ACTAUNIVERSITATIS
UPSALIENSISUPPSALA
2020
Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1640
Effectiveness and cost-effectivenessof indicated preventiveinterventions for depression inadolescents
An application of health economics methods
RICHARD SSEGONJA
ISSN 1651-6206ISBN 978-91-513-0879-1urn:nbn:se:uu:diva-404547
Dissertation presented at Uppsala University to be publicly examined in Bertil Hammersalen,Blåsenhus, von Kraemers allé 1, Uppsala., Thursday, 16 April 2020 at 09:00 for the degree ofDoctor of Philosophy (Faculty of Medicine). The examination will be conducted in English.Faculty examiner: Professor Jeffrey Hoch (University of California, Davis campus).
AbstractSsegonja, R. 2020. Effectiveness and cost-effectiveness of indicated preventive interventionsfor depression in adolescents. An application of health economics methods. DigitalComprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1640.84 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-0879-1.
Unipolar depressive disorders are commonly encountered conditions in clinical practicewith about 25% reporting their first episode during adolescence. Longitudinal studies showadolescent depression to be associated with an increased risk of mental illness in adulthood,healthcare usage, poor educational outcomes, unemployment, and dependency on welfarerecipiency. Therefore, adolescent depression contributes a high disease burden and impinges aconsiderable financial strain on society’s limited resources.
Several preventive interventions have been developed to prevent adolescent depression. Ofthese interventions, group-based cognitive behavioural therapy (GB-CBT) indicated preventiveinterventions for depression are well studied and accepted. However, evaluations of theireffectiveness and cost effectiveness have yielded conflicting and incomplete results mainly dueto methodological limitations.
Therefore, the overarching aim of this thesis was to investigate the effectiveness and cost-effectiveness of GB-CBT indicated preventive interventions for depression in children andadolescents, and estimate the additional healthcare costs and welfare burden in early to mid-adulthood associated with adolescent depression.
In study I, GB-CBT indicated preventive interventions for depression in children andadolescents were protective against developing a depressive disorder and also reduceddepressive symptoms. The intervention relative effect was noted to decay over time. StudyII revealed that adolescent depression was associated with considerable increased healthcareconsumption in mid-adulthood compared to the non-depressed peers at a population level. Thisfinding was more pronounced in females with persistent depressive disorder (PDD). Study IIIshowed adolescent depression to be associated with all the forms of social transfer payments(welfare) especially in individuals with PDD and those with early comorbid psychopathologies.
Study IV demonstrated that GB-CBT indicated preventive interventions for adolescentdepression are not only effective but also have the potential to be cost-effective compared toleaving adolescents with subsyndromal depression unattended.
Findings from studies II and III, emphasized the large financial burden to society in termsof additional healthcare costs and welfare expenditure associated with adolescent depression.Whereas, study I and IV bring forward the message that it is possible to change the course ofsubsyndromal depression by offering GB-CBT indicated preventive interventions to adolescentswith subsyndromal depression. Such an initiative was not only cost-effective compared to notintervening, but also largely cost-saving. Therefore, GB-CBT indicated preventive interventionscould be used as part of a stepped care program linking into more specialized care services. Theresults of this thesis will be useful in decision-making concerning the resource allocation relatedto adoption and implementation of such preventive measures.
Keywords: Health economics; adolescent depression; indicated prevention; direct costs;Markov models; welfare burden
Richard Ssegonja, Department of Public Health and Caring Sciences, Social medicine/CHAP,Box 564, Uppsala University, SE-751 22 UPPSALA, Sweden.
© Richard Ssegonja 2020
ISSN 1651-6206ISBN 978-91-513-0879-1urn:nbn:se:uu:diva-404547 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-404547)
List of Papers
This thesis is based on the following papers, which are referred to in the text by their Roman numerals.
I. Ssegonja, R., Nystrand, C., Feldman, I., Sarkadi, A., Langenskiöld,
S., Jonsson, U., 2018. Indicated preventive interventions for depres-sion in children and adolescents: a meta-analysis and meta-regression. Prev. Med. 118, 7–15.
Ssegonja, R., Alaie, I., Philipson, A., Hagberg, L., Sampaio, F., Möller, M., von Knorring, L., Sarkadi, A., Langenskiöld, S., von Knorring, A.L., Bohman, H., Jonsson, U., Feldman, I., 2019. De-pressive disorders in adolescence, recurrence in early adulthood, and healthcare usage in mid-adulthood: A longitudinal cost-of-illness study. J. Affect. Disord. 258, 33 – 41.
III. Alaie, I., Ssegonja, R., Philipson, A., Hagberg, L., von Knorring, A.L., Möller, M., von Knorring, L., Ramklint, M., Bohman, H., Feldman, I., Jonsson, U. Adolescent depression, early psychiatric comorbidities, and adult welfare burden:
IV. Ssegonja, R., Sampaio, F., Alaie, I., Philipson, A., Hagberg, L.,
Sarkadi, A., Langenskiöld, S., Murray, K., Jonsson, U., Feldman, I. Cost-effectiveness of an indicated preventive intervention for de-pression in adolescents: a model to support decision making. (sub-mitted
Reprints were made with permission from the respective publishers.
Contents
Introduction ................................................................................................ 9
Background ...............................................................................................11Health economics and decision making in healthcare .............................11
Decision making ...............................................................................11Health economics .............................................................................11
Economic evaluation evidence and its use in healthcare .........................12Economic evaluation of mental health interventions ..............................13The spectrum of prevention in mental health .........................................14Depressive disorders in adolescents .......................................................15Effectiveness of indicated preventive interventions for depression in adolescents ............................................................................................17Cost burden of depression in adolescents and later adulthood ................17Cost-effectiveness of indicated preventive GB-CBT interventions for depression .............................................................................................18
Rationale for this thesis .............................................................................20
Overall aim ...............................................................................................21
Specific aims .............................................................................................22Study I ..................................................................................................22Study II .................................................................................................22Study III................................................................................................23Study IV ...............................................................................................23
Methods and results ...................................................................................24Summary of the studies included in this dissertation ..............................25Study I: Indicated preventive interventions for depression in children and adolescents: a meta-analysis and meta-regression ............................26
Aims.................................................................................................26Methods ...........................................................................................26Results..............................................................................................27
Study II: Depressive disorders in adolescence, recurrence in early adulthood, and healthcare usage in mid-adulthood: A longitudinal cost-of-illness study ..............................................................................33
Aims.................................................................................................33
Methods ...........................................................................................33Results..............................................................................................38Conclusions ......................................................................................42
Study III: Adolescent depression, early comorbid psychiatric disorders and adult welfare burden: A 25-year longitudinal cohort study ..............42
Aims.................................................................................................42Methods ...........................................................................................42Results..............................................................................................46Conclusions ......................................................................................51
Study IV: Cost-effectiveness of an indicated preventive intervention for depression in adolescents: a model to support decision-making ........51
Aims.................................................................................................51Methods ...........................................................................................51Results..............................................................................................55Conclusions ......................................................................................62
General Discussion ....................................................................................63
General Conclusions ..................................................................................71
Acknowledgements ...................................................................................73
References .................................................................................................75
Abbreviations
CBA Cost-benefit analysis CBT Cognitive behavioral therapy CCA Cost-consequences analysis CEA Cost-effectiveness analysis CMA Cost-minimization analysis CUA Cost-utility analysis D Depressed DALYs Disability adjusted life years De Dead DRG Diagnosis related group GB-CBT Group based cognitive behavioral therapy GDP Gross domestic product MDD Major depressive disorder PDD Persistent depressive disorder QALY Quality adjusted life year QoL Quality of life R Remission Re Recovered RCT Randomized controlled trial SEK Swedish krona sD Subsyndromal/subthreshold depression WTP Willingness to pay SMD Standardized mean difference RR Relative risk
9
Introduction
Health care expenditure as a percentage of the gross domestic product (GDP) has been increasing over the past decades globally (WHO, 2017). This is partly due to the increase in disease burden, and treatment alternatives on the market but yet a disproportionate increase in healthcare resources (Kernick, 2003). Thus, governments’ and health systems’ decision-makers are facing an ongoing challenge of achieving efficiency in resource allocation within the healthcare sector. The increased application of economics principles to decision-making regarding health and healthcare provision, that is, health economics, has aimed to improve efficiency within the healthcare sector. The purpose of health economics is to produce evidence and support deci-sion-making to achieve good value for the investments made within the healthcare sector (M. Drummond, Sculpher, Torrance, O'Brien, & Stoddart, 2005). However, this kind of health economics evidence and its use is still limited within the mental health sphere, especially pertaining to children and adoles-cents (Skokauskas et al., 2018; Wykes et al., 2015). Recent research reports mental health problems, such as depression (depressive disorders) in chil-dren and adolescents, to be on the rise (Avenevoli, Swendsen, He, Burstein, & Merikangas, 2015; Del Vecchio, 2018), and to contribute considerably to the overall disease burden (Kyu et al., 2016; Reiner et al., 2019). In addition, there is documentation of the high cost burden of depressive disorders in adolescents (Bodden, Stikkelbroek, & Dirksen, 2018) and adults in general (Coretti, Rumi, & Cicchetti, 2019; Konig, Konig, & Konnopka, 2019; M. Luppa, Heinrich, Angermeyer, Konig, & Riedel-Heller, 2007; Olesen, Gustavsson, Svensson, Wittchen, & Jönsson, 2012; Sobocki, Jonsson, Angst, & Rehnberg, 2006) both in the short and long-term. The cost burden spans across both healthcare usage, labour market performances and other societal sectors including welfare systems. One way to withstand this negative trend is to implement effective and cost-effective preventive measures (Cuijpers, Beekman, & Reynolds, 2012; Del Vecchio, 2018). There is a growing body of research on the prevention of depression in adolescents (Rasing, Creemers, Janssens, & Scholte, 2017; Stockings et al., 2016). Preventing depression in adolescents is needed given its associated long-term negative consequences, including decreased produc-
10
tivity, suicide, increased all-cause mortality, reduced quality of life, academ-ic and financial under achievement (Clayborne, Varin, & Colman, 2019; Cuijpers et al., 2014; Naicker, Galambos, Zeng, Senthilselvan, & Colman, 2013; OECD, 2010; Wulsin, Vaillant, & Wells, 1999) and welfare depend-ency (Bardone, Moffitt, Caspi, Dickson, & Silva, 1996; McLeod, Horwood, & Fergusson, 2016). There is both effectiveness (Hetrick, Cox, & Merry, 2015; Rasing et al., 2017; Stockings et al., 2016; Yang et al., 2017) and some cost-effectiveness evidence (Lee et al., 2017; C. Mihalopoulos & Vos, 2013; Cathrine Mihalopoulos, Vos, Pirkis, Smit, & Carter, 2011) asserting that indicated preventive interventions for depression in adolescents are pre-ferred compared to a do-nothing alternative. This is especially true for cogni-tive behavioral therapy (CBT) delivered in a group setting in routine care or schools. Though, the vast evidence is conflicting and leaves several knowledge gaps yet to be addressed regarding the effectiveness and cost-effectiveness of indicated preventive group-based cognitive behavioral therapy (GB-CBT) interventions for depression (Brent et al., 2015; Corrieri et al., 2013; Holmes et al.; Khan, Faucett, Lichtenberg, Kirsch, & Brown, 2012; Merry & Spence, 2007; Rasing et al., 2017). The evidence on effectiveness of such interven-tions when compared to other forms of prevention rather than a do-nothing alternative is yet to be explored. The actual cost burden and the social finan-cial assistance (welfare) burden in form of disability pension, sickness ab-sence (sick-leave), unemployment and public assistance due to depression (including subsyndromal disease) compared to non-depressed individuals, has not received much attention in the costing studies on adolescent depres-sion. Not having dependable costs and effectiveness estimates limits the evaluations of the cost-effectiveness of these interventions in the long-term using modelling techniques. The studies included in this thesis attempt to address the current knowledge gaps by investigating the effectiveness and cost-effectiveness of indicated preventive GB-CBT interventions for depres-sion in children and adolescents, and estimating the additional healthcare costs and welfare burden in early to mid-adulthood associated with adoles-cent depression.
11
Background
Health economics and decision making in healthcare Decision making Decision making occurs at different levels and in different fields every day. However, it is a complex field especially when the consequences of the deci-sions undertaken impact a part of, or the entire population. It is difficult pri-marily because of the benefits forgone by choosing one alternative over an-other (opportunity cost) amidst the uncertainties involved. Therefore, deci-sion making has to be supported by a lot of background evidence to help reduce the level of uncertainty, and by the theory or ideology of what consti-tutes the best value for the individual/people. Quantitative approaches to produce evidence to aid decision making are increasingly being adopted (Claassen, Hendriks, & Hendrix, 2007). The generated evidence has to be benchmarked to other factors like equity, justice, political atmosphere, feasi-bility and resource availability in form of a network to inform the decision task at hand. At a societal level, this applies to all publically funded sectors like transport, education, national security, agriculture, and, in some coun-tries, the health sector. Thus, evidence generated by health economics work and its use to inform decision-making in healthcare is increasingly becoming important (EUnetHTA, 2015; NICE, 2013a; Utvärdering, 2013).
Health economics Health economics is a discipline concerned with the application of econom-ics principles to health and healthcare (Morris, Devlin, & Spencer, 2012). Healthcare provision involves goods and services intended to promote health, and prevent, alleviate or eliminate ill-health (Culyer, 2008). It in-volves a complex interplay between the consumers (patients), suppliers (healthcare system) and buyers (third party payers, government, insurance companies or patients). Therefore, health economics deals with the issues related to the efficiency, effectiveness, value, and behavior in the production and consumption of health and healthcare among others (M. Drummond et al., 2005).
