ease of access to data records.  A number of the studies listed in Table 2‑8 are from private research groups or otherwise not in the public domain.  Access to these data records will require specific information requests to the research organisation.  Costs will likely be associated with data provision, if it is permitted.

International differences between social, policy and economic environments mean that Australian data are strongly preferred.  That said, Björklund et al (2006) found that estimates of the negative relationship between living in a non-intact family and child outcomes were remarkably similar between the USA and Sweden (two countries with significantly different social and policy structures).  If Australian data sources do not provide the information required, international data sources could be examined for their ability to provide appropriate information, however this scoping study has only analysed the appropriateness of the identified Australian databases:

q  the Longitudinal Study of Australian Children (LSAC) – Section 2.2.1;

q  the Longitudinal Study of Australian Youth (LSAY) – Section 2.2.2;

q  the Australian Temperament Project (ATP) – Section 2.2.3; and

q  the Household, Income and Labour Dynamics in Australia (HILDA) study – Section 2.2.4.

Table 2 ‑ 8 :  Child and youth related longitudinal studies

Source:  www.aifs.gov.au/growingup/links.html and Access Economics

2.2.1                   Longitudinal Study of Australian Children (LSAC)

The LSAC database aims to provide evidence for a better understanding of Australian children’s development in the current social, economic and cultural environment, as an input to policy and practice regarding children and their families.

The LSAC measures child outcomes as: behavioural and emotional adjustment; language and cognitive development; and social competence.  The study examines the role of FF or processes over time and interactively, to determine the extent to which these vary within the population and how they impact on child development outcomes.  This information is collected through self-reporting and also observational measures (such as parent-child interactions).

There are 14 specific research questions in the LSAC.  The first key research question focuses particularly on the impacts of family relationships, composition and dynamics on child outcomes, and changes to these over time.  The question includes the analysis of:

q  the size and make-up of family;

q  the involvement of extended family;

q  roles of family members;

q  character of parental relationships and level of conflict in the family;

q  parenting practices;

q  child’s temperament;

q  impact of family break-up and re-formation; and

q  family coping strategies, particularly in times of stress.

The data pertaining to this first key research question in the LSAC are likely to be very relevant for quantifying the value of goods and services as a result of PFF, as they help to provide information on FF and child behaviour.  The LSAC is particularly useful for identifying opportunities for prevention, early intervention or other changes to policy. 

Other relevant research questions in the LSAC include:

q  Key research question 3: the impact of the family’s economic status, workforce status and poverty on the child;

q  Key research question 4: how beliefs and expectations of children impact on child outcomes;

q  Key research question 6: how connected families are to wider social networks, and how important family and child social connections are to child outcomes;

q  Key research questions 7-10: the wellbeing and health of children and families;

q  Key research questions 12-13: education opportunities for children; and

q  Key research question 14: how the interactions among factors in FF, health, non-parental care and education affect child outcomes.

LSAC data cover factors influencing developmental outcomes that relate to:

q  the child, such as health, temperament and literacy experiences;

q  the parents, such as health, socioeconomic status and parenting style; and

q  the broader family, child care, school and community environments.

In addition, LSAC provides information on key life events that can lift or depress a developmental trajectory (such as illness or injury, entry to child care or school, separation or divorce of parents).

A longitudinal design is essential to explore the developmental sequences that place children at risk and to illuminate protective factors.  A multiple cohort cross-sectional design has been selected, providing for coverage of two cohorts expected to be as follows;

q  a minimum of 5,000 children aged less than 12 months, selected in 2003 and followed at least every two years until they reach six to seven years of age in 2009; and

q  a minimum of 5,000 children aged four to five years when selected in 2003, and also followed at least every two years until they reach the age of 11-12 years of age in 2009.

The LSAC is funded by the Australian government and managed by FaHCSIA – $20.2 million over nine years with research to be completed in 2009 (although there is potential for an extension beyond Wave 3, with Wave 4 due to take place in 2009-2010).  It is intended that the data will be widely available to researchers following the conclusion of the study.  Waves of datasets from each two year observation period can be obtained from the Australian Institute of Family Studies (AIFS) website at http://www.aifs.gov.au/growingup/data.html.  Wave 2.5 has already been released:  Wave 3 research is still being undertaken.  Box 2‑1 outlines relevant specific research questions being investigated in LSAC.

Source:  Sanson et al (2002)

The age of the study cohorts by Wave 4 in 2010 of the data collection will be 6‑7 years in the infant group and 10‑11 years in the child cohort.  This database would therefore be useful in determining the short to medium term impacts of FF on children in the family, although it is currently limited in its usefulness to determine long term productivity impacts through completion of Year 12, further (tertiary) education, participation in the labour force and employment.  The same limitations apply to health and social impacts.  Inferences from LSAC would be able to inform some of the intermediate and transitional outcomes from FF but could not (until further Waves of the study are completed) quantify final outcomes directly.

Funding arrangements for LSAC have been substantially extended, which will see the current cohorts progress into and complete secondary education levels.  LSAC will be a critical study to inform long term outcomes of potential policy decisions in future years.

2.2.2                   Longitudinal Survey of Australian Youth (LSAY)

The LSAY database provides information on Australian youths transitioning through secondary school, into further education or training as well as into the labour market and adult life.  This study is funded by the Department of Education Science and Training (DEST) and jointly managed by DEST and the Australian Council for Educational Research (ACER).  The aim of this study is to determine which factors influence the participant’s pathways through education and employment.

This study began in 1995 with the selection of a national sample of 13,000 students in Year 9, with further cohorts selected in 1998 and 2003.  Participants were interviewed annually from approximately 15 years of age until around 25 years of age.  Information was provided through telephone surveys on their education, training and jobs as well as other facets of their lives. 

Unfortunately, no measure of FF was included in this survey, so LSAY does not have a direct use for this project.

2.2.3                   Australian Temperament Project (ATP)

The ATP is longitudinal by design, studying the psychological development of a large and representative cohort of Australian children born in Victoria between September 1982 and January 1983.  The aim of this study is to determine pathways of psychological adjustment and maladjustment across the cohort’s lifespan, along with the interplay of personal, familial and environmental factors with development and wellbeing.

This project is currently being funded by AIFS, but has received support from NHMRC and ARC grants as well as the Murdoch Foundation of the Royal Children’s Hospital.  Currently 13 waves of information have been collated through mail surveys from the age of 4‑8 months until 19‑20 years.  The first four waves of data were collected on an annual basis from infancy until three to four years of age.  Data collections from then on (from primary school) have been collected biennially along with an additional assessment during the first year of secondary school. 

There are approximately two thirds of families still participating in the study after 21 years.  Further information about this project is outlined in Box 2‑2.

Source:  Logie et al (2004).

As part of this project researchers have already established a number of longitudinal pathways in which FF plays a part.  Two examples of these pathways for depression (Figure 2‑6) and antisocial behaviour (Figure 2‑7), were presented by Prior et al (2005).  FF measures[6] in these pathways work alongside temperament, school, behavioural problems as well as peer based influences, to result in final outcomes such as depression and anti-social behaviour (which includes criminality).

Other outcome measures that have been examined in this study include substance abuse, risky driving and measures of social capital, such as positive social functioning, engagement in the community, as well as positive and anti-social peer relationships.

Figure 2 ‑ 6 :  Longitudinal pathways to adolescent depression (15-16 years), example only

Figure2 ‑ 7 :  Longitudinal pathways to adolescent antisocial behaviour (15-16 years), example only

Source (both charts):  Prior et al (2005).

Measures of FF and parenting have only been collected since the cohort of children were in their early teens (approximately 12‑13 years of age) and do not extend to the earlier years of childhood and infancy.  This aspect limits the ability of ATP to provide information on the short to mid-term outcomes of FF.

Through personal communication with Diana Smart, General Manager (Research) at the AIFS, it was established that, in order to access these data sources a research collaboration would be required with the analysis to be undertaken by AIFS.  This was the standard approach used for this particular study, which is required in order to protect the confidentiality of the study participants. 

There are some parallels between the measures of FF in ATP and in LSAC. However, these are limited and would require one or two more additional waves of data to bridge the age gaps. 

2.2.4                   Household, Income and Labour Dynamics inAustralia(HILDA)

HILDA is an annual survey that was initiated in 2001, commissioned and funded through FaHCSIA and managed through the Melbourne Institute at the University of Melbourne.  The study follows a longitudinal design with a representative sample of Australian residents (approximately 14,000 respondents from wave 1).  Information on four consistent overlapping themes are recorded in each wave of the study, with additional information collected depending upon government policy and particular issues around that time period.  The four consistent themes covered in HILDA are;

q  households and family life;

q  incomes and economic wellbeing;

q  labour market outcomes; and

q  life satisfaction, health and wellbeing.

The first of these areas has potential relevance for this project.  It incorporates a description of changes in individuals’ family structures over time, changes in marital status and marital satisfaction, child care issues and major life events experienced by respondents in the year leading up to the interview.  Information on parenting stress as well as work-family stress is also collected in this study. 

