Journal of Child Psychology and Psychiatry 56:9 (2015), pp 1026–1033

doi:10.1111/jcpp.12438

The population cost-effectiveness of a parenting intervention designed to prevent anxiety disorders in children Cathrine Mihalopoulos,1 Theo Vos,2 Ronald M. Rapee,3 Jane Pirkis,4 Mary Lou Chatterton,1 Yu-Chen Lee,1 and Rob Carter1 1

Deakin Health Economics, Deakin University, Melbourne, Vic., Australia; 2Institute of Health Metrics and Evaluation, University of Washington, Seattle, WA, USA; 3Centre for Emotional Health, Macquarie University, Sydney, NSW, Australia; 4Centre for Health Policy Programs and Economics, University of Melbourne, Melbourne, Vic., Australia

Background: Prevention and early intervention for anxiety disorders has lagged behind many other forms of mental disorder. Recent research has demonstrated the efficacy of a parent-focussed psycho-educational programme. The programme is directed at parents of inhibited preschool children and has been shown to reduce anxiety disorders at 1 and 3 years following intervention. The current study assesses the cost-effectiveness of this intervention to determine whether it could provide value-for-money across a population. Method: A cost-utility economic framework, using Disability-Adjusted-Life-Years (DALYs) as the outcome, was adopted. Economic modelling techniques were used to assess the incremental cost-effectiveness ratio (ICER) of the intervention within the Australian population context, which was modelled as add-on to current practice. The perspective was the health sector. Uncertainty was measured using multivariate probabilistic testing and key assumptions were tested using univariate sensitivity analysis. Results: The median ICER for the intervention was AUD$8,000 per DALY averted with 99.8% of the uncertainty iterations falling below the threshold value-for-money criterion of AUD$50,000 per DALY averted. The results were robust to sensitivity testing. Conclusions: Screening young children in a preschool setting for an inhibited temperament and providing a brief intervention to the parents of children with high levels of inhibition appears to provide very good value-for-money and worth considering in any package of preventive care. Further evaluation of this intervention under routine health service conditions will strengthen conclusions. Acceptability issues associated with this intervention, particularly to preschool staff and parents, need to be considered before wide-scale adoption is undertaken. Keywords: Economic evaluation, anxiety disorders, children, prevention.

Introduction Anxiety disorders are the most common mental health conditions affecting both children and adults – with international point prevalence estimates suggesting that up to 5% of children and adolescents meet the criteria of an anxiety disorder at any one time (Rapee, Schniering, & Hudson, 2009). Anxiety disorders in childhood can cause impairment in school performance, family and social functioning. Ezpeleta and colleagues (Ezpeleta et al., 2001) suggest that the impairment caused by childhood anxiety is as great as that caused by depression and even greater than that caused by Attention Deficit Hyperactivity Disorder. Furthermore, longer term epidemiological studies suggest that childhood anxiety disorders may predict adult mental disorders, particularly anxiety and depression (Bittner et al., 2007; Last, Perrin, Hersen, & Kazdin, 1996; Pine et al., 1998). While the treatment of childhood anxiety has received extensive attention in the literature (Rapee et al., 2009), the prevention of such disorders is only

Conflict of interest statement: See Acknowledgements for disclosure.

beginning to receive attention (Bayer et al., 2009). Preventive interventions for mental disorders are commonly classified as universal, selective or indicated (Mrazek & Haggerty, 1994). Universal interventions are delivered to entire populations (or subpopulations) regardless of individual risk. The FRIENDS programme, initially developed in Australia is the best known universal intervention for the prevention of internalising disorders. It is based on a cognitive behavioural paradigm which is used in a school setting to teach children strategies to cope with anxiety and challenging situations (Lowry-Webster, Barrett, & Dadds, 2001). While evaluations of this intervention have found promising results (LowryWebster et al., 2001), others have not (Lock & Barrett, 2003). Although universal trials show some promise, the overall effect-sizes have not been large (Neil & Christensen, 2009). Selective interventions are delivered to populations with known risk factors for the development of mental disorders. Such interventions are fairly rare in the prevention of mental disorder literature, possibly due to the infancy of the literature investigating risk factors and the development of mental disorders (Lyneham & Rapee, 2011). A wellevaluated intervention by Rapee et al. (2005) screened young preschool-aged children (aged between 3 and 5)