12
Health economics work encompasses several domains: description, explana-tion (justification), evaluation and prediction. Descriptive work in health economics helps to paint a picture of, for example, the cost or health out-comes related to ill health or health events. Work exploring questions that seek justifications or explanations aims to address the occurrence of different health events, influences of certain economic factors on health, insurance aspects, patients’ behavior or trends in healthcare utilization. Evaluation work seeks to answer questions of whether the benefits of adopting an inter-vention relative to another are worth the investment. To complement this, prediction work helps to show or ascertain the future health and economic impact of undertaking a certain health intervention or making a decision, through the use of economic and statistical modelling. The work of this dis-sertation falls under the domains of description, evaluation and prediction within health economics.
Economic evaluation evidence and its use in healthcare The aim of economic evaluation is to support decision-making by creating evidence on the cost of the marginal benefits of undertaking a new approach compared to its alternative (opportunity cost) (M. Drummond et al., 2005). Within most societal sectors, economic evaluations have been used in deci-sion-making for a long time, while its application within healthcare has come about rather late. The foundational activities and pillars within eco-nomic evaluation include costing of resources and estimation of health bene-fits. The costing process includes activities such as identification, quantifica-tion (estimating frequencies) and valuation (attaching unit costs) of resource use. Health benefits can be expressed as solid clinical outcomes, such as depression symptom free days, cases of depression prevented, or composite generic measures like quality adjusted life years (QALYs) and disability adjusted life years (DALYs). Differences in costs and health benefits for the different alternatives being compared are then combined to estimate an in-cremental cost-effectiveness ratio (ICER). The ICER is judged based on the willingness to pay value (WTP), determined by society, for a unit of health benefit to ascertain if it is cost-effective. However, the ultimate decision to adopt a given intervention also depends on other factors, such as disease severity, human dignity and solidarity principles as well as equity considera-tions. Broadly, economic evaluations can be termed as partial or full evaluations. Full evaluations are preferred as they present comprehensive information on both costs and health outcomes. The common forms of full economic eval-uations include cost-benefit analysis (CBA), cost-utility analysis (CUA), cost minimization analysis (CMA), cost-effectiveness analysis (CEA) and
13
cost-consequences analysis (CCA) (M. Drummond et al., 2005; M. F. Drummond, Sculpher, Claxton, Stoddart, & Torrance, 2015; Julia & John, 2005; Neumann et al., 2017). The different forms of economic evaluations differ in the unit of measurement of health outcomes but all value costs in monetary terms. The cost-utility analysis approach considers composite out-comes combining both morbidity and mortality. It is the commonly recom-mended approach by many healthcare technology assessment agencies like the Swedish dental and pharmaceutical benefits agency and the National institute for health and care excellence in the United Kingdom. The CUA approach is recommended since it considers a comprehensive outcome, that is, both the clinical and non-clinical benefits of the interventions. It also pre-sents an opportunity to be able to compare cost effectiveness results across different disease groups for planning purposes.
Economic evaluation of mental health interventions In spite of the growing body of knowledge pointing to the increasing disease (IHME, 2015; Kyu et al., 2016; Reiner et al., 2019) and cost burden of men-tal health problems and disorders (Andlin-Sobocki, Jonsson, Wittchen, & Olesen, 2005; Coretti et al., 2019; Gustavsson et al., 2011; IHME, 2015; Kessler et al., 2009; Konig et al., 2019; M. Luppa et al., 2007; Olesen et al., 2012), the use of economic evidence in mental healthcare is still limited (McDaid, Park, & Wahlbeck, 2019; Skokauskas et al., 2018; Zechmeister, Kilian, McDaid, & group, 2008). In turn, there is an influx of treatment and preventive strategies delivered via different platforms entering the market (Farah et al., 2016; Holmes et al.; Khan et al., 2012; Linde et al., 2015; Rasing et al., 2017; Stockings et al., 2016; Tindall et al., 2017). Of these strategies, economic evaluations of new drug therapies (pharmacotherapy) are the most common (Karyotaki, Tordrup, Buntrock, Bertollini, & Cuijpers, 2017). Evaluations of early treatment and preventive approaches have not received much attention by health economists, thus the literature on their cost-effectiveness is limited (Brettschneider et al., 2015; Karyotaki et al., 2017; Lee et al., 2017; C. Mihalopoulos & Vos, 2013; C. Mihalopoulos, Vos, Pirkis, & Carter, 2012). The above studies are both reviews of econom-ic evaluations as well as model-based evaluations that report early preven-tive interventions for depression to be cost-effective. However, the model-ling studies possess shortcomings (Lee et al., 2017) especially regarding the presentation of the natural history of depression, hence the estimates report-ed maybe incomplete. The review studies also note the paucity of economic evidence on the cost-effectiveness of preventive strategies for depression in general and thus advocate for further studies to be done.
14
The spectrum of prevention in mental health Preventive interventions are measures aimed at hindering the occurrence or progression of a health event aimed at populations with various levels of risk. Prevention can be primary, secondary or tertiary. Primary prevention pertains to hindering the initial occurrence of a health event, whereas sec-ondary prevention entails early treatment of the health event in order to re-store health and hinder the occurrence of complications. At a tertiary level, that is, in an instance where a health event and the associated complications have occurred, it is important to help rehabilitate the patient to slow down the progress of disease and restore functioning (Clarke, 1974; Culyer, 2008). In public mental health, primary prevention is preferred, as demonstrated by the predicted future benefits and the avoided negative consequences of poor mental health (D. Chisholm et al., 2016; Institute of Medicine Committee on Prevention of Mental, 1994; McDaid et al., 2019; Zechmeister et al., 2008). Primary preventive interventions in public health can be universal or target-ed. Universal interventions are those delivered to the whole population irre-spective of underlying level of risk factors for the health event of interest (Institute of Medicine Committee on Prevention of Mental, 1994). These interventions have received critics for having low effects and thus targeted approaches are preferred as they focus on a population with an increased risk or symptoms. Targeted preventive interventions include selective and indicated interven-tions. Selective interventions target a population with a heightened risk to develop a health event but with no symptoms yet. Indicated interventions target a population with symptoms of a given health event, but they are ei-ther too few or not severe enough to warrant a diagnosis i.e. they are sub-threshold, in other words subsyndromal (Institute of Medicine Committee on Prevention of Mental, 1994). Indicated preventive interventions have demonstrated to be highly effective relative to the other forms of primary prevention, and thus increasingly adopted in the prevention of mental health problems and disorders, such as depression in children and adolescents (Hetrick et al., 2015; Rasing et al., 2017; Stockings et al., 2016).
15
Figure 1. The continuum of prevention and treatment of mental disorders (Institute of Medicine Committee on Prevention of Mental, 1994)
Depressive disorders in adolescents Depression is a mental health disorder characterized by persistent low or irritable mood and loss of interest in pleasurable activities (APA, 2013). In children and adolescents, depression may present with only low or irritable mood. A diagnosis of depression is symptom criteria based. The statistical manual for mental health disorders (DSM-5) classifies depressive disorders as: disruptive mood dysregulation disorder, major depressive disorder (MDD) (including major depressive episode), persistent depressive disorder(including dysthymia), premenstrual dysphoric disorder, sub-stance/medication-induced depressive disorder, depressive disorder second-ary to another medical condition, other specified depressive disorder, andunspecified depressive disorder. The etiology of depression is hypothesizedto be a multifactorial problem, with theories on attachment, trauma, geneticpredisposition and inadequate levels or dysregulation in the monoamineneurotransmitters i.e. serotonin, norepinephrine and Dopamine in the brain(Fauci & Harrison, 2008).
In this thesis, unipolar depression was the subtype of interest. Depression is one of the diseases which contributes the most to the overall burden of dis-ease even in the children and adolescents population (Kyu et al., 2016; Reiner et al., 2019). This contribution is predicted to be even higher by 2030
16
(Mathers & Loncar, 2006). Depression is a common disorder in children and adolescents with the global point prevalence estimated as 5-6 % (Costello, Erkanli, & Angold, 2006; Erskine et al., 2017) with a female to male ratio of 2:1 (Bromet et al., 2011). Present research knowledge supports an increase in the prevalence of depression from 5% in the early teens to about 20% in late teenage years (Lewinsohn, Rohde, Klein, & Seeley, 1999; Merikangas et al., 2010), thus it is not surprising that a considerable proportion of the popu-lation report their first episode of a depressive disorder during adolescence (Kessler et al., 2005; Thapar, Collishaw, Pine, & Thapar, 2012). In this re-gard, Sweden is no exception. In a community sample of adolescents in Uppsala, the estimated lifetime prevalence was as high as 11.4% (Olsson & von Knorring, 1999). Therefore, depression in children and adolescents is a public health problem. It is associated with long-term disability, poor general health, decreased quality of life, and increased risk of all-cause mortality including suicide (Creed et al.; Cuijpers et al., 2014; Osby, Brandt, Correia, Ekbom, & Sparen, 2001; Thapar et al., 2012). There is also evidence linking depression to re-duced overall productivity, psychosocial functioning and achievement in life (Clayborne et al., 2019). Reduced productivity in form of productivity losses due to absenteeism from work or school and presenteeism amount to a sig-nificant proportion of the overall indirect costs of depression (Bodden et al., 2018). Further, depressed adolescents are also prone to other psychiatric illnesses, including substance abuse, violence and anxiety disorders (Jonsson, 2010), and have an increased dependency on social financial assis-tance (Bardone et al., 1996; Fazel et al., 2015; McLeod et al., 2016). Generally, there seems to be a possibility that the health consequences, cost and social financial assistance related burden of depressive disorders have been underestimated. This is because sometimes the subsyndromal disease group is rarely captured in descriptive or explanatory studies and no compar-ison to non-depressed peers is attempted. There are several management (prevention and treatment) approaches to depression, including pharmacological and psychotherapeutic approaches recommended by various international authorities, such as the National Health Services (NHS) in the United Kingdom and the Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU) (NICE, 2013b; Swedish Council on Health Technology, 2004). Prevention of depression using indicated preventive interventions is a priority in view of the future negative consequences. Importantly, there is still a need to assess the effectiveness and cost-effectiveness of such interventions (Bertha & Balazs, 2013; Isobel M. Cameron, Kenneth Lawton, & Ian C. Reid, 2011).
17
Effectiveness of indicated preventive interventions for depression in adolescents There are several indicated preventive interventions with different theoreti-cal underpinnings that target depression in adolescents. Of these back-grounds, there are extensive studies on the effectiveness and acceptability of interventions based on CBT principles (NICE, 2013b; Swedish Council on Health Technology, 2004). Existing studies present effectiveness estimates of preventive CBT interventions as relative risks ranging from 0.29 to 0.78 of developing a depressive disorder compared to controls (Horowitz & Garber, 2006; Rasing et al., 2017; Stice, Shaw, Bohon, Marti, & Rohde, 2009; Stockings et al., 2016). The CBT interventions evaluated in the differ-ent studies differ in some structural aspects, however they are usually based on similar principles. Therefore, these interventions conform to a general format that can be adapted to a given particular setting. Research reports that indicated preventive group-based CBT interventions (GB-CBT) are relatively effective to prevent depressive disorders (Stockings et al., 2016), although the results are conflicting (Brent et al., 2015; Corrieri et al., 2013; Holmes et al.; Merry & Spence, 2007; Rasing et al., 2017). Ad-ditionally, the results do not distinguish between the comparisons of GB - CBT to active or passive controls, which presents a challenge when drawing clinically relevant conclusions, as results may be over- or underestimated in relation to current practice. The existing studies also do not distinguish or clarify the impact of additional booster sessions, which may have both clini-cal (Gearing, Schwalbe, Lee, & Hoagwood, 2013) and economic conse-quences.
Cost burden of depression in adolescents and later adulthood Depression impinges a large cost on society, both in the short- and long-term (Andlin-Sobocki et al., 2005; Coretti et al., 2019; Gustavsson et al., 2011; Hu, 2006; Kind & Sorensen, 1993; M. Luppa et al., 2007; Melanie Luppa, Heinrich, Angermeyer, König, & Riedel-Heller, 2003; Olesen et al., 2012; Sobocki, Ekman, et al., 2007; Sobocki et al., 2006; Sobocki, Lekander, Borgstrom, Strom, & Runeson, 2007; von Knorring, Åkerblad, Bengtsson, Carlsson, & Ekselius, 2006). These costs include direct health care costs like inpatient care, outpatient care visits and medication as well as indirect costs like productivity losses. Indirect costs in the form of productivity losses con-tribute the most to this cost (Kind & Sorensen, 1993; Sobocki et al., 2006).
18
A similar large cost burden is also demonstrated in adolescent depression (Bodden et al., 2018), although there is still a paucity of literature regarding cost burden of depression in adolescents. Recent research emphasizes that as the negative consequences of adolescent depression are vast and costly, pre-vention should be highly prioritized (Cuijpers et al., 2012). This is also sup-ported by economic evidence on preventive interventions showing promising returns on investment for preventing future depressive disorders in a scenar-io analysis. (Dan Chisholm et al., 2016). Considering that the young popula-tion is the future productive work force, and the documented association between depression in adolescence and an increased risk of depression in adulthood (Copeland et al., 2013) as well as poor psychosocial functioning (Clayborne, Varin, & Colman), it may be advisable to target adolescents when offering preventive strategies, especially as they transit to adulthood. This is because there is a significant proportion of individuals reporting their first episode of depression during this period (Kessler et al., 2005) and this is a period of vast growth as well as acquiring marketable skills that necessary in the workforce later in life. Notably, much of the costing studies of depression pertain to adult popula-tions and group together different forms of depressive disorders, rarely in-cluding subsyndromal disease (Andlin Sobocki & Wittchen, 2005; Chisholm et al., 2003; Löthgren, 2004) and yet subsyndromal disease has been noted to be disabling to a significant extent (I. M. Cameron, K. Lawton, & I. C. Reid, 2011). Thereby they may underestimate or overestimate the cost burden of direct health care utilization. This in turn would make the cost-effectiveness evidence of indicated preventive GB-CBT interventions based on these cost and effectiveness estimates incomplete and worth revisiting.