In HILDA it would be difficult to control for confounding influences on the relationship between the family environment variables and outcomes – influences such as school environment, peers and the quality of parent-child relationships.  In addition, the relatively short time period of the study would pose a problem for examining the full impact of the family environment on later outcomes for the children.

HILDA could be used to examine parenting and work-family stresses on the parents’ overall participation in the labour force, but is otherwise limited in its applicability for this purpose.

2.2.5                   Summary of Australian longitudinal data sources

Two of these studies (LSAC and ATP) have the potential to be useful for the determination of causal impacts of FF.  As LSAY did not collect information on FF variables, this longitudinal survey is unfortunately of little use here and HILDA would be limited to examining relationships between the family environment and parent’s participation in the labour force.  LSAC has the advantage of containing a breadth of data, valuable for controlling purposes.  However, a disadvantage is the relatively short timeframe of data collection.  LSAC would only be useful in providing information for causal impacts in the short to medium term, as the eldest participants from the child cohorts are currently 10‑11 years of age.

ATP is currently the only study in Australia that could be used to determine long term impacts of FF on health, economic and social outcomes.  Participants in this study are currently 25 years of age and would provide sufficient data to determine long term outcomes.  Indeed, preliminary works have been published which map the inter-relationships of a number of variables that interact over time and result in outcomes such as anxiety, depression, substance abuse, antisocial behaviour, academic achievement, risky driving and measures of social capital.  The main limitation of ATP is the length of time that measures of FF and parenting have been recorded.  These measures have only been taken since the participants were in their early teens, with no measures during infancy or early childhood.

2.2.6                   Proposed method of determining outcomes

Outcomes would be mapped from the ATP and LSAC databases using a similar methodology to that developed by Prior et al (2005) (see Figure 2‑6 and Figure 2‑7).  Outcomes would be mapped for each wave of the study so that the likelihood of one of the events of interest occurring could be established across different age groups.  For example, as shown in Figure 2‑6 the outcome being measured is depression (on the right of the diagram).  The mapping process would determine the relationships between family, peer, school, and other environments (on the left hand side of the diagram) that lead to this depression outcome.  These determining environmental factors build up over time so information on each individual from previous waves of the study would be used to determine outcomes in older age cohorts.

The ability to join information from LSAC and ATP is limited by slightly different methods of measuring FF.  There are, however, some consistencies, although one or two more waves of information would ideally be required to successfully bridge the gap in ages between the two studies.  Figure 2‑8 shows the approximate age of the study cohorts, as well as the information gap that would exist as a result of conducting this evaluation with the current data.  It is possible to utilise the current information to estimate outcomes in the gap period.  This would be done on a last observation carried forward basis (in other words the mapping for outcomes for children aged 10‑11 years would be applied to those aged 11 to 13 years.

Figure2 ‑ 8 :  Approximate age of study cohorts and bridging the current information gap

While the option of waiting for additional study waves to be completed to bridge the gap in information would provide a more accurate estimate for mapping purposes, completing this analysis with the current information using a last observation carried forward would not introduce vast inaccuracies into the estimates.  The last observation carried forward method could be tested using sensitivity analysis to inform the impact of these assumptions on the overall costing methodology. 

In addition to this there would be the opportunity to update the study with the remaining waves of information as they become available.  This process would allow the current set of data to be appropriately utilised, for the most accurate estimation of costs.

In order to complete the study in this way, a partnership with AIFS and in particular the ATP study group would be required.  This is the result of data access restrictions for the ATP study due to privacy concerns.  AIFS would be responsible for conducting the analysis of the ATP database and would have inputs behind the methodology used as well as which variables would most appropriately be controlled. 

A number of benefits would result from such a partnership.  These benefits stem from previous knowledge and work completed by AIFS in this area, as well as their extensive knowledge of both the ATP and LSAC databases.  These benefits would extensively enhance the quality of an evaluation through the cohesion of the datasets.

Risks associated with this partnership, might occur through the inclusion of a third party completing part of the analysis and the loss of sovereignty in relation to control of timelines that this would involve.  It is expected that the analysis by AIFS would be of a very high standard (given previous work in this area), however cross checking of analyses may require additional time to ensure consistency and accuracy of the mapping process.

A partnership with AIFS/ATP is a key element to the success of this project.  Alternatives are limited and either result in extensively longer timeframes or potentially greater restrictions of data access as well as uncertainty surrounding accuracy.  The alternative that would maximise the accuracy of this study would be to wait for an additional 15 years until further information was collected through LSAC and complete the study using one accessible data source.  Another alternative that would potentially allow a report to be completed in a reasonable timeframe would to search for similar international longitudinal studies.  This however, may not alleviate difficulties with database access and may in fact result in greater difficulties in ensure consistency between datasets, as well as being foreign rather than Australian.

2.3                        Cost components and considerations

The valuation exercise would incorporate a number of cost components related to the three broad areas of outcomes (health, productivity and social/criminality) previously discussed.  A number of these cost components will be attributable to more than one of the impacts discussed above (for example, health system expenditure can be attributable to both health specific outcomes and criminal behaviour).  These components are:

q  health system expenditure;

q  productivity and carer costs/gains;

q  criminality costs;

q  taxation revenue forgone and associated deadweight losses (DWLs);

q  other financial impacts; and

q  burden of disease (measured in disability adjusted life years or DALYs).

Cost components map to FF as shown in the diagram below (Figure 2‑9).

Figure 2 ‑ 9 :  Conceptual map for valuing costs of NFF, example only

In addition, this chapter also addresses measuring NPV and choosing discount rates. Each of the cost components is discussed in further detail in the following sub-sections.  First, however, there are some conceptual issues relating to the classification of costs that are briefly addressed.

q  Direct and indirect costs:  Although literature often distinguishes between direct and indirect costs, the usefulness of this distinction is dubious, as the specific costs included in each category vary between different studies, making comparisons of results somewhat difficult.

q  Real and transfer costs:  Real costs use up real resources, such as capital or labour, or reduce the economy’s overall capacity to produce goods and services.  Transfer payments involve payments from one economic agent to another that do not use up real resources.  For example, if a person loses their job, as well as the real production lost there is also less income taxation, where the latter is a transfer from an individual to the government.  This important economic distinction is crucial in avoiding double-counting (Laing and Bobic 2002, Laurence and Spalter-Roth 1996).

q  Financial and non-financial costs:  Financial costs encompass loss of goods and services that have a price in the market or that can be assigned an approximate price with some certainty.  ‘Non-financial’ costs include the loss of wellbeing of a person. Greater uncertainty tends to surround the valuation of non-financial costs, so results should be presented and interpreted cautiously.

q  Prevention and case costs:  Costs following from NFF are distinguished from costs directed towards promoting PFF.  The latter represent the cost of interventions such as those discussed in Chapter 3.

There are essentially two ways of estimating each element of cost for each group:

q  top-down: these data may provide the total costs of a program element (eg, health system); or

q  bottom-up: these data may provide estimates of the number of cases in the category (‘n’) and the average cost for that category; the product is the total cost (eg, the wage rate for lost earnings multiplied by the reduction in employment, and the number of people to whom this applies).

It is generally more desirable to use top-down national datasets in order to derive national cost estimates for large and well-studied social aspects, rather than extrapolate bottom-up data from smaller partial datasets.  Hence a top-down approach is generally applicable to community and government programs, as well as health system expenditures.  A bottom-up approach generally applies in other cases.  Whenever obtaining parameters required for implementing the bottom-up approach, statistical analysis of datasets and a literature review (focussing on Australian literature but sometimes supplemented by international material) are preferable. 

Uncertainty from data sources can be addressed by controlling for confounding factors or other problems where possible, and conducting sensitivity analysis on findings.

2.3.1                    Health system expenditure

Health system expenditure directed to treating conditions caused by NFF represents a diversion of resources that otherwise (in the absence of those health problems) could be used to produce other goods and services and hence optimise the social welfare function in the Australian economy.

Health system expenditure would be attributed to downstream health outcomes, in particular to the direct effects of resulting mental illnesses, substance abuse (drug and alcohol) as well as suicide and self-harm.  The Australian Institute of Health and Welfare (AIHW) produces health system expenditure costs that can be isolated to specific ICD-10 codes.  The health system expenditures cover all sources of funding within the health sector (federal government, state government, private health insurer and out-of-pocket expenses).  These expenditures also incorporate a number of health system components such as:

q  hospitalisation costs (inpatient and outpatient);

q  out of hospital medical services (unreferred attendances, imaging, pathology, other medical specialist services);

q  pharmaceuticals (prescription and over the counter); and

q  other health professionals.

The AIHW produce these expenditure estimates by ICD-10 code, split by age and gender.  These estimates of expenditure are relevant for the whole of the health sector (that is, they have not been siloed to one particular funding source).  With this in mind, these health system expenditures will need to be converted into a cost per case, using prevalence estimates of the specific condition linked to FF.  Evidence from the literature specifying the expected levels of downstream health impacts (such as anxiety, depression, drug abuse, alcohol abuse and suicide) will need to show evidence of causation (as opposed to correlation).  These aspects were previously discussed in Section 2.1 and Section 2.2.