© 2015 Association for Child and Adolescent Mental Health. Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA

doi:10.1111/jcpp.12438

for an ‘inhibited’ or shy temperament. The parents of children who met the criteria for an inhibited temperament were then offered a six-session parenting course. The evaluation of this intervention found a beneficial effect including reductions in the numbers of diagnosed anxiety disorders at 12 months with a recent study demonstrating that this effect is sustained at 3 years postintervention (Rapee et al., 2010). A recent longitudinal follow-up of this evaluation found that while the overall effects were not significant for the whole sample, effects appeared to persist for girls (Rapee, 2013). Lastly, indicated interventions refer to the selection of individuals who are already showing signs of the disorder. Dadds and colleagues developed an indicated intervention for slightly older children (7–14 years) who were screened for anxiety problems and then offered a 10-week child and parent-focussed psychosocial intervention (Dadds et al., 1997). A high percentage of the children participating in the trial had a diagnosable anxiety disorder (55%) and by 6 months the intervention group appeared to have less anxiety disorders than the control group; however, these differences had disappeared by 12 months (Dadds et al., 1999). The differences between the two groups reemerged at 2 years but only for children who had more severe anxiety problems at intake. Recently, the effectiveness and cost-effectiveness of two indicated interventions for the prevention of childhood disorders has been evaluated for children aged 8–12 (Simon, Dirksen, B€ ogels, & Bodden, 2012). The interventions evaluated were a child-focused intervention, a parent-focused intervention and no intervention. The study’s main outcome was the incremental cost-effectiveness per ‘ADIS improvement’. This study found no differences in outcomes between the child intervention and the parent intervention, but there were significant differences between both these interventions and no intervention. As the outcomes of this study were not expressed in generic units, such as DALYs or QALYs, it is difficult to determine whether the reported ratios represent good value-for-money or not. In summary, the efficacy of universal interventions for the prevention of anxiety is not well established with respect to an impact on diagnosed anxiety at follow-up. The results for selective or indicated interventions show a more positive effect in terms of diagnosis up to 3 years postintervention. The intervention developed by Rapee and colleagues has particular merit as the effect of the intervention in terms of reducing diagnosable anxiety disorders persisted for at least 3 years postdelivery and seemed to strengthen over time. Furthermore, the intervention can be easily rolled out into standard practice. The aim of the current study was to evaluate the population cost-effectiveness of the parenting intervention designed by Rapee et al. (2005), using a cost-utility framework, to determine whether such an intervention is a worthwhile use of healthcare resources. © 2015 Association for Child and Adolescent Mental Health.

Cost-effectiveness of parenting intervention

1027

Method The current study was undertaken within the context of a larger study, called Assessing Cost-Effectiveness in Prevention (ACE-Prevention). The aim of ACE-Prevention was to evaluate the cost-effectiveness credentials of over 100 preventive interventions for noncommunicable disease to provide value-formoney advice to health policy decision-makers. The project tried to minimise methodological confounding as all economic evaluations were undertaken using a standardised framework (Carter, Vos, Barendregt, & Mihalopoulos, 2005; Carter et al., 2008). The key features of the ACE-Prevention approach relevant to the current study are described in Table 1. Notably, key parameters have been updated to a 2013 reference year. A ‘value-for-money’ threshold of AUD$50,000/DALY averted was used in ACE-prevention. Interventions whose incremental cost-effectiveness ratios (ICERs) fell below this threshold were considered ‘good’ value-for-money and interventions whose ICERs fell above this threshold were considered ‘poor’ valuefor-money. This threshold has been used in previous Australian economic evaluation studies and is a ‘rule of thumb’ within the Australian context (Carter et al., 2008). Not all issues which are important to health decisionmakers can be easily captured in the technical aspects of an economic evaluation. Therefore, to try to capture other issues not inherent within each of the ICERS, all results were subjected to a second-stage filter analysis based on the following criteria: 1. the strength of the evidence base; 2. effects on equity; 3. ‘feasibility’ and ‘sustainability’ in terms of work force considerations, financing mechanisms and health system structure; 4. ‘acceptability’ of the intervention to key stakeholders including clinicians, patients and policy-makers; and 5. other important beneficial or harmful effects associated with the intervention not identified in the analyses. All technical ICERS were discussed qualitatively in terms of their performance on these second-stage filter considerations. The second-stage filter criteria were developed by the ACEPrevention stakeholder reference group which comprised stakeholders from various levels of the Australian government (both Commonwealth and States), clinical experts across the various diseases and risk factors considered by the project (including mental health and nongovernment agencies) and other stakeholders from Australian health agencies (e.g. public health associations). Important issues for each intervention across the different criteria were firstly specified by the key researcher in the main project written communication documents (briefing papers and reports) and presented to all stakeholder reference group members via the ACE-Prevention website or at stakeholder meetings. Input (both written and oral) by any other technical advisors used for the evaluations (such advisors did not sit on the project reference group) was included in the initial written project briefing materials. All written (or oral) comments from the stakeholder reference group was then synthesised by the researcher in the final written documents and recirculated to all stakeholders, thus ensuring all views were appropriately captured. This was considered the most efficient approach to reflect important second-stage filter considerations by multiple stakeholders The proforma for all written materials is available from the ACE-Prevention Protocol (Carter et al., 2005).