Cost-effectiveness of indicated preventive GB-CBT interventions for depression There is a handful of cost-effectiveness evaluations for indicated preventive GB-CBT interventions. The majority of the studies report that indicated in-terventions are cost effective compared to a do-nothing alternative (Karyotaki et al., 2017; Lee et al., 2017; Lynch et al., 2005; C. Mihalopoulos & Chatterton, 2015; C. Mihalopoulos & Vos, 2013; C. Mihalopoulos et al., 2012; Cathrine Mihalopoulos et al., 2011). These studies, however, present less evidence about the long-term consequences of undertaking such inter-ventions, and could be using inflated or underestimated effect and cost esti-mates. The studies are also mainly population based models (Lee et al., 2017; C. Mihalopoulos et al., 2012) focusing on the states, well, diseased or dead, thus, may not fully represent the natural history of depression. There-
19
fore, more predictive modelling-based studies or studies with different strat-egies to capture the natural history of depression are necessary.
20
Rationale for this thesis
The existing literature on preventive medicine demonstrates the increasing advocacy for and adoption of preventive interventions in mental health ser-vices (Del Vecchio, 2018; McDaid et al., 2019; Zechmeister et al., 2008). However, the evidence base for these decisions needs to be further investi-gated as the effectiveness results of indicated preventive GB-CBT interven-tions for depression are conflicting and their cost-effectiveness in the long-term remains deficient. The current studies don’t capture the full impacts of depression on healthcare and other societal sectors, neither do they capture subthreshold cases (Bertha & Balazs, 2013). This thesis contributes to the body of knowledge in the research community for both decision makers and researchers. It informs on the differences in effectiveness estimates based on the nature of different comparators, and updates information on the cost and social financial assistance burden due to depression (including subthreshold disease) in depressed adolescents compared to their non-depressed counter-parts as they transit into adulthood. Further, it also presents the cost-effectiveness of an average indicated preventive GB-CBT intervention, using economic modelling techniques. Decision makers will be able to use the results of these studies to support decision-making and priority setting in the appropriate allocation of resources pertaining to the funding of such preven-tive strategies.
21
Overall aim
The overarching aim of this work was to investigate the effectiveness and cost-effectiveness of indicated preventive GB-CBT interventions for depres-sion in children and adolescents, and estimate the additional healthcare costs and welfare burden in early to mid-adulthood associated with adolescent depression.
22
Specific aims
Study I The aim of this study was to synthesize evidence on GB-CBT indicated pre-ventive interventions for depression in children and adolescents with a par-ticular focus on:
i) Addressing the inconsistencies in effectiveness of GB-CBT in-dicated interventions for the prevention of depression in children and adolescents
ii) The effectiveness of GB-CBT indicated preventive interventions in relation to an active or passive comparator and the impact of booster sessions on intervention effectiveness
iii) Synthesizing and reporting effect sizes in a useful form for use in cost-effectiveness assessments and decision-analytic model-ling of these interventions.
Study II The aim of study II was to investigate the association of adolescent depres-sion with subsequent healthcare utilization and costs in mid-adulthood, ad-dressing the following concerns:
i) To estimate additional cumulative healthcare utilization and costs (psychiatric and somatic) in mid-adulthood (ages 31–40 years) in females and males with adolescent depression (persis-tent depressive disorder (PDD), episodic major depressive dis-order (MDD), subthreshold depression(sD), compared to indi-viduals without a history of depression in adolescence
ii) To ascertain to which extent any additional healthcare utilization and costs (ages 31–40 years) in males and females with a history of adolescent depression are mediated by depressive episodes occurring in early adulthood (ages 19–30 years).
23
Study III The aim of study III was to investigate the association between adolescent depression and government transfer payments (social financial assistance or welfare) related to unemployment, work disability, and public assistance across early to middle adulthood (20 - 40 years). The following specific aims were examined:
i) To estimate the magnitude of the additional government transfer payments associated with the adolescent depression (PDD, epi-sodic MDD, subthreshold depression) compared to no depres-sion in early to mid-adulthood (age 20 to 40 years)
ii) To ascertain whether such associations if any, were impacted by comorbid psychopathology occurring in childhood or during ad-olescence.
Study IV Study IV aimed to assess whether an indicated GB-CBT preventive interven-tion for depression offered to adolescents with subsyndromal depression provides good value for money in the long-term, compared to a do-nothing alternative, using modelling techniques.
24
Methods and results
The studies used a range of research methods, from evidence synthesis and register based comparative analysis, to economic evaluation and modelling techniques, in order to address the aims of the work. Study I used published literature to collect outcome data and hence calculate the weighted effective-ness estimate of indicated preventive GB-CBT interventions for depression in children and adolescents, using a meta-analytic approach. Study II and III, were register based comparative analyses of the additional healthcare cost and social financial assistance burden of depression in adolescents as they transit into adulthood compared to their non-depressed peers. The register data from a previously conducted community-based cohort study of adoles-cent depression in Uppsala, Sweden (1991/2) (Olsson & von Knorring, 1999), were complemented by national registers (Alaie et al., 2019), thus we could follow the cohort up to adulthood. In study IV, we used a single cohort Markov model to conduct an economic evaluation comparing an indi-cated GB-CBT preventive intervention for depression offered to adolescents with subsyndromal depression to a no-intervention option.
26
Study I: Indicated preventive interventions for depression in children and adolescents: a meta-analysis and meta-regression Aims To synthesize evidence on GB-CBT indicated preventive interventions for depression in children and adolescents with particular focus on: addressing the inconsistencies in reported intervention effectiveness, the role of the nature of the comparator, booster sessions and reporting the effect sizes in a useful form for inputs in cost-effectiveness assessments and decision-analytic modelling.
Methods This study was a systematic literature review with a meta-analysis based on the Cochrane Handbook for Systematic Reviews of Interventions (Higgins & Green, 2011) and the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement guidelines (Moher, Liberati, Tetzlaff, & Altman, 2009). The work stemmed from meta-analyses conducted by Stock-ings et al. (Stockings et al., 2016), and Rasing et al. (Rasing et al., 2017). These studies were deemed to be of good quality, according to the Assessing the Methodological Quality of Systematic Reviews (AMSTAR) framework (Shea et al., 2007). The study analyses included randomized controlled trials (RCT) of indicated preventive GB-CBT interventions for depression, offered to children and adolescents aged 12-19, reporting cases of depressive disorder as the primary outcome. The studies were retrieved from an electronic literature search in PubMed, Web of Science, PsycINFO, Medline and the Cochrane Library of Systematic Reviews using a combination of search terms including adoles-cents, depression, preventive interventions and cognitive behavioral therapy, for the period 01/09/2014 to 28/02/2018 to complement the two recent re-views on the topic. Two researchers independently assessed the eligibility of the articles, extracted and coded the relevant data. Disagreements were re-solved through consulting with a third party. An agreement statistic (Cohen’s Kappa) was calculated to examine the extent of agreement in the selection of articles (Higgins & Green, 2011). The eligible articles were assessed for bias using The Cochrane Collaboration’s tool for assessing risk of bias (Higgins & Green, 2011). The assessment included bias within and between studies. The total bias scores were used in a meta-regression as an explanatory varia-ble to assess its impact, alongside the occurrence of booster sessions, on the intervention effectiveness.
27
A random effects model (REM) (Borenstein, Hedges, Higgins, & Rothstein, 2009) was used to estimate the pooled intervention effectiveness, expressed as a relative risk or a Cohen’s d. The REM was chosen as it incorporates both the within-study and between-study variance in the weighting process of the studies. The Cochrane’s Q statistic and I2 were estimated as measures of heterogeneity across the included studies. The pooled effectiveness results were estimated with 95% confidence intervals and illustrated using forest-plots. The calculations were performed using the “metafor” package in R, version 3.4.3.
Results
Study selection From all the electronic and hand search of existing literature, a total of 38 articles with 49 unique comparisons from 34 trials were eligible for inclusion in this work as illustrated in figure 2. The measure of agreement (kappa statistic) in the selection of studies was 0.84.
28
Figure 2. Flow chart showing the selection of studies Study featuresOverall, almost half of the studies (45%) were carried out in the US. The populations studied were children and adolescents with a mean age of 14 years. The interventions were delivered by a wide range of professionals e.g. psychologists, therapists, counsellors, nurses, teachers and therapists. The average number of participants per group was 8 with an average attendance of 6.5 out of 9 sessions with each session lasting on average 70 min. Deliv-ery of each group intervention required 1–2 facilitators who received super-vision weekly. The intervention duration was on average 9 weeks with fol-low-up times of up to 13 months. The comparator was mostly passive but a few studies compared GB-CBT interventions to active comparators. Eight trials included active comparators designed to control for non-specific aspects of psychological treatment.
29
These active comparators included a group intervention focusing on stressors associated with adolescent depression, an individual or group support pro-gram (one educative interview), a supportive-expressive group intervention, a group-based attention control intervention, a structured activity group, and an expressive writing/journaling intervention. Three trials included com-pared GB-CBT with CBT delivered as either bibliotherapy or as an interac-tive fantasy game as a control condition. Finally, one trial used the comple-mentary treatment Reiki as comparator. The interventions were delivered in schools (99%) and health centers (1%). The facilitators received training of a mean duration of 2.65 working days (21.24 h) before delivering the intervention. Supervision to the facilitators of the interventions was done face-to-face 87% of the time, the remaining done via phone calls. Supervision lasted on average one hour per week. Only 16% of the studies had included booster sessions with the interventions. A majority of the studies used the Children's Depression Inventory (CDI) (Kovacs, 1985), Becks Depression Inventory II (BDI II), The Centre for Epidemiological studies on Depression tool (CESD), and the Schedule for Affective disorder and Schizophrenia for School Age Children tool (K-SADS) to screen. In the reviewed studies, a case of depression was defined as a diagnosis of MDD or/and dysthymia using a diagnostic interview, and/or a score on a symptoms scale indicative of severe depression.
Risk of bias in individual studies and across studies There was always a possibility of performance bias due to the difficulty of blinding the participants and facilitators given the nature of the interventions. However, other forms of bias, especially selection bias and reporting bias, were unclear, as per the subjective scoring system in this work. There was also some evidence of risk of publication bias as shown by the asymmetry noted in the funnel plots. Studies with small samples sizes and negative out-comes (results showing the intervention as being ineffective) were less rep-resented.
Intervention effectiveness Results from the random effects model showed GB-CBT indicated preven-tive interventions to significantly reduce the incidence of depressive disorder at all follow-up time points, see table 1. Intervention effectiveness was noted to decay between end of the intervention and 12-month follow-up, and seems to decay more beyond that period. Reducing the sample to studies with a passive comparator maintained the intervention effectiveness at 6 and 12 months after the end of the intervention, whereas when comparing to active comparators, the results became non-significant for the first six months, but significant at twelve months or more. The same pattern of find-
30
ings was observed when considering standardized mean differences as the outcome.
33
Additional analyses Meta-regression analyses showed significant results for the secondary out-come (Cohen’s d or SMD) but not the primary outcome (cases of depression prevented). The results of the exploration of the number needed to treat to prevent a case of depressive disorder in adolescents with subthreshold de-pression at the different time points ranged from nine to forty-two.
Conclusions GB-CBT indicated preventive interventions for depression in children and adolescents are effective, but the effect seems to decay over time. Therefore, there is a need to further explore the benefit of booster sessions and the ap-propriate timing for them to be offered. Our results also emphasize the value of making a clear distinction between active and passive comparators in order to guide decision-making.
Study II: Depressive disorders in adolescence, recurrence in early adulthood, and healthcare usage in mid-adulthood: A longitudinal cost-of-illness study Aims To investigate the association between adolescent depression and subsequent healthcare consumption and related costs in mid-adulthood, with the interest to answer the following questions:
1. What is the estimated cumulative healthcare utilization and costs (psychiatric and somatic) in mid-adulthood (ages 31–40 years) in females and males with clinical subtypes of depressive disorders in adolescence (PDD, episodic MDD, subthreshold depression), com-pared to individuals without a history of depression in adolescence?
2. To what extent is any additional healthcare utilization and costs (ag-es 31–40 years) in males and females with a history of adolescent depression mediated by depressive episodes occurring in early adulthood (ages 19–30 years)?
Methods This was a register-based comparative study whereby data pertaining to a cohort of depressed and non-depressed participants screened in 1990-92 were followed up using national registers. The study received ethical ap-proval from the Regional Ethical Review Board in Uppsala (2015/449/1-2).