Health system expenditures can also be attributed to the occurrence of criminal activity.  Mayhew (2003) estimate medical costs per crime for homicide offences ($7,600), assault offences ($200), sexual assault offences ($200) as well as robbery offences ($300).  These health system expenditures could be applied (using a specific health cost inflation rate) to predicted crime events in future years linked to FF.  This approach will be conservative as it does not include all sources of health system funding in its cost calculation.  This approach is used in the estimations of second generational crime associated with child abuse and neglect in Taylor et al (2008).

2.3.2                    Productivity and carer costs

Both the productivity losses and informal carer costs represent less than optimal participation and employment in the labour force.

q  Productivity losses would best be measured using a frictional approach in the short term and a human capital approach in the long term[7], given relatively low rates of unemployment in Australia over future decades[8].  Productivity losses for children experiencing NFF derive from all three sources – health impacts, productivity impacts (through the educational pathway) [9], and social/criminal impacts.  Multiplicative probabilistic adjustments could be used to ensure there is no double-counting of productivity impacts where more than one impact affects a person.

q  Informal carer costs are a result of health impacts.

2.3.2.1             Productivity losses

In the short term, most people with a specific health outcome (such as anxiety or depression) require time off work to attend medical appointments, receive treatment in hospital, or simply recover from their symptoms at home.  In the case of criminality, court appearances can also require short term absences.

In the long term, health impacts, educational disruption and crime/incarceration can lead to reduced labour force participation (including through death).  The reduction in the labour pool thus decreases the capacity of the economy to produce goods and services at any given level of unemployment.  An affected person may also choose to retire early and remain permanently out of the workforce (‘permanently disabled’), or return to work at a lower level of productivity through:

q  reduced hours of work;

q  decreased efficiency;

q  changed responsibilities; and/or

q  changed employment.

Factors that influence whether a person returns to work, and at what level of productivity, include the following (Main et al, 2005, Spelten et al, 2002):

q  the severity of physical and mental health impacts on functionality;

q  economic needs to survive and provide for others; and

q  attitudes to work and the workplace environment.

The length of time that a person is out of work can have a substantial impact on the attainment of experience and hence career development.

There are two components to productivity losses which are measured in this section:

q  the initial short-run disruption until production is restored to former levels (due to temporary absences); and

q  the loss of the labour resource (if there is permanent disability or fatality) over the longer term, which reduces the capacity of the economy to produce at any given level of unemployment.

Figure 2‑10 illustrates some potential variations in productivity losses, and whether they are ‘funded’ through sick leave, Sickness Allowance, the Disability Support Pension or lost/reduced earnings.

Figure2 ‑ 10 :  Productivity losses

Short term productivity costs

The time period persists until the patient returns to work, or is replaced, if they become unable to work.  In the meantime, employers often choose to make up lost production through overtime or employment of another worker that attracts a premium on the ordinary wage.  The overtime premium represents lost employer profits.  On the other hand, the overtime premium also indicates how much an employer is willing to pay to maintain the same level of production.  Thus, if overtime employment is not used, the overtime premium also represents lost employer profits due to lost production.  Thus, while productivity remains at the same level, the distribution of income between wages and profits changes[11]. 

According to traditional microeconomic theory (in particular the work of Gary Becker in the 1960s), people will work until they are indifferent between the marginal value of the income earned relative to the personal value of the leisure sacrificed.  However no-one else tends to value the individual's leisure similarly.  The typical approach to overcome this problem is to value leisure time at a discounted proportion of earnings which takes into account taxes that reduce the effective income from work and restrictions on the amount of time that can be used for work (for both biological and governmental regulation reasons).

Average employment rates and average weekly earnings (AWE) would be based on ABS data for all calculations of productivity losses (Table 2‑9).

Table 2 ‑ 9 :  Average e mployment r ate and AWE in the Australian p opulation

Source: ABS 6105.0, ABS 6310.0 (July 2008).

Long term productivity costs

Avenues through which NFF can lead to the long term reduction in the productive capacity of the labour force include long term absence from employment or reduction in hours of work, long term reduction in the productivity per hour worked, premature retirement[12] and premature mortality (i.e. some people may die before retirement age).

A full economic analysis of the effects of a disease on the economy would also examine the long-run situation where the lost productive capacity of the labour force (incurred via the worker or the employer) is passed on to society through adjustments in wages and prices.  A reduction in the supply of labour would increase wages, which would be passed on to consumers through price increases.  At the same time a decrease in the demand for goods and services would decrease prices, which would push down wages.  The overall impact on the economy depends on a complex array of elasticities.  However, this methodology would assume that, in the absence of the disease, people with an illness would participate in the labour force and obtain employment at the same rate as other Australians, and earn the same average weekly earnings.  The implicit and probable economic assumption is that the numbers of such people would not be of sufficient magnitude to substantially influence the overall clearing of the labour market.

The following methodology would be used to estimate lost long-run productivity costs.

q  The expected retirement age by the current age of the worker is calculated based on the participation rates at each age group.  Similar to life expectancy, the older the person, the less time it is expected that the person will remain in the workforce but the older they are when they do leave the workforce (see Table 2‑10).  Note that this methodology takes into account the probability that the person is working.

q  As the person ages, the annual income (based on average weekly earnings) is multiplied by the average employment rate at each age group while alive.  Income earned at each age is then summed to calculate the expected total income over a person’s lifetime (discounted back to present values).

Table2 ‑ 10 :  Expected r etirement a ge and r emaining l ifetime e arnings ( $ 2005)

Sources: ABS 6105.0, ABS 6310.0 (Indexed to $2005).

Consequently:

q  for permanent disability: the expected remaining lifetime earnings are reduced by the percentage reduction in employment during the period the individual has a lower level of employment; and

q  for premature death: the entire expected remaining lifetime earnings for the individual are lost.  The productivity costs of premature mortality are allocated to the year that the person died.

If an employed person stops working or prematurely dies from their illness, there are also staff turnover costs borne by the employer.  These turnover costs are estimated to be equal to 26 weeks salary of the incumbent worker (Access Economics, 2004a).  However, this cost is merely ‘brought forward’ a number of years because there would be some normal turnover of persons – approximately 15% per annum (which implies that people change jobs, on average, approximately once every 6.7 years (Access Economics, 2004b).

2.3.2.2             Informal carer productivity losses

Carers are people who provide informal care to others in need of assistance or support.  For example, carers may take time off work to accompany people with an illness to medical appointments, stay with them in hospital, or care for them at home.  Carers may also take time off work to undertake many of the unpaid work the person with the illness used to do – such as provide childcare, do the housework, yard work, shopping and so on.

Informal care is distinguished from services provided by people employed in the health and community sectors (formal care) because the care is generally provided free of charge to the recipient and is not regulated by the government.  Most informal carers are family or friends of the person receiving care.

While informal care is provided free of charge, it is not free in an economic sense, as time spent caring is time that cannot be directed to other activities such as paid work, unpaid work (such as housework or yard work) or leisure.  As such, informal care is a use of economic resources.

There are three potential methodologies that can be used to place a dollar value on the level of informal care:

q  opportunity cost method – measures the formal sector productivity losses associated with caring, as time devoted to caring responsibilities is time which cannot be spent in the paid workforce;

q  self-valuation method – measures that carers themselves feel they should be paid; and

q  replacement cost method – measures the cost of ‘buying’ an equivalent amount of care from the formal sector, if the informal care were not supplied.

The self-valuation method is not commonly used, and there are few reliable Australian studies of the amount Australian carers feel they should be compensated. 

Estimates of the value of informal care are very sensitive to the estimation methodology used.  In general, the opportunity cost method is used as data about the age and sex of the carers are available from SDAC and the aim of this report would be to estimate the current impacts on people with a resulting health outcome (from NFF) and their families, rather than a ‘what if’ scenario if this care were instead provided by formal care providers.

2.3.3                    Criminal justice system costs

Greater rates of participation in crime and delinquency due to NFF lead to greater expenditures within the criminal justice system.  Criminal justice system expenditures would utilise a top-down approach for cost estimation.  Total governments expenditures for the justice system are provided in the report for government services, produced by the Steering Committee for the Review of Government Service Provision (SCRGSP).  Expenditures are reported by police services, court administration and corrective services.

These government expenditures encapsulate total recurrent costs for both federal as well as state and territory programs.  To attribute these expenditures to the impacts of NFF, these recurrent costs need to be converted to per case expenditures, that is, divided by the total number of people accessing these services.  For example, the ABS produces statistics on court appearances (Cat. no. 4513.0), which can be used to derive an average cost per appearance.  Evidence from the literature showing the causal effects of a NFF environment (discussed previously in Section 2.1 and Section 2.2) would be used to estimate an attributable cost of these government programs and services.

2.3.4                    Taxation revenue forgone and associated DWLs

Where there are workplace productivity losses, there is also associated taxation revenue forgone.  While taxation revenue is a transfer payment rather than a real economic cost, it is important to estimate these losses for two reasons.  First, they affect the distribution of the costs (i.e. whether they are borne by the individual, Government or society).  Second, taxation has associated real DWLs that arise due to the administration of the taxation system and the distortionary impacts on production and consumption choices of raising taxation.