Intervention description To model the population cost-effectiveness of such an intervention some important adjustments to the original intervention design were required to make it appropriate for broader implementation nationally. As with the original intervention study, screening children between the ages of 3–5 for inhibition

1028

Cathrine Mihalopoulos et al.

J Child Psychol Psychiatr 2015; 56(9): 1026–33

Table 1 Main methodological considerations of the ACE-prevention study Economic Evaluation Design

Cost-utility analysis – costs are measures in dollars and outcomes are measured using a generic outcome metric (refer below)

Intervention selection

Sourced from the published literature ensuring available evidence of effectiveness

Economic perspective

Health sector – with costs divided into government and private

Population

Target population in ACE-Prevention was the Australian population in 2003 but has been updated to 2013 for the current study

Outcomes

Outcomes are evaluated using Disability-Adjusted-Life-Years (DALYs). The DALY is conceptually similar to a quality-adjusted-life-year (where all years of life lived in a health state are weighted for the quality of life). Both mortality and morbidity effects of interventions can be measured by the one metric making across-disease comparisons possible. The weights used to value the health states within the DALYs come from the Australian burden of disease studies (Begg et al., 2007)

Costs

Costs and cost-offsets (costs saved by diseases/disorders avoided) of interventions are measured in Australian dollars and expressed in the base year of 2013. Unit cost prices not available for the 2013 reference year are adjusted using published Australia health price deflators (Australian Institue of Health and Welfare, 2014). The inclusion of time and travel costs using standardised unit costs developed for the project (Carter et al., 2005) and updated using 2013 estimates are also included though reported separately to the main results

Discount rate

A 3% annual discount rate is applied to all costs and outcomes, as is commonly used in many Australian studies

would occur in the preschool setting, whereby preschool teachers distribute the brief screening questionnaire to parents to complete at home. Once the questionnaire is completed it is returned to the preschool setting. From there, it is assumed that the questionnaires are centrally assessed by a coordinating unit, primarily staffed by project officers with psychology qualifications located in each Australian state. The proportion of children who meet the inhibition temperament criteria on the short screening questionnaire (16%) was sourced from the Rapee et al. (2005) study. The original effectiveness trial included a secondary laboratory assessment of children to validate the questionnaire screen. However, only a small percentage of children (18%) who met questionnaire screening were excluded based on direct observation and such secondary assessment is unlikely to occur in routine practice. We therefore assumed the intervention would be offered to all parents of children who screen positive after initial questionnaire completion – though we have not included a benefit for the 18% of children who were screened out at the second stage – that is we have only modelled a benefit to 82% of children whose parents attend the intervention. In this way we are not assigning a benefit to children which were not included in the original evaluation of the intervention. The comparator was ‘do-nothing’ because no alternative interventions are currently being routinely offered in Australia.

Efficacy/effectiveness of the intervention The proportions of children with one or more anxiety diagnoses for the intervention and control group in the Rapee trial were: 50.8% versus 63.5% at 12 months; 37.8% versus 68.4% at 24 months; and 39.5 versus 68.8% at 36 months (Rapee et al., 2005, 2010).

Modelling to health outcomes A 3-year model was developed to assess the cost-effectiveness of this intervention. An 11-year model was also developed as a sensitivity analysis (refer below). A lifetime model was not developed because the long-term epidemiology of anxiety disorders in children is not well investigated. The only two epidemiological studies which included younger children in a follow-up cohort were not community samples (Cantwell & Baker, 1989; Last et al., 1996). One study was of children referred to a speech and hearing clinic and used DSM-III nosology (Cantwell & Baker, 1989), the second was a follow-up

study of clinically referred children. The Great Smoky Mountains Study (Costello et al., 2003; Ezpeleta et al., 2001) was a well-conducted study into the prevalence and development of psychiatric disorders into childhood and adolescence but the youngest children comprising the sample were 9. One other study by Caspi et al. assessed children for behavioural patterns at 3 years of age and then retested them at 21 to determine whether certain behavioural patterns predicted the latter onset of mental disorder (Caspi, Moffitt, Newman, & Silva, 1996). However in this study, inhibition was not clearly defined and the children were not assessed for the presence of disorders until they reached the age of 21. The limiting factor in modelling the long-term course of anxiety is the lack of information on the number and length of episodes for children with anxiety disorders. Even in adults the epidemiology is not well documented (Yonkers et al., 2003). The starting population of the model is the total number of 3- and 4-year olds in 2013 adjusted for the proportion who attend preschool (total population of 602,152 with 95% attending preschool). Five-year olds were not included as many 5-year olds are already attending school in Australia. The 2003 Australian Burden of Disease Study jointly modelled depression and anxiety (Begg et al., 2007) and used the Child Behaviour Checklist (CBCL) to assess the presence or absence of these internalising disorders in children. However, the CBCL does not differentiate well between the different types of internalising disorders (i.e. depression and anxiety) and certainly does not differentiate between the subcategories of anxiety disorders. The CBCL is also not very sensitive to the identification of anxiety disorders in preschool children. Existing longitudinal studies demonstrate that children have different types of anxiety disorders with a greater proportion meeting the criteria for anxiety rather than depression. As the most recent burden of disease study used composite measures of depression and anxiety, we used a number of sources to estimate the key epidemiological parameters for the economic evaluation (including the older burden of disease studies which separated depressive and anxiety disorders). A number of plausible epidemiological estimates largely stemming from the Australian burden of disease studies for 2003 and 1996 were used in the current study (Begg et al., 2007; Mathers, Vos, & Stevenson, 1999). The more recent Burden of Disease studies were not used as childhood anxiety disorders are not well differentiated in these studies (Whiteford et al., 2013). For duration of time symptomatic in a year the range of values for time symptomatic (26%–56%) for the two anxiety disorders modelled for children in the 1996 study © 2015 Association for Child and Adolescent Mental Health.