34
Study sample and procedure The initial epidemiological investigation was done in Uppsala, Sweden, where all first-year high school students (16-17 years) (n=2300) were screened for depression. The depression screening was done using two self-reported questionnaires, the Beck Depression Inventory (BDI-C) (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961; Olsson & von Knorring, 1997a) and the Centre for Epidemiological Studies-Depression scale for Children (CESDC) (Olsson & von Knorring, 1997b; Schoenbach, Kaplan, Grimson, & Wagner, 1982). Adolescents (n=355) meeting the clinical cut-offs (BDI-C≥16 or CESDC ≥ 30+BDI-C≥11), or reporting a prior suicide attempt, were invited for an in-person diagnostic assessment. For every screen-positive, a non-depressed peer (n=355) matched for sex, age and school year was invit-ed to an identical assessment, and thus increasing representativeness of the study sample to the general adolescent population. The in-person assessment included the Diagnostic Interview for Children and Adolescents in the re-vised form according to DSM-III-R for adolescents (DICA-R-A) (W. Reich, B. Herjanic, Z. Welner, & P. R. Gandhy, 1982). A total of 631 of the 710 invited adolescents participated in these in-person baseline assessments. A second wave of data collection included an assessment of depressive epi-sodes in early adulthood (age 19–30) using the Mini International Neuropsy-chiatric Interview 5.0.0 (MINI-PLUS) in conjunction with a life-chart. Par-ticipants from the baseline assessment who consented to be contacted for a follow-up were invited (n=609), of which 409 participated. Against this background, the present study included 539 participants who took part in the baseline assessment and were eligible for the third, register-based, wave of data collection as illustrated in figure 3. Individuals who did not provide informed consent to be followed up, or who explicitly refused extraction of register-based data for future research, were excluded (n=52). Consistent with the diagnostic criteria for depressive disorders, individuals with a histo-ry of manic or hypomanic episode in adolescence were excluded (n=40). The description of the initial epidemiological study, waves of data collection and preliminary results are described in detail elsewhere (Alaie et al., 2019).
Variables To study the relationship between adolescent depression and future healthcare consumption in early to mid-adulthood, we defined the variables as follows: Adolescent depression (age 16–17): the participants were recategorized into the clinical subtypes as either having a persistent depressive disorder (PDD) n= 175, episodic major depressive disorder (MDE) n=82, subthreshold de-pression (sD) n= 64 or no depression n=213, according to the current Diag-
35
nostic and Statistical Manual of Mental Disorders 5th ed. (DSM-5) criteria for depressive disorders (APA, 2013), to the original DICA-R-A diagnoses. Depression in early adulthood (ages 19–30): Depressive episodes in early adulthood (ages 19–30) were assessed at follow-up using the MINI-PLUS, both in participants with a history of depression in adolescence and those without such history. In order to create a dichotomized mediator variable of clinical relevance, depression in early adulthood was defined as either hav-ing a long episode (i.e., ≥6 months) or recurrent episodes (i.e., ≥2) between ages 19 and 30.
Covariates From the Swedish National Patient Register, Longitudinal Integration Data-base for Health Insurance and Labor Market Studies (LISA) and Statistics Sweden, we were able to access information about the participants’ parents’ mental illness history (dichotomized in our analysis), education level, and disposable income at baseline. Also, from the baseline data, we obtained information about whether the participants had a disruptive behaviour disor-der (dichotomized) e.g. conduct disorder, oppositional-defiant disorder, and/or attention deficit/hyperactivity disorder according to DICA-R-A or/and anxiety disorder (dichotomized) e.g. anxiety disorder, overanxious disorder, and/or avoidant disorder according to DICA-R-A.
Healthcare costs In this work, we included cost information for inpatient care, specialized outpatient care and medication costs. The cost information was retrieved from the Swedish National Patient Register for the care consumption and the Swedish Prescribed Drug Register for the medication consumption. The Swedish National Patient Register contains information on care visits, ad-missions, International Classification of Disease (ICD) code for the health condition and Diagnosis Related Group (DRG) code. Therefore, the inpatient and outpatient care resource use was collected using a bottom-up approach, and the related costs were estimated based on yearly DRG weights (provided by the Swedish National Board of Health and Welfare) multiplied by their average unit cost per DRG weight (provided by the Swedish Association of Local Authorities and Regions) for the period 2005 to 2015 (where all the register data overlapped). Prescribed medication costs were retrieved directly from the Swedish Pre-scribed Drug Register. The registry provides information about the Anatom-ical Therapeutic Classification (ATC) code of each drug, amount dispensed, dosage, total cost, and date of dispensing. Based on ATC codes, the medica-
36
tion was further classified into anti-depressants (N06A), other-psychotropics (N05-N07, excluding N06A) and non-psychotropic (all other codes). All the unit costs were collected in Swedish krona and adjusted to 2015 pric-es in USD using purchasing power parities and inflation indices (Shemilt, Thomas, & Morciano, 2010).
Statistical analyses The results for the different cost and frequency variables (i.e., number of outpatient visits, hospital admissions and hospital length of stay) were esti-mated over time for the matched sample of depressed and non-depressed groups using propensity scores. The groups were defined in accordance to the baseline categorizations. The differences in the mean number of hospital admissions, outpatient visits, length of hospital stay, prescribed medication and related costs between the individuals with and without a history of ado-lescent depression were assessed using generalized linear modelling analysis techniques and appropriate distributions, taking the cumulative costs over time for the period 2005–2015. To improve precision, we controlled for the observed explanatory variables including major psychiatric comorbidities at baseline across all analyses. Given the differences in healthcare seeking hab-its (Thompson et al., 2016) and sex specific conditions (e.g., pregnancy), separate analyses were conducted for males and females, and the results were therefore primarily presented by sex. The study also examined the ICD code diagnostic groups for which the par-ticipants had consumed healthcare, and assessed whether there were differ-ences using Mann–Whitney tests with a Bonferroni correction on the p-value. Following a notice of significant differences between a subtype of adolescent depression and healthcare costs in mid-adulthood, a mediation analysis was undertaken to ascertain whether recurrent depression in early adulthood, ages 19–30 years, may drive the healthcare consumption differ-ences. This was further tested in a sensitivity analysis with imputed data for the variable on depression status at the age 19-30 years (mediator variable). All analyses were done in R version 3.5.1, with the aid of packages “MatchIt” (Ho, Imai, King, & Stuart, 2007) and “cobalt” (Greifer, 2018). Statistical significance was 2-tailed and set at p < 0.05. To obtain a population level understanding of the magnitude of the con-sumption difference, the analysis results for the difference in total cost over the ten years (2005 to 2015) were multiplied by the proportion of the popula-tion aged 16 years and by the prevalence of depression (PDD), then divided this by 10 to get the annual equivalent.
37
Figure 3. Flow diagram showing the participants in study II
38
Results Over the period 2005 to 2015, the previously depressed female ado-lescents with persistent depressive disorder made significantly more visits and consumed more healthcare (inpatient care, specialized out-patient care and prescribed medication) in their mid-adulthood (ages 31-40) compared to their counterparts with no history of depression in adolescence. The differences were more pronounced for the individu-als with a history of persistent depressive disorder in their adolescence period as shown in table 2. The results from the male participants also showed a tendency for previously depressed male adolescents to con-sume more health care than the non-depressed peers though non-significant. The cost differences were significantly mediated by the recurrence or occurrence of depression later in early adulthood (19-30 years), and the cost differences at a population wide level were considerable as illustrated in table 3.
42
Conclusions In light of the findings, there should be specific concern for adolescents with persistent depressive disorder because the future cost consequences are con-siderable. Therefore, efforts to break the cycle of chronicity and aggressively treating or preventing depression early enough should be prioritized.
Study III: Adolescent depression, early comorbid psychiatric disorders and adult welfare burden: A 25-year longitudinal cohort study Aims To investigate the association between adolescent depression and govern-ment transfer payments (social financial assistance/welfare) related to unem-ployment, work disability, and public assistance across early to middle adulthood (20 - 40 years). The following research questions were examined:
1. What is the magnitude of the government transfer payments associ-ated with the clinical subtypes of depression in adolescence (PDD, episodic MDD, subthreshold depression) compared to no-depression in early to mid- adulthood (age 20 to 40 years)?
2. To which extent are these potential associations influenced by com-mon psychiatric comorbidities in early childhood or adolescence?
Methods This was a register-based study where yearly total social transfer payments received by individuals with a history of adolescent depression compared to those with no such history. The comparisons were done both with and with-out adjustments for covariates thought to influence such a relationship.
Study sample and procedure The same study sample as in study II was used. However, in the present study, no re-matching of participants was done, and thus the initial 539 par-ticipants were included in the analysis as illustrated in figure 4. For further details about the waves of data collection and the sample can be found in the cohort profile describing this study sample (Alaie et al., 2019).
Exposure variable The exposure variable, adolescent depression, was defined based on both self-questionnaires followed by a diagnostic interview (Alaie et al., 2019).
43
Further, the depressed adolescents were classified according to the new DSM V diagnostic criteria as:
1. Persistent depressive disorder (PDD): having a depressed mood most of the day, for more days than not, for at least one year (n=175);
2. Major depressive disorder (MDD): A current or life-time major de-pressive episode lasting shorter than one year (n=82);
3. Subthreshold depression: A positive screening but no other past or current depressive disorder (n=64);
4. No depression: Negative screening, no past or current depressive disorder (n=218).
Comorbidities In accordance to the DICA-R-A (Wendy Reich, Barbara Herjanic, Zila Welner, & P. R. Gandhy, 1982) interviews were conducted at baseline, and the study sample was assessed for anxiety disorders (ADs); separation anxie-ty, overanxious, and avoidant disorder) and disruptive behavior disorders (DBDs); oppositional defiant disorder, conduct disorder, and attention defi-cit/hyperactivity disorder (ADHD) of childhood/adolescence. The depressed adolescents (n=321) were further categorized based on the presence or ab-sence of anxiety and/or disruptive behaviour disorder as follows:
1. Non-comorbid adolescent depression (i.e., no ADs or DBDs) (n=132) 2. Adolescent depression + ADs (i.e., no DBDs) (n=93); 3. Adolescent depression + DBDs (i.e., no ADs) (n=44); 4. Adolescent depression + DBDs + ADs (n=52).
Covariates Variables to paint a picture on the family socioeconomic status at the base-line assessments were retrieved from Statistics Sweden (Ludvigsson, Svedberg, Olen, Bruze, & Neovius, 2019). The highest level of education of either parent was taken as a categorical variable (i.e., university education, yes/no). Disposable family income of either parent was assigned to quintiles defined from the distribution of income of a reference population born in the same year (n=200,000). Thereafter, the quintiles were categorized as earn-ings within or above the 3rd quintile, yes/no.
Outcomes The main outcome measure of interest was differences in the total annual transfer payments of individuals with a history of adolescent depression and those with no such history. The transfer payments studied in this work were: (1) unemployment benefits, (2) work disability benefits, (3) public assistance,
44
and (4) overall payments combined. The data on these outcomes for the peri-od 1994-2016 was retrieved from the Longitudinal Integration Database for Health Insurance and Labor Market Studies (LISA), held by Statistics Swe-den (Ludvigsson et al., 2019). The outcome transfer payments were defined as follows: Unemployment benefits referred to transfer payments for individuals regis-tered as full- or part-time unemployed at the Swedish Public Employment Service (Arbetsformedlingen) (Ludvigsson et al., 2019). To qualify for un-employment benefits, the individual needed to be unemployed, however, able to work and ready to accept work if offered. The beneficiary should have worked for at least six months before becoming unemployed. Work disability benefits were considered as all payments an individual re-ceives due to sickness absence and disability pension (Ludvigsson et al., 2019; Swedish Social Insurance Agency, 2016). In the first fortnight of ab-sence from work due to sickness, the employer is responsible to continue paying the employee. However, any periods of time beyond two weeks, the employee can seek sickness financial benefits/assistance amounting to ap-proximately 80% of lost income from the state. Public assistance was defined as financial assistance in form of social wel-fare assistance, housing supplement, housing allowance, and maintenance support. Social welfare assistance referred to a cash income allowance from the local social authorities, that is counties and municipalities, that is offered only after the individual’s means to earn and access/finance the minimum living standards is assessed (Ludvigsson et al., 2019). The counties and mu-nicipalities also provide maintenance support in form of alimony for children with separated parents where the parents cannot afford to meet the costs for the child’s housing, food, clothes and leisure interests. Overall payments combined were defined as the total cash benefits and fi-nancial assistance described above. Statistical analyses Descriptive statistics for the outcomes, the baseline characteristics for the study sample and the categorized groups were calculated and presented as means and standard deviations or percentages where applicable. The relationship of interest between the exposure variable (adolescent de-pression) and the total social transfer payments was modeled using a gener-alized estimating equations (GEE) (Liang & Zeger, 1986). In the GEE mod-els, an auto-regressive working correlation structure and robust standard
45
errors were chosen and used based on model fit. The associations were ana-lyzed by fitting a tweedie distribution with an identity link function due to the excess of zeros in each of the outcomes (Kurz, 2017). Statistical signifi-cance was set at 0.05 and results presented as mean annual social transfer payments with 95% confidence intervals. There was a small percentage of missing values mainly due to death (n=3) and relocation to a different coun-try (n=66), thus all the analysis used only available data. All analyses were conducted both with and without adjustment for covariates, i.e. socioeco-nomic variables, sex and comorbidities. All the social transfer payments were converted to the value of January 2019 using Consumer Price Index (CPI) (Statistics Sweden, 2019). Thereafter, the estimates were converted from Swedish krona (SEK) to US Dollar (USD) using an exchange rate of 1 SEK = 0.1113 USD, valid as of January 2019.