DWLs are not only related to taxation revenue forgone, but also to:

q  other transfer payments, notably the additional induced social welfare payments required to be paid; and

q  the value of government services provided (including the Government-funded component of health system costs).

If the act of taxation did not create distortions and inefficiencies in the economy, then transfers could be made without a net cost to society.  However, through these distortions, taxation does impose DWLs which represent lost goods and services from NFF.

DWL is the loss of consumer and producer surplus, as a result of the imposition of a distortion to the equilibrium (society preferred) level of output and prices.  Taxes alter the price and quantity of goods sold compared to what they would be if the market were not distorted, and thus lead to some diminution in the value of trade between buyers and sellers that would otherwise be enjoyed.  The principal mechanism by which a DWL occurs is the price induced reduction in output, removing potential trades that would benefit both buyers and sellers.  In a practical sense, this distortion reveals itself as a loss of efficiency in the economy, which means that raising $100 of revenue requires consumers and producers to give up more than $100 of value (Figure 2‑11).

Figure2 ‑ 11 :  DWL of t axation

The rate of DWL recommended is 27.5 cents per $1 of tax revenue raised plus 1.25 cents per $1 of tax revenue raised for Australian Taxation Office administration, based on Productivity Commission (2003) in turn derived from Lattimore (1997) i.e. 28.75% overall. 

2.3.5                    Other indirect financial impacts

Educational assistance

For a variety of reasons, children coming from a NFF environment may require more assistance at school.  Lower levels of literacy and academic performance have been associated with children from these family environments resulting in potentially higher utilisation rates of government funded education programs.  Educational assistance due to NFF that is provided through Government programs would be most suited to a top-down approach for cost estimation.  Total governments expenditures for early childhood education and training are provided in the report for government services, produced by SCRGSP.  Expenditures are reported by children’s services, school education and vocational education and training.  These would be converted to costs per case and attributed to NFF on the same basis as for criminal justice costs.

Welfare transfers

Lower academic performances and Year 12 completion rates may, in turn, lead to lower rates of participation in the labour force and thereby a greater reliance on welfare payments such as NewStart.  Welfare and support payments, such as NewStart, Sickness Allowance and DSP will require a bottom-up approach for estimation.  These costs are defined and treated as transfer costs, as they are a transfer of wealth from one person to another and therefore have a net economic cost of zero.  Although, as discussed in Section 2.3.4, these transfer payments occur with friction and result in distortions in the economy imposing further DWLs. 

Centrelink collects information on the amount, number, and reason that people receive welfare payments.  The most appropriate method in determining this costing component is to use data from Centrelink that provides welfare and support payment information by disease or injury group.  It should be noted that these are an underestimate because:

q  people aged over 65 years would be eligible for the aged pension instead[13];

q  Sickness Allowance and Disability Support Pension would not include people who did not indicate a particular illness (caused by NFF) as a reason for receiving the payment; and

q  some individuals after recovering from an illness (caused by NFF) may still experience prolonged absence from the labour force or premature retirement and continue to receive welfare payments such as NewStart.

The proportion of welfare payments caused by NFF can be estimated from the Centrelink data and the NFF attributable fractions estimated from causal linkages.

Aids and home modifications

Although the cost of some health aids and appliances are captured within health costs, other ‘non‑health’ equipment (such as movement devices) or home modifications may not be.  These cost components could be estimated as one-off expenses, with a market rate used for valuation, and may include:

¬  movement devices :  wheelchairs (manual and electric), crutches, scooters, walking frames and sticks;

¬  home modifications :  ramps, hoists, mechanical lifters, specific room modifications, hand held showers, rails and grips;

¬  medical aids :  ventilators, tube feeding equipment, respirator masks and accessories; and

¬  personal care :  shower chairs or commodes, continence aids, pressure care cushions, beds and mattresses.

Travel and accommodation

Travel and accommodation costs to obtain services required as a result of NFF may also be substantial, especially for people living in rural and remote areas.  Measurement of these costs elements can be based on average unit cost measures multiplied by the number of people incurring such costs.  It is doubtful that there will be enough relevant information to estimate these amounts.

Other services for people experiencing the impacts of NFF

Government and non-government service programs aimed to assist with the consequences of NFF (e.g. child protection services intervening in abuse cases) would be determined through an investigation and liaison with the relevant departments and would include other services provided by government not captured in the cost elements itemised in the previous sections.  While most of these service costs will be borne by state and territory governments, some are co-funded federally or require co-payments, as well as services provided by the not-for-profit sector.

Bring forward of funerals

The ‘additional’ cost of funerals borne by family and friends of patients is based on the likelihood of death due to illness in the time that a person has it.  However, some patients (particularly older patients) would have died during this time anyway, from other causes.  Eventually everyone must die and this incur funeral expenses – so the true cost is the cost brought forward (adjusted for the likelihood of dying anyway in a given year).  The BTRE (2002) calculated a weighted average cost of a funeral across all States and Territories, to estimate an Australian total average cost of $3,200 per person for 1996.  This could be inflated to current prices and applied to the number of premature deaths due to NFF.

2.3.6                    Burden of disease

To those experiencing NFF, impacts such as mental anguish, loss of leisure, physical pain and disability are often as (or more) important than the tangible financial costs discussed so far.

There are a number of methods that can be applied to determine measures of the quality of life (QoL) or wellbeing that result from aspects of FF.  One common measure of lost wellbeing is the ‘burden of disease’ associated with the health outcomes resulting from NFF.  This measure is limited in that it only measures illnesses and injuries as QoL and does not include aspects of overall family happiness and wellbeing that are not associated with health outcomes.  However, no data were identified on the latter in this scoping exercise.

2.3.6.1                       Burden of disease — DALYs and QALYs

In the last decade, a non-financial approach to valuing human life has been derived, where loss of wellbeing and premature mortality – called the ‘burden of disease and injury’ – are measured in DALYs.  This approach was developed by the World Health Organization (WHO), the World Bank and Harvard University for a study that provided a comprehensive assessment of mortality and disability from diseases, injuries and risk factors in 1990, projected to 2020 (Murray and Lopez, 1996). Methods and data sources are detailed further in Murray et al (2001) and the WHO continues to revisit the estimates for later years.

A DALY of 0 represents a year of perfect health, while a DALY of 1 represents death.  Other health states are attributed values between 0 and 1 as assessed by experts on the basis of literature and other evidence of the quality of life in relative health states.  For example, the disability weight of 0.18 for a broken wrist can be interpreted as losing 18% of a person’s quality of life relative to perfect health, because of the inflicted injury.  Total DALYs lost from a condition are the sum of the mortality and morbidity components – the Year(s) of Life Lost due to premature death (YLLs) and the Year(s) of healthy life Lost due to Disability (YLDs). 

The DALY approach has been successful in avoiding the subjectivity of individual valuation and is capable of overcoming the problem of comparability between individuals and between nations, although some nations have subsequently adopted variations in weighting systems, for example age-weighting for older people.  This report treats the value of a life year as equal throughout the lifespan.

As these approaches are not financial, they are not directly comparable with most other cost and benefit measures.  In public policy making, it is often desirable to apply a monetary conversion to ascertain the cost of an injury, disease or fatality or the value of a preventive health intervention, for example, in CBA.  Such financial conversions tend to utilise ‘willingness to pay’ or risk-based labour market studies as described in the next section.

2.3.6.2                       Willingness to pay and the value of a statistical life year

The burden of disease as measured in DALYs can be converted into a dollar figure using an estimate of the Value of a ‘Statistical’ Life (VSL).  As the name suggests, the VSL is an estimate of the value society places on an anonymous life.  Since Schelling’s (1968) discussion of the economics of life saving, the economic literature has focused on willingness to pay (WTP) – or, conversely, willingness to accept – measures of mortality and morbidity, in order to develop estimates of the VSL.

Estimates may be derived from observing people’s choices in situations where they rank or trade off various states of wellbeing (loss or gain) either against each other or for dollar amounts e.g. stated choice models of people’s WTP for interventions that enhance health or willingness to accept poorer health outcomes or the risk of such states.  Alternatively, risk studies use evidence of market trade-offs between risk and money, including numerous labour market and other studies (such as installing smoke detectors, wearing seatbelts or bike helmets and so on).

The extensive literature in this field mostly uses econometric analysis to value mortality risk and the ‘hedonic wage’ by estimating compensating differentials for on-the-job risk exposure in labour markets; in other words, determining what dollar amount would be accepted by an individual to induce him/her to increase the probability of death or morbidity by a particular percentage.  Viscusi and Aldy (2002), in a summary of mortality studies, found the VSL ranged between US$4 million and US$9 million with a median of US$7 million (in year 2000 US dollars), similar but marginally higher than the VSL derived from studies of US product and housing markets.  They also reviewed a parallel literature on the implicit value of the risk of non-fatal injuries.

Weaknesses in the WTP approach, as with human capital approaches to valuing life and wellbeing, are that there can be substantial variation between individuals.  Extraneous influences in labour markets such as imperfect information, income/wealth or power asymmetries can cause difficulty in correctly perceiving the risk or in negotiating an acceptably higher wage in wage-risk trade off studies, for example.