doi:10.1111/jcpp.12438 (separation anxiety disorder and posttraumatic disorder) were used. In Burden of Disease studies, disability weights were used to represent the morbidity component of the various diseases and disorders. These weights ranged from 0, representing no morbidity issues (or good health) to 1 which represented death. The disability weight for mild/moderate social phobia was used (0.17) in the current study as this is the most common diagnostic category for this group of children (Rapee et al., 2010). This is a lower weight than that used in the 2003 burden of disease studies for children aged 5–9 of 0.19. The weights and duration of illness were both varied in the uncertainty analysis. The flow of the model along with the associated probabilities of events occurring is described in Figure 1.

Costs of intervention and cost-offsets The screening costs included: 1 hr of teachers’ time which includes the costs of administering the screening instruments plus costs of some brief training (administered by the dedicated project officers in an online or telephone based capacity) in the administration and purpose of the screening questionnaire; and processing the screening questionnaires (assumed to require 16 project officers across Australia with psychology training plus postage and stationary costs for results). Intervention costs consist of up to six, 1.5 hr group sessions with a psychologist valued at the group therapy Medicare Benefits Schedule rate for psychologists (this amount includes preparation time for the sessions as well as administrative allowance). Details of intervention costs are in Table 2. The costoffsets were valued at $312 – the cost of treating a case of anxiety and depression between ages 5 and 14 from the 2001 Disease Costs and Impacts Study (DCIS) study (Australian Institute of Health and Welfare, 2005) inflated to 2013 values. The costs of time and travel for parents to complete the screening questionnaire and attend the intervention have also been included. Importantly, it is assumed one hour is required for the screening instrument – which is probably generous as the questionnaire usually takes less than 10 min

Cost-effectiveness of parenting intervention

to complete, though this time also includes return of the questionnaire, reading of results and time spent deciding on whether to attend the intervention. For parents who do attend the intervention, we also included 1-hr preparatory time to undertake activities such as contact with the intervention delivery staff to organise time and dates. Travel costs were based on a standardised weighted average cost of a trip to a health professional used across all ACE-Prevention evaluations. Further details regarding specifics of unit costing are available from the first author.

Key assumptions. Some of the key assumptions of the current analysis included: 1. Seventy-five per cent of preschools agree to participate in the intervention (based on estimates from willingness of preschools to participate in other screening initiates); 2. Teachers are able and willing to distribute the screening tool; 3. The screening programme is centrally administered by psychology-trained project officers. These officers may sit in either general health or specific prevention-focused organisations in either government or nongovernment sectors. 4. All parents are given the parenting intervention in group format and one therapist provides the group therapy sessions; 5. Return rates of questionnaires and participation rates are similar to those in the trial; 6. Intervention participants are similar to those in the trial in terms of background characteristics; and 7. Adherence rates to the intervention are similar to those observed in the trial. Uncertainty and sensitivity analysis. To incorporate uncertainty associated with all cost and outcome data, we use Monte Carlo simulation modelling and present uncertainty ranges as well as point estimates for benefits, costs, cost-offsets and cost-effectiveness ratios. Monte Carlo simulation modelling basically means that the model is recalculated 2000 times using different values of key model inputs as defined in Table 3. The distributions used around each parameter are based on the

Figure 1 Flow of model – references (Rapee et al., 2005, 2010; Wake, Gerner, & Gallagher, 2005; Wake et al., 2008) © 2015 Association for Child and Adolescent Mental Health.

1029

1030

Cathrine Mihalopoulos et al.