46
Figure 4. Flow diagram showing the participants in study III
ResultsAdolescent depression was associated with all the different forms of social transfer payments studied. The additional annual overall payments amounted to 938 USD (95% CI: 551 – 1326) in the depressed adolescents group com-pared to the non-depressed adolescents, see table 4. This relationship was
47
evident in the PDD group amounting to 1172 USD (95% CI: 673 –1670) annually. However, this observation was rather less obvious in the MDD and the subthreshold depression groups, with the exception of the public assis-tance. The relationships remained unchanged after adjustment for the ob-served covariates. Further, depressed adolescents with comorbid psychiatric disorders received significantly higher social transfer payments than their peers without comor-bidities. Individuals with comorbid DBDs and ADs received higher social transfer payments across all outcomes, that is, unemployment benefits (566 USD, 95% CI: 60–1071), work disability benefits (545 USD, 95% CI: 9–1081), and public assistance (634 USD, 95% CI: 237–1031), thus emphasiz-ing the role played by comorbidities, see table 5. Adjusting for comorbidities did not change the results. The attempt to examine the different clinical subtypes of depression revealed significant findings for the PDD (1058 USD, 95% CI: 529–1588) and MDD (711 USD, 95% CI: 67–1356) groups, but not the subthreshold depression group.
51
Conclusions The magnitude of the additional long-term societal/public financial assis-tance (transfer payments) associated with adolescent depression are substan-tial. This makes the need to prioritize effective and cost-effective preventive interventions even more important in order to counteract this financial bur-den.
Study IV: Cost-effectiveness of an indicated preventive intervention for depression in adolescents: a model to support decision-making Aims To assess whether an indicated GB-CBT preventive intervention for depres-sion offered to adolescents with subsyndromal depression provides good value for money in the long-term compared to a no-intervention alternative.
Methods Economic evaluation framework This study used a decision analytic modelling approach to assess the long-term cost-effectiveness of an indicated GB-CBT intervention for the preven-tion of depression in adolescents. The evaluation utilized a model-based cost utility and effectiveness analysis conducted over a 5-and 10 years’ time hori-zon from both the societal and healthcare perspectives. Both costs and health benefits (cases of depression prevented and QALYs) were discounted at an annual rate of 3%. Data to populate the model were sourced from existing literature, and where data was missing, assumptions were made.
Study population The study modelled a hypothetical homogeneous cohort of 15-years old ado-lescents with subsyndromal depression (sD) (Rodríguez, Nuevo, Chatterji, & Ayuso-Mateos, 2012). Participants with sD were modelled since indicated GB-CBT preventive interventions have demonstrated effectiveness for this group (Ssegonja et al., 2018). The study population was assumed to have been screened and accepted to take the intervention.
Model description Given the chronic and recurrent nature of depression, the cost-effectiveness of an indicated GB-CBT preventive intervention for adolescents with sub-syndromal depression was modelled using a single cohort Markov model. The model consisted of six health states: healthy (H), subsyndromal depres-
52
sion (sD), depressed (D), remission (R), recovered (Re), and dead (De) as shown in figure 5.
Figure 5. Model structure
The state healthy (H), was defined as individuals with no depressive disorder and no diminished functioning due to depressive symptoms. The remission state (R), was defined as previously depressed individuals with minimal or absence of depressive symptoms and have regained premorbid functionality (Kennedy, 2002). Recovered individuals (Re) were defined as individuals in remission that have been symptom free and regained premorbid functionality for four to six months, and possibly still on or not on medication (Furukawa et al., 2008). The model assumed all participants started in the sD state be-fore moving through the rest of the natural history of the disease. The model used one-year cycles, modelling a cohort of adolescents with sD from a start-age of 15 years over a time horizon of 5- and 10 years from a societal and healthcare perspectives.
Intervention and comparators
GB-CBT indicated preventive intervention The GB-CBT indicated preventive intervention was constructed using char-acteristics from a recent systematic review and meta-analysis (Ssegonja et al., 2018). The average characteristics are summarized in table 6.
53
Table 6. GB-CBT indicated preventive intervention
Intervention feature Mean (SD)
Number of participants per group 8 (1.69) adolescents
Average number of sessions 9 (3.21) sessions
Average attendance 6.5 (2.14) sessions
Average follow-up 13 (10.82) months
Average duration per session 70 (20.44) minutes
Supervision Once weekly
Number of facilitators 1 to 2
Intervention duration 9 (3.18) weeks
Comparator A no-intervention option was considered as the comparator.
Model Inputs
Effectiveness estimates The intervention effect was derived from a recent systematic review and meta-analysis (Ssegonja et al., 2018) studying the effectiveness of indicated GB-CBT preventive interventions for depression in adolescents (Ssegonja et al., 2018). Where the outcome was standardized mean differences (SMDs), these were transformed into RRs for use in the model using the Cochrane conversion method (Higgins & Green, 2011; Lee et al., 2018).
Health benefits Health benefits were expressed as cases of depression prevented and QALYs gained. The cases of depression prevented were estimated by taking the dif-ference in the proportions of the study cohort in the depressed state per cycle between the intervention and comparator groups. The QALYs were estimat-ed as a product of the health-related quality of life (HRQoL) utility weights derived from existing literature and the life years gained from the modelling process. The HRQoL utility weight for the health state sD was considered to be 0.62 (0.58 – 0.65) (Kolovos et al., 2017). In view of absence of symptoms and regaining premorbid functioning, the utility weight for the health state “Recovered” was assumed to be equal to that of a health individual in the general Swedish population, 0.89 (0.78 – 0.95) (Burstrom, Johannesson, & Diderichsen, 2006). The utility weight for the state “Remission” (R) was assumed as 0.70 (0.67 – 0.73) and 0.39 (0.35 – 0.43) for the state “De-pressed” (D) (Kolovos et al., 2017). Costs
54
All original costs were converted to USD and uprated to 2018 prices, using purchasing power parities and inflation indices (Shemilt et al., 2010) .
Intervention cost The intervention cost was estimated from the characteristics of the indicated GB-CBT preventive intervention multiplied by unit costs derived from Sweden’s municipalities and counties and government documents. All costs were based on a group size of eight participants as shown in table 7 below.
Table 7. Cost items and estimation of intervention cost
Direct healthcare costs and indirect costs Cost inputs were retrieved from the literature. The average yearly direct healthcare cost and productivity losses for an individual in the health- and recovered state were assumed to be zero/negligible. For situations where data on direct non-medical costs like patient transportation was available, it was considered in the overall calculations.
Transition probabilities and mortality Transition probabilities between the different health states were derived from existing literature. Mortality data was derived from Swedish life tables for year 2018, from Statistics Sweden (SCB, 2019). The mortality rates were adjusted with hazard ratios or relative risks derived from the literature where necessary. Individuals in remission were considered to have the same elevat-ed mortality risk as those in the state sD. The individuals in the recovered state were assumed to have the same mortality risk as the general population.
Base case and uncertainly analyses The base case analysis involved comparing the GB-CBT indicated preven-tive intervention to the no-intervention option over a follow-up period of 5-
Cost item Frequency Total cost/USD
Training costs
Facilitor training 7 3810.80 Rent of the venue 7 665.28 Participants' allowances 7 165.00 Trainers' accomodation _ _ Trainers' travel allowances _ _
Delivery costs
Facilitator payment 9 1429.05 Supervision costs 9 714.53 Rent of the venue 9 124.74 Costs of materials _ 405.00
Total cost 7314.40 Average total cost per child 914.30
55
and 10 years post intervention. In this analysis, the intervention effect was estimated to decay by 40% annually (Ssegonja et al., 2018). The intervention cost was applied to only the first year of follow-up. The total costs, cases of depression prevented and QALYs gained during each cycle were summed up for the different time horizons. An incremental cost-effectiveness ratio (ICER) was computed and judged at an implied threshold value of $0 – $100000 USD/QALY, as reflected in the international literature (Shiroiwa et al., 2010). It was assumed that the threshold for reimbursement for preven-tive approaches may be somewhat lower but not far different. In all analyses, Markov chain Monte Carlo simulations with 1000 iterations based on values drawn from the parameter assigned distributions were performed to produce 95% uncertainty intervals around the net cost differences, cases of depres-sion prevented, QALYs gained and ICERs estimates
Sensitivity analyses As a sensitivity analysis, the following scenarios were modelled: 1) the in-tervention effectiveness was applied to the first cycle only, 2) increased in-tervention costs by 50%, 3) used a discount rate of 0% for the health benefits and 6% for the costs, 4) collapsing the health states remission and recovered into one state (also acted a form of structural validation), 5) treated the utility weights deterministically, 6) varied the proportion of intervention uptake and 7) used a cycle length of 6 months. All the analyses and modelling were conducted in Microsoft EXCEL version 2016. For more details of the meth-ods and the assumptions made, see the supplementary appendix 1.
Results Base case analysis The base case and uncertainty analyses demonstrated that a GB-CBT indi-cated preventive intervention for depression resulted into more cases of de-pression prevented, increased QALYs gained and decreased costs compared to the control condition. This was true when taking both the healthcare and societal perspectives. The intervention was cost-saving resulting into lower costs and increased health benefits as illustrated in table 8 and 9. The proba-bility of the intervention being cost-effective was over 95%.
Sensitivity analyses Changing different parameters in the sensitivity analysis did not change the direction of the findings. However, increasing the intervention cost by 50% and discounting costs at 6 % while leaving the health benefits undiscounted, resulted into the intervention being less cost saving and ultimately a lower probability for the intervention to be cost-effective.
62
Conclusions GB-CBT indicated preventive interventions for depression in adolescents seem to have a potential to be good value for money. Further studies to rep-licate these findings are needed before a large-scale adoption and implemen-tation is undertaken.
63
General Discussion
The overarching aims of this doctoral thesis were to: 1) investigate the effec-tiveness of a GB-CBT indicated preventive intervention for depression in adolescents, 2) estimate the cost- and societal financial assistance burden of adolescent depression, and 3) evaluate the cost-effectiveness of a GB-CBT indicated preventive intervention for adolescent depression compared to a no intervention option.
Interpretation of results Study I investigated the effectiveness of GB-CBT indicated preventive inter-ventions for depression in children and adolescents as well as explored the impact of the nature of the comparator, booster sessions and bias. The results demonstrated that the interventions significantly reduce depressive symp-toms (d –0.21, 95% CI –0.31 to –0.11), and in some cases protected the indi-viduals from developing a depressive disorder altogether (RR 0.43, 95% CI 0.21–0.87) at post intervention. Although there was a noticeable decrease in the relative intervention effec-tiveness between post-intervention and 6 months, there was also a noticeable rise in the relative effectiveness between 6- and 12 months of follow-up post intervention before falling again. This observed biphasic pattern could be as a result of the immediate intervention effects (first spike) by virtue of partic-ipation and teachings, while the second spike as a result of a delayed effect due to the nature of the intervention, since learning and sustaining new be-havior or change takes time. Another explanation could be that since there was improvement in both the intervention and control conditions, the effect or intervention benefit in the treatment group is achieved faster and main-tained over a longer period compared to the benefit in the control group.
Our results also showed the intervention effect to decay steadily over time but at different rates when comparing to passive and active control condi-tions. Overall, the gradual decay of the relative intervention effect may be explained by the thought that over time individuals’ relapse/recidivate back to their previous patterns of thoughts and behaviors. Another possible expla-nation could be the persistence of some risk factors for depression or ex-pressing depressive symptoms that are not impacted by CBT. Examples of
64
such risk factors may include genetic predisposition, lack of social support or ongoing environmental stressors. Considering the comparison of GB-CBT interventions to active controls, a meaningful relative effectiveness difference was noted after 12 months or more. This could mean that the effect of GB-CBT interventions is sustained longer than the active comparators. However, there was great heterogeneity in the active comparators. The meta-regression results showed bias and the nature of the intervention to impact the intervention relative effectiveness when studying the SMD. A possible justification for these findings could be the increased power to capture these effect differences, as more studies used continuous outcome measures rather than reporting RR. The results of study I are similar to the findings from (Hetrick et al., 2015; Rasing et al., 2017; Stockings et al., 2016) in terms of the effectiveness esti-mates. However, it also uplifts some different aspects like the pattern of de-cay of the treatment effect and the differences in the RRs in relation to active and passive controls which is more meaningful from a policy making per-spective. In this work, the booster sessions showed no impact on the inter-ventions effectiveness as demonstrated by Gearing et al. (2013). One expla-nation for this finding can be attributed to the small number of studies with booster sessions included in this work and that Gearing’s work studied a sample of all mood disorders, as well as combining primary and secondary prevention CBT interventions despite the mode of delivery, which makes it hard to disentangle the observed effects. Study II examined the cumulative long-term healthcare consumption differ-ences between individuals with a history of adolescent depression and their matched non-depressed counterparts. Additionally, it investigated whether the observed healthcare consumption differences are mediated by subsequent depression in early adulthood (19-30 years). The results showed that the female participants with a history of persistent depressive disorder (PDD) in adolescence had increased healthcare utilization and related costs compared to the non-depressed. A proportion of these costs was mediated by subse-quent depression in early adulthood especially for the psychiatric disorders but not the somatic disorders. The results of the male participants were in-consistent possibly due to the small proportion of the male population in the study. The noted differences in the healthcare consumption costs could be inter-preted as small or an underestimate since data on general primary care con-sumption was not included, but they were substantial when demonstrated at a population level.