As DALYs are enumerated in years of life rather than in whole lives it is necessary to calculate the Value of a ‘Statistical’ Life Year (VSLY) based on the VSL.  This is done using the formula: [14]

Clearly there is a need to know n (the years of remaining life), and to determine an appropriate value for r (the discount rate). There is a substantial body of literature, which often provides conflicting advice, on the appropriate mechanism by which costs should be discounted over time, properly taking into account risks, inflation, positive time preference and expected productivity gains.  In reviewing the literature, Access Economics (2008) found the most common rate used to discount healthy life was 3% and this had a strong rationale (see next section 2.3.7).  Access Economics (2008) recommended an average VSL of $6.0 million in 2006 Australian dollars ($3.7 million to $8.1 million).

q  This equates to an average VSLY in 2006 of $252,014 ($155,409 to $340,219), using a discount rate of 3% over an estimated 40 years remaining life expectancy. [15]

q  Inflating the 2006 VSLY value to 2008 dollars by multiplying it by two years of inflation (2.9% in each year, from the Access Economics Macroeconomic model) results in a base case of $266,843 with lower and upper bounds of $164,553 and $360,238.

q  However, from this gross value, Access Economics deducts all costs borne by the individual, reflecting the source study VSL estimates, to avoid double counting.  This provides a different net VSLY for different conditions (and for different age-gender groups).

Since Access Economics (2008) was published, the Department of Finance and Deregulation have also provided an estimate of the VSLY, which appears to represent a fixed estimate of the net VSLY.  This estimate was $151,000[16] in 2006, which inflates to $157,795 in 2008 dollars.  This is very similar to the average net VSLY estimated using the Access Economics (2008) meta-analysis, and is suggested for calculations in modelling here.

2.3.7                    Net present value and the discount rate

Where future costs are ascribed to the year 2007 throughout the report the formula for calculating the NPV of those cost streams is provided below.

NPV = ΣC_i/(1+r)^ⁱ where i=0,1,2….n              where

C_i = cost in year i, n = years that costs are incurred and r = discount rate.

Choosing an appropriate discount rate is a subject of some debate, as it varies depending on what type of future income or cost stream is being considered.  The discount rate needs to appropriately take into account risks, inflation and positive time preference.

q  Risk and positive time preference : The minimum option that one can adopt in discounting future expected healthy life streams and other costs is to set future values on the basis of a risk free assessment about the future i.e. assume the future flows are similar to the certain flows attaching to a long term Government bond. From recent history, the long term nominal bond rate has averaged 5.8% per annum (Figure 2‑12). If there were no positive time preference, people would be indifferent between having something now or a long way off in the future, which applies to all goods and services.

Figure2 ‑ 12 :  Ten-year government bond yields (daily)

Source: Bloomberg, Reserve Bank of Australia.

q  Inflation :  The Reserve Bank has a clear mandate to pursue a monetary policy that delivers 2% to 3% inflation over the course of the economic cycle.  This is a realistic longer run goal and an inflation rate in this range (2.8%) is used in arriving at the discount rate for healthy life below.  It is important to allow for inflation in order to derive a real rather than nominal rate.

3.          Efficiency of current Australian and International interventions

Policy interventions targeting FF have historically been difficult to measure and implement, primarily as a result of the large number of interacting and confounding factors involved in raising functional children from functional families (Carneiro and Heckman, 2003).  Interventions targeting FF aim to reduce risk factors that are the product of a complex web of prenatal, social, cultural, demographic, family, neighbourhood, economic and political factors (AIHW, 2005).

Interventions, whether they be through the provision of social services or long term social marketing campaigns, fall into one of three categories (Donovan et al, 2007) depending on where they are designed to take effect (Figure 3‑1).  Primary prevention is largely aimed at influencing public opinion and attitudes, with the aim of reducing the overall distal risk factors for NFF within society as a whole.  Secondary prevention strategies are aimed at providing services to those who are most at risk of NFF, focusing on proximal risk factors.  Tertiary prevention is aimed at reducing risks of recurrence where an adverse NFF outcome has occurred (such as child abuse or neglect).

Figure3 ‑ 1 :  Prevention continuum

Source:  Donovan et al (2007).

There are a number of points in the family life cycle where a FF intervention may be implemented, from the pre-birth stage through to the child’s adolescent years, with some interventions providing a specific focus on risky transition periods and others targeting overall family and society environment changes.  For example, Carneiro and Heckman (2003) showed that interventions made early in the lifecycle provided the most optimal investments in terms of human capital outcomes (Figure 2‑3).

Interventions at the FF level can be thought of as targeting adverse outcomes (health, productivity or criminality) through generally distal risk factors as opposed to the more common proximal risk factors.  This is in line with a recent shift in paradigm in the examination of causal relationships of diseases, illnesses and injury towards the associated distal risk factors.  In brief:

q  proximal risk factors are those that are close to the individual or close in time to the origins of NFF, typically providing the strongest correlations; and

q  distal risk factors are those that are typically distant from the individual, for example social or cultural factors or factors occurring earlier in time before NFF is experienced.

A 2002 report by the AIFS stated that:

‘…it is important to study [the] historical and social context and how diversity or cultural norms influence pathways.  Mathematical models concentrating on proximal risk factors actually preclude causal pathway analyses that involve more distal risk factors.  This is because in a multivariate model proximal risk factors are usually so powerful that the more distal risk factors do not appear significant (Stanley et al, 2000).  By ‘controlling for confounding’ we not only ignore the earlier part of the causal path, we also often ignore the causal sequences and interactions which may be so important.’ Sanson (2002)

This chapter explores the effectiveness of a broad range of interventions targeting FF, sourced from the literature.  These interventions were chosen, in close consultation with the Expert Reference Group (ERG) and FaHCSIA, as a representative sample selected on the basis of achieving a balance of:

q  Australian Government, state/territory and international;

q  focus on parent, child, family and community/environment;

q  different life cycle stages (pre-birth, infancy, later); and

q  existing evidence compared to future potential evidence.

General families

q  Family assistance programs and income support payments.

q  Family Relationship Services (FRS) as well as Children and Parenting programs, focusing on a couple of individual services e.g. Family Education and Skills Training around parenting, Adolescent Mediation for young people.

Young children

q  Stronger families and communities strategy 2004-2009

o  Communities for children

o  Invest and grow (a mostly early years’ intervention, targeted at socially disadvantaged areas)

q  positive parenting programs (Triple-P).

q  early education programs

q  Peel Child Health Study (Our Children, Our Family, Our Place): looks at biological factors interacting with community (school, daycare, recreation) and how it influences health and development.

q  REACh, a program intended to improve the capacity of families and caregivers to respond appropriately to children's needs for care, development and safety through timely access to community resources that can support them in their parenting role.

Youth

q  Reconnect uses community-based early intervention services to assist young people aged 12 to 18 years who are homeless, or at risk of homelessness, and their families. Reconnect assists young people stabilise their living situation and improve their level of engagement with family, work, education, training and their local community.

q  YouthLinx - is prevention and early intervention program that supports young people aged 11 to 16 and their families build self reliance, strengthen their family relationships and encourage community involvement.

q  TILA - helps young people exiting formal/informal care to transition to independent living through enhanced access to accommodation, employment/education and the development of life skills/knowledge.

Other

q  SURE start (from the UK).

q  Parents Under Pressure (PUPs) – in relation to parental substance abuse.

Each of these interventions has been designed to intrinsically affect a number of domains of FF from Table 1‑1.  In particular the first five domains are typically impacted by the majority of interventions with significant overlap in each domain.  The last domain (social connectedness) is typically directly affected by broad interventions such as those under the ‘Stronger families and communities strategy 2004-2009’. 

In addition, different aspects of each intervention are implemented at different points along the prevention continuum.  For example, the Family Relationship Services is a broad program of service based interventions that have both elements of secondary and tertiary interventions, depending on the service provided.

3.1                        Family assistance programs and income support payments

Program services and target groups

Family assistance programs are offered by the Australian government and provide payments that assist with the costs of parenting.  Income support payments under these programs are listed below. [17]

q  Family Tax Benefit A :  helps families with the cost of raising children.

q  Family Tax Benefit B :  gives extra assistance to families with one main income including sole parents.

q  Child Care Benefit :  helps families with the cost of child care.

q  Child Care Tax Rebate :  additional help for working families with the cost of child care.

q  Baby Bonus :  helps with the extra costs of a new child (including adopted and stillborn children).

q  Maternity Immunisation Allowance :  is a non-income tested payment that is usually payable as two separate amounts for children who are fully immunised between the ages of 18-24 months and between 4 and 5 years, or who have an approved exemption from immunisation.

q  Large Family Supplement :  for people receiving Family Tax Benefit for three or more children.  An amount is added to the Family Tax Benefit for each child after the second.

q  Multiple Birth Allowance :  for families with three or more children born at the same time, an amount is added to the Family Tax Benefit until the children turn 16 years of age, or if at least three of the children are in full time study, until the end of the calendar year in which they turn 18 years of age.

q  Rent Assistance :  additional help for those eligible and receiving more than the base rate of Family Tax Benefit Part A while paying rent to private landlords.

q  Health Care Cards :  issued to those receiving the maximum rate of Family Tax Benefit Part A.

q  Double Orphan Pension :  provision of additional assistance to raise children that are orphans.

q  Jobs, education and training child care fee assistance :  as part of an activity agreement, this payment contributes to the costs of approved child care for eligible parents undertaking activities such as work, job searching, training, study or rehabilitation, to enter or re-enter the workforce. 