J Child Psychol Psychiatr 2015; 56(9): 1026–33

Table 2 Intervention unit costs Cost description

Unit cost

Teacher costs

$44.34 per hour

Inhibition Scale plus postage costs Psychologist costs (group therapy) Time costs

$1.30

Travel costs

$8.90

Source OECD, Table X2.4a. Teachers statutory salaries at different points in their careers Mihalopoulos, Vos, Pirkis, and Carter (2011)

$30.13 ($25.62 government cost, $4.52 out of pocket)) $23.49

ACE-Prevention protocol (Carter et al., 2005) and Briggs, Claxton, and Sculpher (2006). The point-estimate results along with the associated uncertainty are based on these 2000 iterations. Sensitivity analysis is a useful addition to probabilistic analysis, as it allows the impact of individual assumptions or key design features of an intervention to be addressed. We explored in a univariate analysis the variations in the disability weight. This was considered necessary as the weight for social phobia was used and there were no childhood weight estimates for the other anxiety diagnoses. We considered extending the time horizon to 11 years to include the most recent outcome data available for this intervention. However, we elected not to do this as the overall results at 11 years were no longer significant and we did not have sufficient information to know at which point this occurred. Furthermore, any benefits beyond 3 years would be favourable to the intervention, so the results from any such analyses would only improve the current cost-effectiveness ratios.

Results Based on the assumptions made in the model, approximately 1000 children less will have anxiety disorders in the first year post the intervention, 3000 less in the second year post the intervention and 3000 less in the third year. Table 4 contains a summary of key results. The DALYs averted by the intervention were modest and the bulk of the intervention costs accrued to the government. The median ICER for this intervention

Medicare Benefits Schedule (Australian Government Department of Health and Aging, 2013) Weighted average of working and nonworking population. Source: Australian Bureau of Statistics Carter et al. (2005)

was $8,000/DALY averted with 99% of the uncertainty iterations falling below the threshold value-for-money criterion of $50,000 per DALY. This means that the intervention is classified as very good value-for-money. Figure 2 shows that the uncertainty iterations predominantly fell in the north-east corner (‘health gain at a cost’) of the cost-effectiveness plane (Figure 2). A cost-effectiveness plane is a four-quadrant plane in which results in the northeast quadrant represent health gain at a cost, results in the south-east quadrant represent health gain at a cost-saving (sometimes called ‘dominant’), results in the south-west quadrant represent health loss at a cost-saving and results in the north-west quadrant represent health loss at a cost (sometimes called ‘dominated’). The exclusion of cost-offsets increased the ICER, though the conclusions of the study did not change. The average cost per child without an anxiety disorder is $540 (with cost-offsets) and $800 without cost-offsets included. The inclusion of time and travel increased the ICER with cost-offsets to $15,000/DALY averted ($6,000–$38,000).

Sensitivity analyses. The results of the evaluation were sensitive to the disability weight used, though the overall conclusions of the analysis did not change. For example, the weight would need to fall

Table 3 Uncertainty parameters used in the simulation analysis Parameter Proportion of schools participating in screening Proportion of parents returning screening questionnaire Proportion of parents taking up intervention Proportion of children scoring above cut-off Proportion of children who are offered the intervention who may not qualify (i.e. incur a cost but no benefit) Session number Proportion of time symptomatic Efficacy – % with anxiety disorders at 12, 24 & 36 months

Private costs (including unit costs of practice nurses, psychologists, screening materials, bibliotherapy)

Values

Uncertainty distribution

Source

75% 25%

Triang (0.5, 0.75, 0.9) Beta

Estimate based on trial Trial

63% 16% 18%

Beta Beta Beta

Trial Trial Trial

Discrete (1, 2, 3, 4, 5, 6) (0.25, 0.25, 1, 3, 4, 5) Uniform Beta distributions

Trial

Triangle

Protocol

Up to 6 26–56% Int Control 12 51% 63% 24 38% 68% 36 40% 69% 20% of unit cost

Burden of Disease studies Trial

© 2015 Association for Child and Adolescent Mental Health.

doi:10.1111/jcpp.12438

Cost-effectiveness of parenting intervention

1031

Table 4 Results of reference case analysis for the psychological intervention Median point estimate (95% uncertainty interval) DALYs averted Cost of the intervention Cost-offsets Total costs ICER (with cost-offsets) ICER (without cost-offsets)

Figure 2 Cost-effectiveness of the parenting interventions designed to prevent childhood anxiety illustrated on a costeffectiveness plane with $50,000 per DALY threshold line

460 Gov: $5.2M Priv: $0.44M $1.8M $3.8M $8,000 $12,000

(210–850) (3.5M–6.7M) (0.09M–0.637M) (0.74–3.1M) (1.6M–5.6M) (2,700–22,000) (6,200–26,000)

attached to the identification of children at risk of anxiety problems.

Discussion to approximately 0.03 for the ICER to approach $50,000/DALY. Lastly, variations in the discount rate made very little difference to results.