65
The results of study II may not be comparable to previous studies investing healthcare costs related to depression that were conducted in the Swedish setting, as such studies investigated clinical samples. Owing to the finding that the consumption and related costs for the subthreshold and major epi-sodic depression did not reach the set significance level, it could imply that the duration/chronicity and continuity of depressive disorders (PDD) is truly central to the extent of future healthcare needs. The increased healthcare consumption and related costs were noted in both psychiatric and a range of somatic disorders. The range of somatic disorders happening in depressed individuals may not be surprising, given the docu-mented relationship between inflammatory processes being associated with depression, thus affecting a number of organ systems and the immune sys-tem (Felger and Lotrich, 2013; Leonard, 2010). These findings further em-phasize the need to intervene early and prioritizing prevention. Study III set out to investigate the longer-term social transfer payments (wel-fare payments) in early to mid-adulthood that are associated with adolescent depression at the population level, while accounting for early comorbid psy-chopathology. Our results demonstrated that there is a significant association between ado-lescent depression and the sum of the different social transfer payments stud-ied, that is, unemployment benefits, work disability benefits, and public as-sistance. The mean total additional social transfer payments were estimated to be approximately 1000 USD per person per year for individuals with a history of adolescent depression. This observation was more pronounced in the PDD group and individuals with early comorbid psychopathology, especially disruptive behavioural disorders. We suppose that such an observation maybe due to the scarring effect on human capital development that depression during adolescence can cause. Thus, individuals do not get to achieve marketable skills to help them earn financial self-sufficiency later on in life. Therefore, having an early comorbid psychopathology simply accentuates this problem. The findings in this study are in line with previous work emphasizing the poor future psy-chosocial functioning associated with adolescent depression (Clayborne et al., 2019) and the extent of future disability that conditions like DBDs, ADs and depression can pose to an individual (Copeland, Wolke, Shanahan, and Costello (2015). These findings further emphasize the concern to intervene early through prevention and early treatment approaches that are both efficacious and cost-
66
effective to counteract the future ill-health and the risk of being dependent on social welfare. Preventive approaches, early treatment and timely contact with healthcare services (National Institute for Health and Clinical Excellence, 2011; Neufeld, Dunn, Jones, Croudace, & Goodyer, 2017) have demonstrated good effects and thus worth a try as shown by a scenario case analysis (D. Chisholm et al., 2016). Therefore, as a result governments and institutions are urged to consider implementation of preventive approaches.
Study IV investigated whether GB-CBT indicated preventive interventions for depression in adolescents provide good value for money in the long-term using a single cohort Markov model. The model was populated with data from existing literature and studied a GB-CBT indicated preventive interven-tion constructed from the average intervention characteristics from study I. The aim of the study was addressed through conducting a cost-utility and effectiveness analysis comparing a GB-CBT indicated preventive interven-tion to the no-intervention option using the effectiveness estimates from study I. The analysis results showed GB-CBT indicated preventive interven-tions for depression to have the potential to be cost saving and cost-effective when considering both a societal and a healthcare perspective over a 5- and 10 years’ time-horizon. In the uncertainty and sensitivity analyses, the over-all direction of the results did not change. There was a proportion of simulations where adopting the new intervention compared to a do-nothing alternative resulted into cost savings without sub-sequent gains in health outcomes in QALYs. This observation might be due to several reasons. Firstly, it could be attributed to the disruption in the ordi-nal nature of the utility weights for the health-related quality of life during the monte carlo simulations. This was also tested in the sensitivity analysis and the results supported the suspected notion. Secondly, it can also be due to the uncertainty in the intervention effectiveness that is used to adjust the transition probabilities in the model. The results of this study support the conclusions from the few existing stud-ies examining a similar question. They all reported the interventions to be cost-effective. However, there are also considerable differences between studies ranging from the setting, population studied, the comparator condi-tion, time horizons, age-groups, model type and health states representing the natural history of the disorder, to the choice of outcome measure (Lee et al., 2017; C. Mihalopoulos et al., 2012).
67
Implications to clinical practice All our findings point to the need to consider intervening early and prioritiz-ing prevention of adolescent depression. In study I, our work demonstrated that GB-CBT indicated preventive interventions for adolescent depression are effective in reducing depressive symptoms as well as prevent cases of depression. It also clarified the distinction in the effectiveness estimates de-pending on the nature of the comparator, as well as highlighting the decay process over time. Despite the inability to demonstrate the influence of booster sessions on the effectiveness estimates (due to power reasons as few studies with interventions including booster sessions were found in the lit-erature), it is important to consider adopting these interventions as opposed to leaving adolescents with subthreshold depression untreated given the long-term negative consequences of early onset depression. However, de-spite the demonstrated effectiveness of these interventions and the known long-term negative consequences of adolescent depression, to adopt these preventive interventions would require much more information, including for instance the additional cost burden related to adolescent depression. This paved the way for studies II and III. Study II and III in this thesis contribute to the evidence pool on the long-term consequences of adolescent depression by demonstrating how much more (additional) direct healthcare costs- and social or public financial assis-tance/benefits are received by individuals with adolescent depression com-pared to the non-depressed in the long-term. In study II, individuals with a history of adolescent depression especially females with persistent depres-sive disorder consumed significantly more healthcare resources than their matched counterparts without a history of adolescent depression in their mid-adulthood (30-40 years). Study II also showed that the long-term healthcare consumption was mediated by subsequent depressive episodes in early adulthood, implying that close follow-up and advocating for prevention especially for individuals with a history of depression is key. Despite the cost difference estimates being substantial at the population level, the prima-ry healthcare costs were not included, thus making the estimates conserva-tive. Hence, the additional burden may even be much larger. The productivity loss related to depression is also well documented in the literature. However, less has been established regarding the additional public transfer payments (welfare) that are received by individuals with a history of adolescent depression while in their early to mid-adulthood compared to their matched counterparts without such history. Study III attempted to quan-tify that burden by estimating the financial assistance related to unemploy-ment, disability pension, public assistance and work disability. The findings showed these transfer payments to be substantial and more pronounced in
68
the individuals with persistent depressive disorders and those that have de-pression with early co-morbid psychopathologies. Therefore, these results point to the need for a society-wide alarm to have interventions in place to combat or prevent depression e.g. at work and in school settings before indi-viduals get to the extent of being welfare or public assistance recipients due to potentially preventable circumstances. To some extent, society has already taken steps to adopt GB-CBT indicated preventive interventions for depression some places. This initiative has been fueled by the demonstrable negative consequences of adolescent depression. Therefore, prevention is a possible approach in some instances to be able to combat this scourge and possibly free some resources that can be reallocated to other sectors. However, less is still known about the cost-effectiveness of such interventions.
In study IV, we took a step further and investigated the cost-effectiveness of a GB-CBT indicated preventive intervention compared to a no-intervention option was examined using a single cohort Markov model. The results of the cost utility analysis conducted showed the intervention to be highly cost saving and cost-effective. The results were stable over a series of sensitivity analyses, thereby strengthening the case that these interventions are worth adopting. However, there were a proportion of simulations where adopting the intervention was cost inefficient and hence this uncertainty has to be taken into account during decision making regarding adoption of preventive interventions. As the results from this work demonstrate GB-CBT indicated preventive interventions for adolescent depression to be effective compared to leaving the individuals untreated (study I), and that the intervention was cost-effective (study IV), it is appropriate to advocate for adopting such interven-tions. Given that even the active comparators such as bibliotherapy and counselling or coaching did show a temporary effect, it is safe to assume that, in cases of resource constraints to rollout GB-CBT indicated preventive interventions, society can implement the active comparators first in a step-wise fashion. Larger emphasis should be put on females with subthreshold depression. As demonstrated in study II and III, this group consumes the largest margin of additional healthcare and welfare recipiency, especially those that go ahead to develop persistent depressive disorders.
Future research directions The work in this thesis has both contributed to the scientific literature with new evidence and unveiled areas that require improvement or further re-search.
69
In study I, despite the promising results that GB-CBT indicated preventive interventions are effective in reducing the symptoms or incidence of depres-sion, there were no conclusive findings on the effect of booster sessions. The results around the role played by booster sessions and the appropriate timing to when they should be offered is yet to be investigated. This will be im-portant because including booster sessions in an intervention package surely has economic consequences, but evidence about their additional benefit is still controversial or lacking. Another aspect worthy of further research is the decay rate of the intervention effect. This has paramount meaningfulness to when the timing for booster sessions is appropriate and the use in modelling studies to assess cost-effectiveness. Study II investigated differences in future healthcare consumption and relat-ed costs considering adolescent depression as the exposure. The work high-lighted important findings, and projecting the cost differences at the popula-tion level revealed the costs to be substantial and worthy of attention. How-ever, it also emphasized other areas that commend close research attention such as: 1) the relationship between depression, cumulative stress and the accompanied immune deficiency, 2) the individual diagnosis categories for which significant healthcare consumption was noted for example infectious diseases, endocrine disorders, and non-specific symptoms to name a few. In study III, individuals with a history of adolescent depression received more public or societal financial assistance in early to mid-adulthood than their counterparts with no such history. This result was more pronounced for individuals with persistent depressive disorder and early comorbid psycho-pathologies. Though, the results of the work also incited more questions worthy revisiting, including, the role played by other psychiatric comorbidi-ties which were not investigated here due to data limitations. The possibility of subsequent depression mediating the differences seen in the amount of societal financial assistance received in presence of a history of adolescent depression versus not. Study IV investigated the cost-effectiveness of a GB-CBT indicated preven-tive intervention for adolescent depression using a markov modelling ap-proach. It demonstrated the intervention to be cost saving and cost-effective despite the uncertainties. The results however showed the need to investigate further the assumptions made around the pattern or rate of decay of the in-tervention effect over time. It also left room to test the model on an actual population with a known extent of treatment acceptance, adherence and dif-ferential treatment benefit as well as investigate the wide-spread roll out of such interventions.
70
Methodological considerations There were several shortcomings in the work included in this thesis. In study I, a few limitations were noted. Firstly, it included a few studies that investi-gated interventions that included booster sessions, thus limited the conclu-sions that could be made about the effect of booster sessions on the interven-tion effectiveness. Secondly, most of the included studies in the meta-analysis had a short follow-up period, and thus it was difficult to study the long-term effectiveness or pattern of decay of the intervention effect over time, that is, beyond 24 months. While in study II, using DRG based costing could have led to either under- or overestimating the actual healthcare consumption costs, this could also have been worsened by not including costs due to primary healthcare con-sumption. Therefore, the presented cost estimates in study II are conserva-tive and provide a good starting point to dig into this field. Furthermore, the small proportion of male participants limited conclusive remarks that could be made about the cost burden investigated for the male population. Overall, study III suffered from the difficulty to distinguish between the pattern of onset of the comorbidities or their severity. The results are also liable to changes in the social transfer payments policies that have changed over the years. This could in turn limit the generalization of the results to other settings with different policies and governance systems. Finally, in study IV, due to the assumptions taken about the decay of the intervention effect over time, the intervention cost-effectiveness estimates are uncertain and prone to review with better data inputs. Since the model-ling work heavily depended on data inputs retrieved from the literature, the inherent bias from the original studies providing the model inputs was intro-duced in the models and results should be interpreted in the light of such limitations.
71
General Conclusions
This thesis set out to investigate and shed more light on the: 1) effectiveness of GB-CBT indicated preventive interventions for adolescent depression, 2) the cost and societal financial assistance burden in early to mid-adulthood that is associated with adolescent depression, and 3) the cost-effectiveness of GB-CBT indicated preventive interventions for adolescent depression in the short- and long-term. Considering the long-term negative consequences of adolescent depression, decision makers at different levels in society are concerned with addressing these issues, and indicated preventive interventions have shown the greatest promise. However, their wide spread adoption still requires research evi-dence that this thesis tempted to provide. From study I, the results of our work support the claims that GB-CBT indi-cated preventive interventions for depression in children and adolescents are effective compared to active control conditions and leaving the subthreshold depressed individuals untreated. It is also important to note that the interven-tion effect decays over time, and thus further exploration of the benefit of booster sessions and their appropriate timing to offer them is warranted. The results also emphasize the value of making a distinct difference between active and passive comparators in decision-making. In study II, the healthcare consumption cost burden in mid-adulthood was pronounced in females with a history of persistent depressive disorder in adolescence, but not subthreshold and major episodic depression. Continued depression into early adulthood was shown to mediate these costs. There-fore, the duration and continuity of depressive disorders in adolescence seem to be pivotal in determining later healthcare requirements. A similar pattern of findings was noted in relation to societal financial assistance recipiency in study III. The burden was more pronounced in individuals with persistent depressive disorders and those with early comorbid psychopathologies (dis-ruptive disorders and anxiety disorders). This as possibly due to the accentu-ated scaring effect that causes stagnation of the human capital development during the formative years of development and acquiring marketable skills in an individual’s life. The problem was more pronounced in individuals that had both comorbidities, emphasizing a possible multiplicative effect. Ex-
72
trapolating to a population level highlights the magnitude of the costs and societal financial assistance burden, and thus strengthens the prevention ad-vocacy claims. In an attempt to put a price or cost on the benefits of GB-CBT indicated preventive interventions for depression in adolescents, we performed a mod-elling study. In study IV, our modelling analyses showed the interventions to have a potential to be good value for money. However, further studies to reproduce our findings are important before a wide spread adoption and im-plementation is undertaken. The work in this thesis is important to the research field and aims to aid de-cision making on how to allocate resources to suggested preventive interven-tions as opposed to the competing alternatives. Although, this is only a step to the overall considerations that have to be taken into account in the process of decision making, as for instance equity considerations and budget con-straints, it is a good starting point.