Income strain on families and households are well known to have negative effects on families.  Ermisch (2008) showed that parental behaviours improved as the household income increased (Figure 3–4).  The mechanism by which this outcome is achieved is unclear.  For example, it maybe through the alleviation of financial pressures at the margin that led to an improvement in family conditions, alternatively it might be related to people with higher levels of education earning more money and that are more informed about parenting practices.

Research on efficacy

Measures of program performance are linked into high level KPIs, with little measurement of the health and social outcomes that are associated with these interventions.  For example, the Family Tax Benefit performance indicators (Table 3‑1 and Table 3‑2) are linked to their expected and actual expenditure amounts.  Actual payments for Family Tax Benefit were under the estimated amount for Part A while expenditure on Part B was over its expected amount.

Table3 ‑ 1 :  Family Tax Benefit Part A

Source:  FaHCSIA Annual Report 2007-08

However, Family Tax Benefit Part B (Table 3‑2) has an additional measure of effectiveness –the proportion of families entitled to family Tax Benefit Part B that take up the benefit over a three year period.

Table 3 ‑ 2 :  Family Tax Benefit Part B

Source:  FaHCSIA Annual Report 2007-08

Similar KPI measurements are used by FaHCSIA for the ‘Double Orphan Pension’ (Table 3‑3), the ‘Maternity Immunisation Allowance’ (Table 3‑4) as well as the ‘Baby Bonus’ (Table 3‑5).  Performance indicators measure the expected and actual payments made, while quality assurance measures indicate the number of children in respect of whom the payments and benefits are made, as well as the number of families that receive the payment.

Table3 ‑ 3 :  Double Orphan Pension

Source:  FaHCSIA Annual Report 2007-08

Table 3 ‑ 4 :  Maternity Immunisation Allowance

Source:  FaHCSIA Annual Report 2007-08

Table3 ‑ 5 :  Baby bonus

Source:  FaHCSIA Annual Report 2007-08

3.2                        Family relationship services

Program services and target groups

The Family Relationship Services Program (FRSP) is a national program which is delivered by over 100 non-Government organisations across Australia providing a variety of services.  The program is funded by FaHCSIA and the Australian Government Attorney-General’s Department.

The purpose of the FRSP (Young, 2004) is to contribute to the development of Australia in terms of:

q  children, young people and adults being enable to develop and sustain safe, supportive and nurturing family relationships; and

q  minimising the emotional, social and economic costs associated with disruption to family relationships.

The program aims to reduce NFF, as per its overall intended benefits (Young, 2004) of:

q  enhancing people’s capacity for achieving and sustaining valued family relationships;

q  enhancing people’s relationships; and

q  improving people’s management of relationship difficulties. 

The broad nature of the goals and aims of the program make it apparent that the focus of the FRSP is not any particular target group of the population but rather on the Australian population in general.  Additionally, the FRSP intends to be non-discriminatory in its focus, supporting youth and adults, with equity of access in terms of income, location, age and culture.  Sub-programs within the FRSP include the following.

q  Family Counselling helps couples and families to manage relationship issues arising from relationship changes, separation or divorce.

q  Adolescent Mediation and Family Therapy services aim to prevent family breakdown or youth homelessness by working directly with young people and their families or caregivers to resolve conflict and improve relationships.

q  Family Relationship Services for Families in Special Circumstances.  Currently funded under this program is the Family Support Drought Response Team initiative, designed specifically for individuals and families, including small business owners, who are affected by the drought.

q  Family Relationships Education and Skills Training services assist couples and families, including those with children, to develop skills to foster positive, stable relationships with their partner or family.

q  Family Relationship Services for Carers (FRSC) is available to family members who are considering the future care of their family member with disability.

q  Family Relationship Services for Humanitarian Entrants are designed specifically for families who have entered Australia under the Humanitarian Entry Program.  Services are open to adults, young people, parents and their children.

q  Men and Family Relationship Services help men to develop and maintain strong family relationships, or deal with conflict or separation.

q  Mensline Australia is a national phone line dedicated to helping men and their families to deal with their family relationship issues.

q  Specialised Family Violence Services support individuals, couples, families, young people and children who have experienced or witnessed family violence, including helping those who use violent behaviour to change.

q  Family Relationship Centres are a first port of call for families.  Through information and referral, the centres help families strengthen relationships and deal with relationship difficulties.

q  Family Relationship Advice Line is a national telephone service providing free information, advice and referral to assist people affected by family relationship or separation issues.

q  Family Relationships Online provides individuals and families with information about family relationship issues, ranging from building better relationships to dispute resolution.

q  Children’s Contact Services provide a neutral venue for safe changeovers when families separate and the transfer of children from one parent to another is a problem.

q  The Parenting Orders Program assists separating families in high conflict over parenting arrangements.

q  Family Dispute Resolution is the legal name for services such as mediation and conciliation that help people affected by separation and divorce to sort out their disputes with each other.

q  Regional Family Dispute Resolution is designed to meet the particular needs of regional communities, providing a range of services to help separating families resolve disputes and reach agreement on parenting arrangements, finances and property.

q  Post Separation Cooperative Parenting services assist separated parents in conflict to work cooperatively over parenting arrangements.

q  Supporting Children After Separation services help children from separating families to deal with issues arising from the breakdown in their parents’ relationship and to participate in decisions that impact on them.

Research on efficacy

A number of separate evaluations of particular FRSP sub-programs have been carried out in the past, however a recent review (Young, 2004) attempted to assess the FRSP as a whole.  It intended to gather information about the effectiveness of the overall FRSP program and its contribution to strong family relationships.  The findings of specific studies and the broader review are summarised below.

Review of FRSP 2004

Research findings on client perspectives on the FRSP program were highlighted by the Review on FRSP (2004).  It was stated that client research was conducted through 10 discussion groups and 46 one-to-one interviews with existing users across 11 sub-programs and non-users/potential clients including parents, couples and people involved in the court system.  Interviews were conducted in the locations of Sydney, Brisbane and Perth, as well as Bunbury, Ballarat and Cairns.

This client research found (Young, 2004):

q  Overall user reaction to FRSP services was overwhelmingly positive, with users expressing a high level of satisfaction with these services and being able to identify positive outcomes resulting from them.

q  Users reported that services accessed had assisted them in achieving:

¬  More effective communication with partners, ex-partners and children.

¬  More effective family relationships and a decrease in family tensions.

¬  Better approaches to repairing and rebuilding relationships.

¬  Agreements on custody, child support and other post-separation decisions which were difficult before intervention.

¬  Calm and safe visitations from children, from the non-residential parent.

¬  Emotional support and encouragement to improve.

¬  Increased self esteem.

¬  More confidence in capacity to deal effectively with other family/relationship issues.

¬  Better communication with a wide range of people and better understanding of children’s needs. 

Study on Family Dispute Resolution

A recent study funded by the Australian Government Attorney General’s Department (McIntosh and Long, 2006) compared outcomes over one year for two groups of separated parents who attended mediation over parenting disputes.  Parents engaged in either a Child Focused intervention of Child Inclusive in one of three Relationships Australia Services in Canberra, Melbourne and Adelaide.  There were 275 parents who took part in the study, reporting on 364 children, and 194 of these children (aged 5-16 years) also directly participated in the research. 

The Child Focused Intervention placed a priority on the psychological and relational elements of parents’ separation and formulating parenting arrangements best supporting the developmental needs of children.  Children were not seen for the purposes of mediation.  The Child Inclusive intervention had the same approach and aim but also involved an assessment of children’s experiences of the separation and their relationships with parents. 

It was reported that entry into mediation was a point of high risk of the groups/families, with both parents reporting high to very high current acrimony with their former or soon-to-be former partners.  Children also reported high rates of conflict between parents, with one third aged 5-16 displaying the clinical range of psychological symptoms (McIntosh and Long, 2006). 

The outcomes that were found to be common to both types of interventions were (McIntosh and Long, 2006):

q  Significant and enduring reduction in the levels of conflict since mediation.

q  The majority of parents reported improved management and dispute resolution of initial disputes that had brought them into mediation.

q  Across all age groups, children in both interventions perceived less frequent and intense conflict between their parents, and a significant lowering of their own distress over parental tensions.

It was reported that no isolated effects were evident for the Child-Focused intervention at three or twelve month follow-ups.  The Child Inclusive intervention had a number of associated effects not evident in the other treatment group, which were reported to be strongest for fathers and children.  It is reported that one-year post intervention showed significantly better outcomes in terms of (McIntosh and Long, 2006):

q  Lower acrimony in fathers to their former spouses.

q  Improvement in the parental alliance for fathers.

q  Increased contentment in children with care and contact arrangements and less inclination to change these.

q  Greater father satisfaction with care and contact arrangements for their children.

q  Greater stability of care and contact patterns over the year. 