Second-stage filter considerations The strength of the evidence upon which the analyses were based was considered sufficient, though trials outside the university in which the intervention was developed are required to ensure generalisability. There are potentially important acceptability issues of this intervention to a number of important key stakeholders. Firstly, some preschools may be unwilling to engage in such a screening procedure. Similarly, there may be broader community concerns associated with the stigmatisation of preschool children who screen ‘positive’. Parents may also be reluctant to attend such an intervention. The way the intervention is introduced may influence attendance rates (e.g. to promote resilience and confidence rather than to prevent anxiety disorders). The availability of a trained and available workforce capable of delivering the intervention in a timely and consistent manner needs to be carefully considered, along with a sustainable financing mechanism (current publically financed psychological services in Australia only extend to diagnosed mental disorders and would largely exclude this type of intervention). There are other effects which were not captured in the technical ICER analyses (called spill-over effects). For example, there may be health benefits to parents associated with such an intervention through less ‘worry’ or ‘stress’ associated with raising an inhibited child as well as greater confidence associated with parenting more generally (which could benefit other children as well, not just the inhibited child identified via the screening). There are unlikely to be many negative spill-over health effects associated with such an intervention, as the intervention is largely about promoting resilience and confidence, except possibly due to stigma © 2015 Association for Child and Adolescent Mental Health.

This study has demonstrated that this well regarded screening and parenting intervention for the prevention of anxiety in children and adolescents represents very good value-for-money. This is the first study, as far as we are aware, which has considered the costeffectiveness of a parenting intervention to prevent anxiety disorders in young children using a costutility frame. Even though Simon et al. (2012) report the cost-effectiveness ratios of a parent-focused intervention, it is difficult to compare the results of this study to the current study as the primary outcome reported in Simon et al. (2012) of ‘incremental cost per ADIS improvement’ does not have an associated value-for-money threshold criterion. There are some limitations of the current study. The efficacy measure as well as adherence rates are based on a single study, albeit a good quality randomised trial. This trial was an efficacy trial and further evidence of effectiveness in a routine setting is required. Cost-offsets included in the study may have been an underestimate as they were based on the cases of disorder averted and do not necessarily include children with multiple episodes in a single year for which help may have been sought. The limited health sector costing perspective also meant that costs, such as child care, have been excluded. The impact of such cost exclusions on the study results is unknown as the inclusion of such costs may increase the costs of the intervention, but at the same time may also reduce the incremental costs of the intervention due to less anxiety disorder cases attributable to the intervention. Many parameters in the economic evaluation are based on either adult populations (e.g. average duration of symptoms or disability weights) which may not be representative of anxiety problems found in very young children or broad estimates of disease duration which do not include the full spectrum of anxiety disorders in children. However, variations in these parameters show that the intervention continues to represent good value-for-money. For example, the disability weight at which the ICER approaches $50,000/DALY is 0.03 which is similar in magnitude to the DW of mild anxiety disorders in the

1032

Cathrine Mihalopoulos et al.

most recent international burden of disease studies (Salomon et al., 2012). The model used in this study only extended to 3 years to mirror the existing evidence base of significant effects. While the 11-year follow-up results showed some impact for girls this subgroup result requires replication before it is appropriate for modelling. In any case, further benefits beyond 3 years will only improve the ICER of the intervention. This shorter time frame is in contrast to other economic evaluations undertaken in ACE-Prevention, whereby lifetime models were developed (with various assumptions around the longer term effectiveness of interventions undertaken). The impact of a limited time frame means that longer term benefits (and costs) are not taken into account. Certainly the effectiveness results upon which the current analysis is based are suggestive of a sustained effect. If this effect is extended beyond 3 years the intervention would be even more cost-effective. There are issues with respect to acceptability of the intervention to preschools, psychologists as well as the end-users of the intervention, that is children with elevated symptoms of inhibition and their parents. While this intervention is a parenting intervention with the aim of building resilience in children rather than a ‘treatment’ for anxiety disorders, there may still be some concern that this may elicit distress in some parents who may feel that there is something ‘wrong’ with their children and/or are being stigmatised. There are also substantial workforce issues as these interventions are ideally delivered by psychologists who are already quite stretched within the current health care system. In order for such interventions to be routinely available governments would need to ensure that public provision of psychological services extends to the provision of preventive interventions such as these which are evidence based as well as cost-effective. While the current study was developed for the Australian context, it is important that international readers appreciate that the results may not be identical in other contexts. For example, the preschool context in Australia means that there is an avenue to screen 3- and 4-year olds as the vast majority attend

J Child Psychol Psychiatr 2015; 56(9): 1026–33

preschools. Secondly, the Medicare system means that there is a financing system available to support the introduction of such interventions, although Medicare is currently targeted towards treatment of mental disorders rather than prevention. Cost-effectiveness of interventions is always context specific, however, it is unlikely that the favourable ICER observed in the current study will become unfavourable in other settings (particularly developed country settings) as the ICER is well below commonly accepted thresholds of cost-effectiveness. Finally, this analysis presents a ‘partial’ picture of the benefit associated with this intervention as the broader benefits (or spill-overs as identified in the second-stage filter analysis) to the parents and siblings of such children have not been included in the measurement of benefits. For example, greater confidence in parenting techniques may improve parent’s own wellbeing as well as the wellbeing of other children. The inclusion of such benefits would further strengthen the economic credentials of this intervention.