73
Acknowledgements
The work in this thesis was conducted at the department of public health and caring sciences at Uppsala university with financial support from Swedish Research Council (grant number 2014-10092), in collaboration with three other government agencies; Forte, Formas and Vinnova. I would like to extend my appreciation to the participants in the Uppsala longitudinal adolescents depression study (ULADS) for the data that was used in study II and III. Supervisors I thank my supervisors, Anna Sarkadi, Inna Feldman, Sophie Langenskiöld and Ulf Jonsson for taking me under their wings and teaching me the re-search trade. I am grateful that you did not only teach or mentor me but you also had my back all the time. Examinator I also want to acknowledge the guidance and help from my examinator, Mar-tin Ingelsson. Thank you for always making the time. Colleagues and collaborations I would also like to express my sincere gratitude to the Child health and par-enting research group members. The endless discussions, consultations, team building activities, helping one another and afterwork hangouts energized me when the going got tough. You are not only a team but a family indeed. Particular thanks to Filipa Sampaio, Raziye Salari and Antonia Tokes, I asked you the most questions and you were always there to help with every-thing. Special thanks to our colleagues at Örebro university, Karolinska Institutet and Uppsala university; Lars Hagberg, Anna Philipson, Margareta Möller, Iman Alaie, Lars von Knorring, Anne-Liis von Knorring and Hannes Bohman, you taught me a lot and I greatly appreciate it.
74
Family and friends My family and friends did stand with me through my doctoral education. The long periods of being away and not indulging in many get together ac-tivities did take a toll on our relationships but thank you for being under-standing.
75
References
Alaie, I., Philipson, A., Ssegonja, R., Hagberg, L., Feldman, I., Sampaio, F., . . . Jonsson, U. (2019). Uppsala Longitudinal Adolescent Depression Study (ULADS). BMJ Open, 9(3), e024939. doi:10.1136/bmjopen-2018-024939
Andlin-Sobocki, P., Jonsson, B., Wittchen, H. U., & Olesen, J. (2005). Cost of dis-orders of the brain in Europe. Eur J Neurol, 12 Suppl 1, 1-27. doi:10.1111/j.1468-1331.2005.01202.x
Andlin Sobocki, P., & Wittchen, H. U. (2005). Cost of affective disorders in Europe. Eur J Neurol, 12(s1), 34-38. Retrieved from https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1468-1331.2005.01195.x. doi:doi:10.1111/j.1468-1331.2005.01195.x
APA. (2013). American Psychiatric Association, Diagnostic and statistical manual of mental disorders Arlington, VA: American Psychiatric Publishing.
Avenevoli, S., Swendsen, J., He, J.-P., Burstein, M., & Merikangas, K. R. (2015). Major depression in the national comorbidity survey-adolescent supplement: prevalence, correlates, and treatment. J Am Acad Child Adolesc Psychiatry, 54(1), 37-44.e32. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/25524788
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4408277/. doi:10.1016/j.jaac.2014.10.010
Bardone, A. M., Moffitt, T. E., Caspi, A., Dickson, N., & Silva, P. A. (1996). Adult mental health and social outcomes of adolescent girls with depression and con-duct disorder. Dev Psychopathol, 8(4), 811-829. Retrieved from https://www.cambridge.org/core/article/adult-mental-health-and-social-outcomes-of-adolescent-girls-with-depression-and-conduct-disorder/E63D689CB740099BBF71F0D80A9053A8. doi:10.1017/S0954579400007446
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An in-ventory for measuring depression. Arch Gen Psychiatry, 4, 561-571.
Bertha, E. A., & Balazs, J. (2013). Subthreshold depression in adolescence: a sys-tematic review. Eur Child Adolesc Psychiatry, 22(10), 589-603. doi:10.1007/s00787-013-0411-0
Bodden, D. H. M., Stikkelbroek, Y., & Dirksen, C. D. (2018). Societal burden of adolescent depression, an overview and cost-of-illness study. J Affect Disord, 241, 256-262. doi:10.1016/j.jad.2018.06.015
Borenstein, M., Hedges, L. V., Higgins, J., & Rothstein, H. R. (2009). How a MetaAnalysis Works. Introduction to Meta-Analysis, 1-7.
Brent, D. A., Brunwasser, S. M., Hollon, S. D., Weersing, V. R., Clarke, G. N., Dickerson, J. F., . . . Lynch, F. L. (2015). Effect of a cognitive-behavioral pre-vention program on depression 6 years after implementation among at-risk ado-lescents: A randomized clinical trial. JAMA psychiatry, 72(11), 1110-1118.
76
Brettschneider, C., Djadran, H., Harter, M., Lowe, B., Riedel-Heller, S., & Konig, H. H. (2015). Cost-utility analyses of cognitive-behavioural therapy of depres-sion: a systematic review. Psychother Psychosom, 84(1), 6-21. doi:10.1159/000365150
Bromet, E., Andrade, L. H., Hwang, I., Sampson, N. A., Alonso, J., de Girolamo, G., Kessler, R. C. (2011). Cross-national epidemiology of DSM-IV major depres-sive episode. BMC Med, 9, 90. doi:10.1186/1741-7015-9-90
Burstrom, K., Johannesson, M., & Diderichsen, F. (2006). A comparison of individ-ual and social time trade-off values for health states in the general population. Health Policy, 76(3), 359-370. doi:10.1016/j.healthpol.2005.06.011
Cameron, I. M., Lawton, K., & Reid, I. C. (2011). Recognition and subsequent treatment of patients with sub-threshold symptoms of depression in primary care. J Affect Disord, 130(1), 99-105. Retrieved from http://www.sciencedirect.com/science/article/pii/S0165032710006245. doi:http://dx.doi.org/10.1016/j.jad.2010.10.010
Cameron, I. M., Lawton, K., & Reid, I. C. (2011). Recognition and subsequent treatment of patients with sub-threshold symptoms of depression in primary care. J Affect Disord, 130(1-2), 99-105. doi:10.1016/j.jad.2010.10.010
Chisholm, D., Diehr, P., Knapp, M., Patrick, D., Treglia, M., & Simon, G. (2003). Depression status, medical comorbidity and resource costs: Evidence from an international study of major depression in primary care (LIDO). British Journal of Psychiatry, 183(2), 121-131. Retrieved from https://www.cambridge.org/core/article/depression-status-medical-comorbidity-and-resource-costs/C4E0CA1B9F1305929C37280C419E8559. doi:10.1192/bjp.183.2.121
Chisholm, D., Sweeny, K., Sheehan, P., Rasmussen, B., Smit, F., Cuijpers, P., & Saxena, S. (2016). Scaling-up treatment of depression and anxiety: a global re-turn on investment analysis. Lancet Psychiatry, 3(5), 415-424. doi:10.1016/s2215-0366(16)30024-4
Chisholm, D., Sweeny, K., Sheehan, P., Rasmussen, B., Smit, F., Cuijpers, P., & Saxena, S. (2016). Scaling-up treatment of depression and anxiety: a global re-turn on investment analysis. The Lancet Psychiatry, 3(5), 415-424.
Claassen, G. D. H., Hendriks, T. H. B., & Hendrix, E. M. T. (2007). Decision Sci-ence : theory and applications. Wageningen, The Netherlands: Wageningen Ac-ademic Publishers.
Clarke, E. A. (1974). What is Preventive Medicine? Canadian Family Physician, 20(11), 65-68. Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2274388/.
Clayborne, Z. M., Varin, M., & Colman, I. Adolescent Depression and Long-Term Psychosocial Outcomes: A Systematic Review and Meta-Analysis. Journal of the American Academy of Child & Adolescent Psychiatry. Retrieved from https://doi.org/10.1016/j.jaac.2018.07.896. doi:10.1016/j.jaac.2018.07.896
Clayborne, Z. M., Varin, M., & Colman, I. (2019). Systematic Review and Meta-Analysis: Adolescent Depression and Long-Term Psychosocial Outcomes. Journal of the American Academy of Child & Adolescent Psychiatry, 58(1), 72-79. Retrieved from https://doi.org/10.1016/j.jaac.2018.07.896. doi:10.1016/j.jaac.2018.07.896
Copeland, W. E., Adair, C. E., Smetanin, P., Stiff, D., Briante, C., Colman, I., . . . Angold, A. (2013). Diagnostic transitions from childhood to adolescence to ear-ly adulthood. J Child Psychol Psychiatry, 54(7), 791-799. doi:10.1111/jcpp.12062
77
Copeland, W. E., Wolke, D., Shanahan, L., & Costello, E. J. (2015). Adult Func-tional Outcomes of Common Childhood Psychiatric Problems: A Prospective, Longitudinal Study. JAMA Psychiatry, 72(9), 892-899. doi:10.1001/jamapsychiatry.2015.0730
Coretti, S., Rumi, F., & Cicchetti, A. (2019). The Social Cost of Major Depression. Systematic Review. Review of European Studies, 11(1).
Corrieri, S., Heider, D., Conrad, I., Blume, A., König, H.-H., & Riedel-Heller, S. G. (2013). School-based prevention programs for depression and anxiety in adoles-cence: A systematic review. Health promotion international, 29(3), 427-441.
Costello, E. J., Erkanli, A., & Angold, A. (2006). Is there an epidemic of child or adolescent depression? J Child Psychol Psychiatry, 47(12), 1263-1271. doi:10.1111/j.1469-7610.2006.01682.x
Creed, F., Morgan, R., Fiddler, M., Marshall, S., Guthrie, E., & House, A. Depres-sion and Anxiety Impair Health-Related Quality of Life and Are Associated With Increased Costs in General Medical Inpatient. Psychosomatics, 43(4), 302-309. Retrieved from http://dx.doi.org/10.1176/appi.psy.43.4.302. doi:10.1176/appi.psy.43.4.302
Cuijpers, P., Beekman, A. T., & Reynolds, C. F., 3rd. (2012). Preventing depression: a global priority. Jama, 307(10), 1033-1034. doi:10.1001/jama.2012.271
Cuijpers, P., Vogelzangs, N., Twisk, J., Kleiboer, A., Li, J., & Penninx, B. W. (2014). Comprehensive meta-analysis of excess mortality in depression in the general community versus patients with specific illnesses. Am J Psychiatry, 171(4), 453-462. doi:10.1176/appi.ajp.2013.13030325
Culyer, A. J. (2008). The Dictionary of Health Economics: Edward Elgar Publish-ing, Incorporated.
Del Vecchio, P. (2018). The Good News About Preventing Adolescent Depression. Prev Sci. doi:10.1007/s11121-017-0860-7
Drummond, M., Sculpher, M., Torrance, G., O'Brien, B., & Stoddart, G. (2005). Methods for the Economic Evaluation of Health Care Programmes (3rd ed. Vol. 234x156mm). New York: Oxford University Press, Inc., New York.
Drummond, M. F., Sculpher, M. J., Claxton, K., Stoddart, G. L., & Torrance, G. W. (2015). Methods for the Economic Evaluation of Health Care Programmes. Ox-ford: Oxford: Oxford University Press.
Erskine, H. E., Baxter, A. J., Patton, G., Moffitt, T. E., Patel, V., Whiteford, H. A., & Scott, J. G. (2017). The global coverage of prevalence data for mental disor-ders in children and adolescents. Epidemiol Psychiatr Sci, 26(4), 395-402. Re-trieved from https://www.ncbi.nlm.nih.gov/pubmed/26786507. doi:10.1017/S2045796015001158
EUnetHTA. (2015). European Network for Health Technology Assessment: Methods for health economic evaluations, a guideline based on current practices in Eu-rope. Retrieved from
Farah, W. H., Alsawas, M., Mainou, M., Alahdab, F., Farah, M. H., Ahmed, A. T., . . . LeBlanc, A. (2016). Non-pharmacological treatment of depression: a system-atic review and evidence map. Evid Based Med, 21(6), 214-221. doi:10.1136/ebmed-2016-110522
Fauci, A. S., & Harrison, T. R. P. o. i. m. (2008). Harrison's principles of internal medicine (17th ed. / editors, Anthony S. Fauci ... [et al.]. ed.). New York ; Lon-don: McGraw-Hill Medical.
Fazel, S., Wolf, A., Chang, Z., Larsson, H., Goodwin, G. M., & Lichtenstein, P. (2015). Depression and violence: a Swedish population study. Lancet Psychia-try, 2(3), 224-232. doi:10.1016/s2215-0366(14)00128-x
78
Furukawa, T. A., Fujita, A., Harai, H., Yoshimura, R., Kitamura, T., & Takahashi, K. (2008). Definitions of recovery and outcomes of major depression: results from a 10-year follow-up. Acta Psychiatr Scand, 117(1), 35-40. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/17986318
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2253703/. doi:10.1111/j.1600-0447.2007.01119.x
Gearing, R. E., Schwalbe, C. S., Lee, R., & Hoagwood, K. E. (2013). The effective-ness of booster sessions in CBT treatment for child and adolescent mood and anxiety disorders. Depress Anxiety, 30(9), 800-808. doi:10.1002/da.22118
Greifer, N. (2018). cobalt: Covariate Balance Tables and Plots. R package version 3.4.1.
Gustavsson, A., Svensson, M., Jacobi, F., Allgulander, C., Alonso, J., Beghi, E., . . . Group, C. D. (2011). Cost of disorders of the brain in Europe 2010. Eur Neuro-psychopharmacol, 21(10), 718-779. doi:10.1016/j.euroneuro.2011.08.008
Hetrick, S. E., Cox, G. R., & Merry, S. N. (2015). Where to go from here? An ex-ploratory meta-analysis of the most promising approaches to depression preven-tion programs for children and adolescents. Int J Environ Res Public Health, 12(5), 4758-4795. doi:10.3390/ijerph120504758
Higgins, J. P. T., & Green, S. (2011). Cochrane Handbook for Systematic Reviews of Interventions [5.1.0 [updated March 2011]. Available from www.cochrane-handbook.org]. In J. P. T. Higgins & S. Green (Eds.).