Displayed in Table 3‑6, Table 3‑7 and Table 3‑8 are the means and standard deviations (SDs) of the children and parent measures at different points in time.

Table 3 ‑ 6 :  Family dispute resolution, means and SDs of parent measures

Source: McIntosh and Long (2006).

Table 3 ‑ 7 :  Family dispute resolution, other parent measures

Source: McIntosh and Long (2006).

Table3 ‑ 8 :  Family dispute resolution, means and SDs of children measures

Source: McIntosh and Long (2006).

Evaluation of Contact Orders Program

The Contact Orders Program commenced as a pilot in 1999 to help the most highly conflicted separated families manage their child contact arrangements and focus on their children’s needs.  The Program assists families having difficulty with enforcement of child contact orders, in particular where these have been breached.  The Program was evaluated by the Sydney Children’s Hospital from July 2000 to April 2002 based on information from evaluators’ visits to sites and data collection (Australian Government Attorney General’s Department, 2003).  The aim of the evaluation was to determine the success of the program in helping separated or divorced families in managing relationship difficulties, to ensure compliance with contact arrangements and ensure children can maintain a healthy relationship with the non-resident parent. 

It was reported that before the program, the parents involved had communication difficulties with each other in terms of discussing their children.  For example, 79% were reported to seldom/never discuss major decisions in their children’s lives with the other parent.  Of parents who had communication, 69% reported an underlying atmosphere of hostility and anger.  Conversation was reported to be stressful and tense in 74% of cases.  Of parents required to pay child support, only 54% were paying regularly. 

The Program used a variety of child-focused interventions to respond effectively and flexibly to families’ needs including:

q  group work;

q  education;

q  counselling;

q  mediation;

q  children’s contact services; and

q  overall case management. 

The evaluation had the following findings (Australian Government Attorney General’s Department, 2003):

q  Of all clients, 88% reported achieving a positive outcome from program participation, ranging from ‘some progress’ to ‘fully resolved’.  The statistical sample comprised clients of the Contact Orders program who completed surveys, with the sample sizes ranging from 403 to 643. 

q  Practitioners reported an even higher rate of client benefits.

q  Clients reported benefits in terms of learning about communication techniques, learning about how their children experience situations, improving dispute resolutions skills and feelings of less isolation. 

3.3                        Stronger Families and Communities Strategy 2004-2009

The Government’s Stronger Families and Communities Strategy 2004-2009 focused on early childhood issues (Soriano et al, 2008) with the aims of:

q  helping families and communities build better futures for children;

q  building family and community capacity;

q  supporting family relationships and communities; and

q  improving the ability of communities to help themselves.

The Communities for Children and Invest to Grow programs are initiatives under the strategy. These will be discussed specifically below.

3.3.1                   Communities for Children

Program services and target groups

The Communities for Children (CfC) program is an Australian Government-funded program under the Stronger Families and Communities Strategy. [18] The program aims to assist children aged 0-5 years to have the best possible start in life and influence the contexts in which they live, their family and community and broader social/economic/cultural environments.  The program was implemented as part of an initiative to improve the coordination of public, private and community activity. 

Members of the program include parents, community leaders, representatives from business, research, academic and non-government sectors and state/local governments. The program aims to provide a national framework for sustainability and capacity building, allowing for individual approaches at the local level, recognising the differing needs of families and communities.  The hybrid model of the program is similar to Sure Start in the UK.

Under the CfC initiative, non-government organisations are funded as Facilitating Partners in 45 community sites around Australia to develop and implement a strategic and sustainable, whole of community approach to early childhood development in consultation with local stakeholders.  Examples of activities that have been implemented at sites are: ^[19]

q  home visitation programs;

q  early learning and literacy programs;

q  early development of social and communication skills;

q  parenting and family support programs;

q  child nutrition programs; and,

q  community events to celebrate the importance of children, families and early years

Research on efficacy

Dubbo, Narromine and Wellington area study

Research was undertaken in on the impact of CfC in the Dubbo, Narromine and Wellington areas (Slack-Smith, 2008).  This research was based on a survey of all CfC committee members.  The survey combined a rating scale and a request for examples/comments.  In relation to each CfC Activity and Facilitating Partner activities, survey participants were asked to rate responses to the questions:  ‘In your view, is the activity making a change?  And ‘Who for?’  Under each rating request, members were asked to give an example and comment on observed changes and benefits.  In terms of ratings, when asked to rate whether an activity was making a change, there was (Slack-Smith, 2008):

q  66 ratings of ‘Don’t know’

q  0 ratings of ‘Negative change’

q  10 ratings of ‘Little change’

q  121 ratings of ‘Positive change’

q  100 ratings of ‘Very positive change’

When asked who changes were affecting:

q  107 stated themselves/their organisation

q  132 stated other service providers

q  119 stated children

q  131 stated families

q  116 stated the wider community

Overall, it was stated that committee members who responded were pleased with the progress of the program and reported many changes/benefits for the community, with no negative change reported.

Breaking Cycles by building neighbourhood hubs

This CfC project aimed to develop community hubs where children’s services could work together to achieve better outcomes for young children and their families in Hume in the Broadmeadows area in Melbourne.  The project commenced in July 2006.

The project targeted families with young children in the 0-5 age group, particularly those from culturally and linguistically diverse backgrounds.  This area was targeted due to high levels of disadvantage and poverty, with limited access to early childhood education programs and affordable childcare.

The project was backed by research indicating that (references were cited by the Communities and Families Clearinghouse Australia, Australian Institute of Family Studies, AIFS 2009a):

q  Families generally receive more comprehensive support from integrated service due to the ‘one-stop-shop’ nature of the centres (Bruner, 1991).

q  Children are prevented from falling through the cracks of service provision especially if they experience multiple disadvantage and social exclusion (Bruner, 1991).

q  Environmental risks in a given community that affect children, such as family violence and poverty, reduce when there are sustained interventions by integrated services (Bruner, 1991).

q  Children are more likely to be socially connected – and stay connected – when they are involved with integrated services throughout their early childhood and into further education (Bertram et al., 2003).

The specific outcome measures used through the projects followed the CfC Strategic Plan 2005 and included (Communities and Families Clearinghouse Australia, Australian Institute of Family Studies, AIFS 2009a)

q  improved child language and literacy;

q  established and/or strengthened community hubs;

q  increased awareness of where to go for help with children;

q  increased social connectedness and support;

q  community members taking an active interest in developing hubs; and

q  shared learning as the model develops

Quantitative and qualitative data were collected across the five projects working under the Hub strategy and some of the findings were as followed (AIFS, 2009a):

q  Improved language and literacy:  Between 2005 and 2006, results for the six targeted suburbs showed improved numbers of children in Prep who had accessed pre-school or child care, and a 9.5% reduction in children’s development vulnerability in two or more domains.  Meadow Fair North Primary School saw an improvement in the number of children starting school that had a pre-school experience, from 53% in 2005 to 86% in 2006.  Key outcomes for children were development of listening, social and engagement skills, and for parents, outcomes were reported to be improved interactions with children.

q  Established/strengthened community hubs:  It was reported that services worked together to share expertise and develop new ways of engaging families through informal activities such as playgroups, family fun days, and transition to school programs.

q  Increased awareness of where to go for help with children:  It was reported that families experienced increased social connectedness due to hub involvement. Parents also reported opportunities for connection such as social meetings outside of organised activities.

Child Care Info Connect

This was a CfC project started in March 2006 as a parent education and support program operating within childcare centres in the Broadmeadows area in Melbourne. The project encompassed information provision on topics relevant to families in the community through parent information sessions and written material distribution translated into community languages.

The overall program outcomes were listed as (AIFS, 2009b):

q  enhanced parent knowledge in the areas of school readiness indicators and positive behaviour guidance strategies;

q  increased parental awareness of early childhood education;

q  improved communication between parents and child care staff; and

q  greater confidence amongst child care staff to offer parent education

One of the project’s objectives was to promote positive parent–child interactions and improved child health and development. 

The project was underpinned by research that parents seek information from people with whom they already have relationships, such as childcare workers.  Additionally, research indicated that programs involving parents or other caregivers of young children can improve outcomes in terms of children’s behaviour, learning and health in later life (McCain and Mustard, 1999, as cited by AIFS, 2009b). 

The approach used by the project promoted relationship building between primary caregivers through joint participation at information sessions, due to research on parenting practice on the importance of connections between parents and a support network as critical protective factors (Ospfsky and Thompson, cited by AIFS, 2009b).

Over the period of the project from 2006-07, 13 information sessions were held at six child care centres, with a total of 114 parents and 80 childcare staff attending.

The following evidence of outcome achievement was reported based on parent evaluation surveys collected after information sessions (AIFS, 2009b).

q  Increased knowledge on school readiness :  Two independent information sessions on school readiness were attended by 23 parents.  Over half reported an improvement in their school readiness knowledge after the session.

q  Increased knowledge of managing challenging behaviour: Two independent information sessions on management of challenging behaviour were attended by 35 parents.  Over two thirds reported an improvement in their knowledge of managing challenging behaviours. The main theme arising from comments left by parents in this area was on improved awareness of addressing behaviour issues in a positive manner. Some parents also reported their improved awareness that positive behavioural guidance was more useful than negative behavioural guidance in the management of challenging behaviours.