Acknowledgements This original article was invited by the journal as part of a special issue; it has undergone full, external peer review. This work was funded by a National Health and Medical Research Council (NHMRC) Health Service Grant (Grant I.D: 351558) and was also supported by a NHMRC Capacity Building Grant (Grant I.D: 456002). CM is supported by a NHMRC Early Career Fellowship (Grant I.D. 1035887). The initial trial of the intervention modelled in this study was funded by NHMRC project grant 167201. R.R is the primary founder of the intervention evaluated in this study. The remaining authors have declared that they have no competing or potential conflicts of interest.

Correspondence Cathrine Mihalopoulos, Deakin Health Economics, Deakin University, 221 Burwood Highway, Burwood, Vic. 3125, Australia; Email: [email protected]

Key points

• • • • •

Anxiety disorders are highly prevalent and burdensome in children. Parenting interventions have been shown to be effective in preventing the onset of anxiety disorders in children. The cost-effectiveness of prevention of anxiety disorders in children is an important policy and practice issue. This study has found that a brief, six session parenting intervention is a very cost-effective use of limited healthcare resources. However, issues surrounding the acceptability of such interventions to preschools, parents and the wider community need to be addressed before widespread adoption can be successfully undertaken.

© 2015 Association for Child and Adolescent Mental Health.

doi:10.1111/jcpp.12438

References Australian Government Department of Health and Aging. (2013). Medicare Benefits Schedule Book, Operating from 01 July 2013, Canberra. Australian Institue of Health and Welfare, A. (2014). Health expenditure Australia 2012–2013. Canberra: Australian Institute of Health and Welfare. Australian Institute of Health and Welfare. (2005). Health system expenditure on disease and injury in Australia, 2000–01. 2 edn. AIHW (Health and Welfare Expenditure Series no. 21). Canberra: Australian Institute of Health and Welfare. Bayer, J., Hiscock, H., Scalzo, K., Mathers, M., McDonald, M., Morris, A., . . . & Wake, M. (2009). Systematic review of preventive interventions for children’s mental health: What would work in Australian contexts? Australian and New Zealand Journal of Psychiatry, 43, 695–710. Begg, S., Vos, T., Barker, B., Stevenson, C., Stanley, L., & Lopez, A. (2007). The burden of disease and injury in Australia, 2003. Canberra: AIHW. Bittner, A., Egger, H.L., Erkanli, A., Costello, J.E., Foley, D.L., & Angold, A. (2007). What do childhood anxiety disorders predict? Journal of Child Psychology and Psychiatry, and Allied Disciplines, 48, 1174–1183. Briggs, A.H., Claxton, C., & Sculpher, M. (2006). Decision modelling for economic evaluation. Oxford: Oxford University Press. Cantwell, D.P., & Baker, L. (1989). Stability and natural history of DSM-III childhood diagnoses. Journal of the American Academy of Child & Adolescent Psychiatry, 28, 691–700. Carter, R., Vos, T., Barendregt, J. J., & Mihalopoulos, C. 2005. ACE-Prevention: Economic Evaluation Protocol. Deakin University & University of Queensland. Carter, R., Vos, T., Moodie, M., Haby, M., Magnus, A., & Mihalopoulos, C. (2008). Priority setting in health: Origins, description and application of the Australian Assessing Cost Effectiveness (ACE) initiative. Expert Review of Pharmacoeconomics and Outcomes, 8, 593–617. Caspi, A., Moffitt, T.E., Newman, D.L., & Silva, P.A. (1996). Behavioral observations at age 3 years predict adult psychiatric disorders. Longitudinal evidence from a birth cohort. Archives of General Psychiatry, 53, 1033–1039. Costello, E.J., Mustillo, S., Erkanli, A., Keeler, G., & Angold, A. (2003). Prevalence and development of psychiatric disorders in childhood and adolescence. Archives of General Psychiatry, 60, 837–844. Dadds, M.R., Holland, D.E., Laurens, K.R., Mullins, M., Barrett, P.M., & Spense, S.L. (1999). Early intervention and prevention of anxiety disorders in children: Results at 2year follow-up. Journal of Consulting & Clinical Psychology, 67, 145–150. Dadds, M., Spence, S.H., Holland, D., Barrett, P.M., & Laurens, K.R. (1997). Prevention and early intervention for anxiety disorders: A controlled trial. Journal of Consulting and Clinical Psychology, 65, 627–635. Ezpeleta, L., Keeler, G., Erkanli, A., Costello, E.J., & Angold, A. (2001). Epidemiology of psychiatric disability in childhood and adolescence. Journal of Child Psychology And Psychiatry, And Allied Disciplines, 42, 901–914. Last, C.G., Perrin, S., Hersen, M., & Kazdin, A.E. (1996). A prospective study of childhood anxiety disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 35, 1502–1510. Lock, S., & Barrett, P.M. (2003). A longitudinal study of developmental differences in universal preventive intervention for child anxiety. Behaviour Change, 20, 183–199. Lowry-Webster, H.M., Barrett, P.M., & Dadds, M.R. (2001). A universal prevention trial of anxiety and depressive symptomatology in childhood: Preliminary data from an Australian study. Behaviour Change, 18, 36–50.

© 2015 Association for Child and Adolescent Mental Health.

Cost-effectiveness of parenting intervention

1033

Lyneham, H.J., & Rapee, R.M. (2011). Prevention of Child and Adolescent Anxiety. In: W. Silverman, & A.P. Field (Eds.), Anxiety disorders in children and adolescents: Research, assessment and intervention (pp. 349–366). Cambridge: Cambridge University Press. Mathers, C., Vos, T., & Stevenson, C. (1999). The burden of disease and injury in Australia: Summary report. Canberra: Australian Institute of Health and Welfare. Mihalopoulos, C., Vos, T., Pirkis, J., & Carter, R. (2011). The economic analysis of prevention in mental health programs. Annual Review in Clinical Psychology, 7, 169–201. Mrazek, P.J., & Haggerty, R.J. (1994). Reducing risks for mental disorders: Frontiers for preventative intervention research. Washington, DC: National Academy Press. Neil, A.L., & Christensen, H. (2009). Efficacy and effectiveness of school-based prevention and early intervention programs for anxiety. Clinical Psychology Review, 29, 208–215. Pine, D.S., Cohen, P., Gurley, D., Brook, J., & Ma, Y. (1998). The risk for early-adulthood anxiety and depressive disorders in adolescents with anxiety and depressive disorders. Archives of General Psychiatry, 55, 56–64. Rapee, R.M. (2013). The preventative effects of a brief, early intervention for preschool-aged children at risk for internalising: Follow-up into middle adolescence. Journal of Child Psychology and Psychiatry, 54, 780–788. Rapee, R.M., Kennedy, S., Ingram, M., Edwards, S., & Sweeney, L. (2005). Prevention and early intervention of anxiety disorders in inhibited preschool children. Journal of Consulting and Clinical Psychology, 73, 488–497. Rapee, R.M., Kennedy, S.J., Ingram, M., Edwards, S.L., & Sweeney, L. (2010). Altering the trajectory of anxiety in atrisk young children. The American Journal of Psychiatry, 167, 1518–1525. Rapee, R.M., Schniering, C.A., & Hudson, J.L. (2009). Anxiety disorders during childhood and adolescence: Origins and treatment. Annual Review of Clinical Psychology, 5, 311– 341. Salomon, J.A., Vos, T., Hogan, D.R., Gagnon, M., Naghavi, M., Mokdad, A., . . . & Jonas, J.B. (2012). Common values in assessing health outcomes from disease and injury: Disability weights measurement study for the Global Burden of Disease Study 2010. Lancet, 380, 2129–2143. Simon, E., Dirksen, C., B€ ogels, S., & Bodden, D. (2012). Cost-effectiveness of child-focused and parent-focused interventions in a child anxiety prevention program. Journal of Anxiety Disorders, 26, 287–296. Wake, M., Gerner, B., & Gallagher, S. (2005). Does Parents’ Evaluation of Developmental Status (PEDS) at school entry predict language, achievement and quality of life two years later? Ambulatory Pediatrics, 5, 143–149. Wake, M., Sanson, A., Berthelsen, D., & Hardy, P., Misson, S., Smith, K., & Ungerer, J. 2008. How well are Australian infants and children aged 4 to 5 years doing? Social Policy Research Paper No. 36. Department of Families, Housing, Community Services and Indigenous Affairs, (FaHCSIA). Whiteford, H.A., Degenhardt, L., Rehm, J., Baxter, A.J., Ferrari, A.J., Erskine, H.E., . . . & Vos, T. (2013). Global burden of disease attributable to mental and substance use disorders: Findings from the Global Burden of Disease Study 2010. Lancet, 382, 1575–1586. Yonkers, K.A., Bruce, S.E., Dyck, I.R., & Keller, M.B. (2003). Chronicity, relapse, and illness–course of panic disorder, social phobia, and generalized anxiety disorder: Findings in men and women from 8 years of follow-up. Depression & Anxiety, 17, 173–179.

Accepted for publication: 13 May 2015 Published online: 26 June 2015