3.3.2                   Invest To Grow

Program services and target groups

Another component of the Stronger Families and Communities Strategy is the Invest to Grow (ItG) initiative. This initiative provided funding for early childhood programs and the development of tools and resource materials to be used by families, communities and professionals supporting families and young children. Similar to Communities for Children, Invest to Grow had a focus on prevention and early intervention, with the aim to develop the Australian evidence base around these issues and support the expansion of successful programs models.

Research on efficacy

Past research has been carried out on the effectiveness of various Invest to Grow initiatives. One of these initiatives and its effectiveness in reducing NFF is discussed.

Through the Looking Glass (TTLG)—A Partnership in Parenting Project

This ItG initiative began in July 2005 to deliver a multi-faceted collaborative intervention for families where there was an identified compromised attachment relationship between the parent and child/ren. It targets families from a range of backgrounds that exhibit multiple risk factors including anxiety, depression and social isolation. The initiative works through providing quality relationship based childcare, engaging parents in a support network, involving parents in therapeutic group work and individual counselling (AIFS, 2009c). Additionally, the project aims to support and build links to community social support and networks. Mothers are targeted as primary caregivers.

Key activities undertaken by the initiative include (AIFS, 2009c):

q  the provision of childcare delivered from a primary care giving approach and as part of the intervention;

q  strength of a multi-disciplinary team approach (i.e. strength of working in partnership where childcare staff are therapeutic partners);

q  video work;

q  combination of the group program and one-to-one individual work; and

q  partnership and collaboration.

This initiative developed out of a number of theoretical models with the base being Attachment Theory (AIFS, 2009c). Past evaluation of the project has been focussed on assessing the following outcomes (AIFS, 2009c):

q  improved child wellbeing and involvement;

q  decreased participant social isolation;

q  increased parenting confidence/competence; and

q  increased staff capacity to be emotionally available.

An external evaluation of the project was undertaken by Adelaide University staff in conjunction with Lady Gowrie Centre in 2007, utilising a series of pre- and post- project measurement tools, surveys, interviews and observations to collect data from others children and staff at the five sites of the project. Methods included qualitative approaches and quantitative approaches.

A number of standardised approaches were used to measure changes for mothers and children including (AIFS, 2009c):

q  The Hospital Anxiety and Depression Scale (HADS) measuring change in a client’s emotional state using anxiety and depression subscales.

q  Parenting Stress Index Short Form (PSI/SF) questionnaire measuring stress in the parent-child system.

q  Emotional Availability (EA) framework measuring changes in the parent–child relationship based on parent dimensions (sensitivity, structuring, non-intrusiveness and non-hostility) and child dimensions (child responsiveness to parent and child involvement with parent).

q  Children’s Wellbeing and Involvement Observations measuring a child’s levels of wellbeing and involvement while attending childcare.

The evaluation resulted in the following findings (AIFS, 2009c):

q  Improved children’s wellbeing and involvement: Mothers reporting improved positive child behaviour on completion of the project with 85% (out of a total of 45) perceiving improved child behaviour. A follow-up survey confirmed sustained changes in improved child behaviour with 88% (out of a total of 44) of responding mothers reporting lasting positive changes in their children’s behaviour since taking part in the project. These results were reflected by standardised testing, with the mean wellbeing score for children rising from 3.1 to 4 and involvement score rising from 2.8 to 3.9. It is stated that the change in these scores was significant with large effect sizes reported for both cases.

q  Increased parenting confidence/competence and decreased participant social isolation.  Parent impacts reported  from the 61 families that had completed the project were:

¬  100% (n=53) learnt more about parenting and attachment.

¬  98% (n = 52) were more confident in responding to their child’s needs.

¬  96% (n = 52) cope better as a parent since taking part.

¬  92% (n = 49) agreed that the TTLG helped them to feel closer to their child (with 35.8% (n = 19) strongly agreeing this was the case).

¬  92% (n = 46) acquired understanding of children’s attachment needs.

¬  89% (n = 46) have acquired understanding of children’s exploration needs.

¬  88% (n = 44) described positive changes in the way they do things with their children.

¬  88% (n = 44) reported increased responsiveness and ability to read cues.

¬  80% (n = 40) reported increased confidence in responding to their children’s attachment needs.

Furthermore, 92% of mothers reported the project had helped them feel closer to their child and 88% noted lasting positive changes in themselves since completion of the project. The pre- and post- application of a number of standardised psychometric instruments revealed statistically significant improvements in parental stress, anxiety, depression, child responsiveness to parent, child wellbeing and involvement, parent sensitivity and parent structuring.

3.4                        Triple-P (positive parenting program)

Program services and target groups

Triple‑P is a behavioural family intervention that is designed to improve family protective factors as well as reduce factors that are associated with severe behavioural and emotional problems in preadolescent children.  The program was developed at the University of Queensland as a multi‑level intervention designed to improve the quality of advice available to parents. 

Triple‑P has been operational for 25 years and is a comprehensive public health model intervention with the specific aims (Sanders, 2003) of:

q  enhancing the knowledge, skills, confidence, self sufficiency and resourcefulness of parents of preadolescent children;

q  promoting nurturing, safe, engaging, non‑violent and low conflict environments for children; and

q  promoting children’s social, emotional, language, intellectual and behavioural competencies through positive parenting practices. 

This program has five levels of intervention (Figure 3‑2) on a tiered continuum of increasing strength for parents and children from birth until the age of 16 years.  These levels of intervention include a ‘media-based parent information campaign’, ‘health promotion, strategy/brief selective intervention’, ‘narrow focus parent training’, ‘broad focus parent training’ as well as ‘intensive family intervention modules’.

Figure 3 ‑ 2 :  Diagram of Triple P system of parenting and family support

Source:  Sanders (2003).

Research on efficacy

A number of reports have found that Triple-P is effective in reducing child disorders, and some reports have studied the cost effectiveness of the program and also found it to be favourable.  Many of these studies discussed the need for a public health approach so that the intervention reaches more parents. 

A meta-analysis of 29 randomised controlled trials measuring the effectiveness of Triple‑P was completed by Nowak and Heinrichs (2008).  This study confirmed the efficacy of Triple‑P in improving parenting skills, child problem behaviour as well as parental wellbeing (Table 3‑9 and Table 3‑10).  Overall effect sizes of Triple P for parenting and child problem behaviour range between 0.35 and 0.48 for between-groups, while within groups effect sizes were between 0.45 and 0.57 in post intervention comparisons.

Table3 ‑ 9 : Triple P, means for unconditional and conditional models

Source:  Nowak and Heinrichs (2008).
Note:  unconditional measures are without explanatory variables and conditional models include all potential explanatory variables.  Confidence intervals are 95%.

Table 3 ‑ 10 :  Estimated effect sizes for different intensities and formats of Triple P

Source:  Nowak and Heinrichs (2008).
Note:  between group effect sizes for different sources of assessment and study designs in controlled comparisons.

Nowak and Heinrichs (2008) concluded that Triple-P is effective for socially and culturally diverse families and that it is a well-researched intervention based on high-quality studies that obtains positive results in meta-analysis and is cost effective.

Another meta-analysis was conducted by de Graaf et al (2008) using a random-effects model.  This study found that Level 4 of the intervention ( Figure 3‑2, broad focus parent training) reduced disruptive behaviour in children, which was maintained over time (an overall effect size of 0.88 at post-intervention and an effect size of 1.00 at the long term follow-up assessment of the child behaviour).  The meta-analysis estimated an overall mean effect range of 0.42 to 0.65 for post-intervention for child behaviour problems and a range of 0.54 and 0.51 for dysfunctional parenting, as shown in Table 3‑11.  These results are for all Levels of the Triple-P program and are comparable to the results from Nowak and Heinrichs (2008). 

Table 3 ‑ 11 :  Results of meta-analyses of Triple P across modalities

Source:  de Graaf et al (2008).

An article by Sanders (2008) discussed how parenting programs can improve the mental health of children and benefit the community, although the traditional clinical models of service delivery of parenting programs such as Triple-P reach relatively few parents.  Sanders (2008) indicate that a public health approach is required to ensure more parents benefit from programs like Triple-P. 

The cost effectiveness of expanding the scope of Triple-P was extrapolated from Mihalopoulos et al (2007) who examined the costs and prevalence of conduct disorders in children and adolescents to society in both financial and emotional terms (Table 3‑12).

Table3 ‑ 12 :  Costs of conduct disorder s ( 2002 - 03 A $)

Source:  Mihalopoulos et al (2007).

Mihalopoulos et al (2007) investigated whether investment in Triple-P is a worthwhile use of government health funds, and identified and measured benefits from Triple-P ( Table 3‑13).  A sensitivity analysis surrounds the data, with the model conducted using different prevalence rates, which was deemed to be the most important uncertainty estimate in the analysis.