Ho, D. E., Imai, K., King, G., & Stuart, E. A. (2007). Matching as Nonparametric Preprocessing for Reducing Model Dependence in Parametric Causal Inference. Political Analysis, 15(3), 199-236. Retrieved from https://www.cambridge.org/core/article/matching-as-nonparametric-preprocessing-for-reducing-model-dependence-in-parametric-causal-inference/4D7E6D07C9727F5A604E5C9FCCA2DD21. doi:10.1093/pan/mpl013
Holmes, E. A., Ghaderi, A., Harmer, C. J., Ramchandani, P. G., Cuijpers, P., Morri-son, A. P., . . . Craske, M. G. The Lancet Psychiatry Commission on psycholog-ical treatments research in tomorrow's science. The Lancet Psychiatry, 5(3), 237-286. Retrieved from http://dx.doi.org/10.1016/S2215-0366(17)30513-8. doi:10.1016/S2215-0366(17)30513-8
Horowitz, J. L., & Garber, J. (2006). The prevention of depressive symptoms in children and adolescents: A meta-analytic review. In: American Psychological Association.
Hu, T. W. (2006). Perspectives: an international review of the national cost esti-mates of mental illness, 1990-2003. J Ment Health Policy Econ, 9(1), 3-13.
IHME. (2015). Global Burden of Disease (1990-2013). Retrieved from http://vizhub.healthdata.org/gbd-compare/
Institute of Medicine Committee on Prevention of Mental, D. (1994). In P. J. Mrazek & R. J. Haggerty (Eds.), Reducing Risks for Mental Disorders: Fron-tiers for Preventive Intervention Research. Washington (DC): National Acade-mies Press (US)
Copyright 1994 by the National Academy of Sciences . All rights reserved. Jonsson, U. (2010). Adolescents with Depression Grown Up. Education, Intimate
Relationships, Mental Health, and Personality., Uppsala University, Uppsala Julia, F.-R., & John, C. (2005). Economic Evaluation: McGraw-Hill Education.
79
Karyotaki, E., Tordrup, D., Buntrock, C., Bertollini, R., & Cuijpers, P. (2017). Eco-nomic evidence for the clinical management of major depressive disorder: a sys-tematic review and quality appraisal of economic evaluations alongside random-ised controlled trials. Epidemiology and Psychiatric Sciences, 26(5), 501-516. Retrieved from <Go to ISI>://WOS:000417439600010. doi:10.1017/s2045796016000421
Kennedy, S. (2002). Full remission: a return to normal functioning. Journal of psy-chiatry & neuroscience : JPN, 27(4), 233-234. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/12174731
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC161656/. Kernick, D. (2003). Introduction to health economics for the medical practitioner.
Postgraduate Medical Journal, 79(929), 147-150. Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1742631/. doi:10.1136/pmj.79.929.147
Kessler, R. C., Aguilar-Gaxiola, S., Alonso, J., Chatterji, S., Lee, S., Ormel, J., . . . Wang, P. S. (2009). The global burden of mental disorders: An update from the WHO World Mental Health (WMH) Surveys. Epidemiol Psichiatr Soc, 18(1), 23-33.
Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E. (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disor-ders in the National Comorbidity Survey Replication. Arch Gen Psychiatry, 62(6), 593-602. doi:10.1001/archpsyc.62.6.593
Khan, A., Faucett, J., Lichtenberg, P., Kirsch, I., & Brown, W. A. (2012). A Sys-tematic Review of Comparative Efficacy of Treatments and Controls for De-pression. PLoS One, 7(7), e41778. Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3408478/. doi:10.1371/journal.pone.0041778
Kind, P., & Sorensen, J. (1993). The costs of depression. International Clinical Psychopharmacology, 7(3), 191-196. Retrieved from https://journals.lww.com/intclinpsychopharm/Fulltext/1993/00730/The_costs_of_depression.10.aspx.
Kolovos, S., Bosmans, J. E., van Dongen, J. M., van Esveld, B., Magai, D., van Straten, A., . . . van Tulder, M. W. (2017). Utility scores for different health states related to depression: individual participant data analysis. Qual Life Res, 26(7), 1649-1658. doi:10.1007/s11136-017-1536-2
Konig, H., Konig, H. H., & Konnopka, A. (2019). The excess costs of depression: a systematic review and meta-analysis. Epidemiol Psychiatr Sci, 1-16. doi:10.1017/s2045796019000180
Kurz, C. F. (2017). Tweedie distributions for fitting semicontinuous health care utilization cost data. BMC Medical Research Methodology, 17(1), 171. doi:10.1186/s12874-017-0445-y
Kyu, H. H., Pinho, C., Wagner, J. A., Brown, J. C., Bertozzi-Villa, A., Charlson, F. J., . . . Vos, T. (2016). Global and National Burden of Diseases and Injuries Among Children and Adolescents Between 1990 and 2013: Findings From the Global Burden of Disease 2013 Study. JAMA Pediatrics, 170(3), 267-287. doi:10.1001/jamapediatrics.2015.4276
Lee, Y. Y., Barendregt, J. J., Stockings, E. A., Ferrari, A. J., Whiteford, H. A., Pat-ton, G. A., & Mihalopoulos, C. (2017). The population cost-effectiveness of de-livering universal and indicated school-based interventions to prevent the onset of major depression among youth in Australia. Epidemiol Psychiatr Sci, 26(5), 545-564. doi:10.1017/s2045796016000469
80
Lee, Y. Y., Le, L. K., Stockings, E. A., Hay, P., Whiteford, H. A., Barendregt, J. J., & Mihalopoulos, C. (2018). Estimation of a Relative Risk Effect Size when Us-ing Continuous Outcomes Data: An Application of Methods in the Prevention of Major Depression and Eating Disorders. Med Decis Making, 38(7), 866-880. doi:10.1177/0272989x18793394
Lewinsohn, P. M., Rohde, P., Klein, D. N., & Seeley, J. R. (1999). Natural course of adolescent major depressive disorder: I Continuity into young adulthood. Jour-nal of the American Academy of Child & Adolescent Psychiatry, 38(1), 56-63. Retrieved from https://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=1999-00128-018&site=ehost-live. doi:10.1097/00004583-199901000-00020
Liang, K.-Y., & Zeger, S. L. (1986). Longitudinal data analysis using generalized linear models. Biometrika, 73(1), 13-22.
Linde, K., Sigterman, K., Kriston, L., Rucker, G., Jamil, S., Meissner, K., & Schnei-der, A. (2015). Effectiveness of psychological treatments for depressive disor-ders in primary care: systematic review and meta-analysis. Ann Fam Med, 13(1), 56-68. doi:10.1370/afm.1719
Ludvigsson, J. F., Svedberg, P., Olen, O., Bruze, G., & Neovius, M. (2019). The longitudinal integrated database for health insurance and labour market studies (LISA) and its use in medical research. European Journal of Epidemiology, 34(4), 423-437. doi:10.1007/s10654-019-00511-8
Luppa, M., Heinrich, S., Angermeyer, M. C., Konig, H. H., & Riedel-Heller, S. G. (2007). Cost-of-illness studies of depression: a systematic review. J Affect Dis-ord, 98(1-2), 29-43. doi:10.1016/j.jad.2006.07.017
Luppa, M., Heinrich, S., Angermeyer, M. C., König, H.-H., & Riedel-Heller, S. G. (2003). Cost-of-illness studies of depression. J Affect Disord, 98(1), 29-43. Re-trieved from http://dx.doi.org/10.1016/j.jad.2006.07.017. doi:10.1016/j.jad.2006.07.017
Lynch, F. L., Hornbrook, M., Clarke, G. N., Perrin, N., Polen, M. R., O'Connor, E., & Dickerson, J. (2005). Cost-effectiveness of an intervention to prevent depres-sion in at-risk teens. Arch Gen Psychiatry, 62(11), 1241-1248. doi:10.1001/archpsyc.62.11.1241
Löthgren, M. (2004). Economic evidencein affective disorders:a review. The Euro-pean Journal of HealthEconomics, 5(1), s12-s20. Retrieved from https://doi.org/10.1007/s10198-005-0283-4. doi:10.1007/s10198-005-0283-4
Mathers, C. D., & Loncar, D. (2006). Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med, 3(11), e442. doi:10.1371/journal.pmed.0030442
McDaid, D., Park, A. L., & Wahlbeck, K. (2019). The Economic Case for the Pre-vention of Mental Illness. Annu Rev Public Health, 40(1), 373-389. Retrieved from https://doi.org/10.1146/annurev-publhealth-040617-013629. doi:10.1146/annurev-publhealth-040617-013629
McLeod, G. F. H., Horwood, L. J., & Fergusson, D. M. (2016). Adolescent depres-sion, adult mental health and psychosocial outcomes at 30 and 35 years. Psychol Med, 46(7), 1401-1412. Retrieved from https://www.cambridge.org/core/article/adolescent-depression-adult-mental-health-and-psychosocial-outcomes-at-30-and-35-years/7CD2843B007D8F23F3E70DE604CDA447. doi:10.1017/S0033291715002950
81
Merikangas, K. R., He, J. P., Burstein, M., Swanson, S. A., Avenevoli, S., Cui, L., . . . Swendsen, J. (2010). Lifetime prevalence of mental disorders in U.S. adoles-cents: results from the National Comorbidity Survey Replication--Adolescent Supplement (NCS-A). J Am Acad Child Adolesc Psychiatry, 49(10), 980-989. doi:10.1016/j.jaac.2010.05.017
Merry, S. N., & Spence, S. H. (2007). Attempting to prevent depression in youth: A systematic review of the evidence. Early Intervention in Psychiatry, 1(2), 128-137.
Mihalopoulos, C., & Chatterton, M. L. (2015). Economic evaluations of interven-tions designed to prevent mental disorders: a systematic review. Early Interv Psychiatry, 9(2), 85-92. doi:10.1111/eip.12156
Mihalopoulos, C., & Vos, T. (2013). Cost-effectiveness of preventive interventions for depressive disorders: an overview. Expert Rev Pharmacoecon Outcomes Res, 13(2), 237-242. doi:10.1586/erp.13.5
Mihalopoulos, C., Vos, T., Pirkis, J., & Carter, R. (2012). The population cost-effectiveness of interventions designed to prevent childhood depression. Pediat-rics, 129(3), e723-730. doi:10.1542/peds.2011-1823
Mihalopoulos, C., Vos, T., Pirkis, J., Smit, F., & Carter, R. (2011). Do Indicated Preventive Interventions for Depression Represent Good Value for Money? Australian & New Zealand Journal of Psychiatry, 45(1), 36-44. Retrieved from http://journals.sagepub.com/doi/abs/10.3109/00048674.2010.501024. doi:10.3109/00048674.2010.501024
Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med, 6(7), e1000097. doi:10.1371/journal.pmed.1000097
Morris, S., Devlin, N., & Spencer, A. (2012). Economic Analysis in Health Care (2nd ed.). Great Britain: John Wiley & Sons, Ltd.
Naicker, K., Galambos, N. L., Zeng, Y., Senthilselvan, A., & Colman, I. (2013). Social, demographic, and health outcomes in the 10 years following adolescent depression. J Adolesc Health, 52(5), 533-538. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/23499382. doi:10.1016/j.jadohealth.2012.12.016
National Institute for Health and Clinical Excellence. (2011). Common Mental Health Disorders: Identification and Pathways to Care (Vol. CG123). Leicester (UK): The British Psychological Society & The Royal College of Psychiatrists.
Neufeld, S. A. S., Dunn, V. J., Jones, P. B., Croudace, T. J., & Goodyer, I. M. (2017). Reduction in adolescent depression after contact with mental health ser-vices: a longitudinal cohort study in the UK. Lancet Psychiatry, 4(2), 120-127. doi:10.1016/s2215-0366(17)30002-0
Neumann, P. J., Ganiats, T. G., Russell, L. B., Sanders, G. D., Siegel, J. E., & Ox-ford University, P. (2017). Cost-effectiveness in health and medicine. New York: Oxford University Press.
NICE. (2013a). National Institute for Health and Care Excellence: a guide to the methods of technology appraisal. Retrieved from London:
NICE. (2013b). Recommendations for the management of Depression. NICE Re-trieved from http://www.nice.org.uk/guidance/cg90/chapter/1-recommendations#treatment-choice-based-on-depression-subtypes-and-personal-characteristics
OECD. Health at a Glance 2015: OECD Publishing. OECD. (2010). Sickness, Disability and Work: Breaking the Barriers.
82
Olesen, J., Gustavsson, A., Svensson, M., Wittchen, H. U., & Jönsson, B. (2012). The economic cost of brain disorders in Europe. Eur J Neurol, 19(1), 155-162. Retrieved from http:https://doi.org/10.1111/j.1468-1331.2011.03590.x. doi:10.1111/j.1468-1331.2011.03590.x
Olsson, G., & von Knorring, A. L. (1997a). Beck's Depression Inventory as a screening instrument for adolescent depression in Sweden: gender differences. (0001-690X (Print)).
Olsson, G., & von Knorring, A. L. (1997b). Depression among Swedish adolescents measured by the self rating scale Center for Epidemiology Studies - Depression Child (CES-DC). Eur Child Adolesc Psychiatry, 6(2), 81-87. Retrieved from https://doi.org/10.1007/BF00566670. doi:10.1007/BF00566670
Olsson, G., & von Knorring, A. L. (1999). Adolescent depression: prevalence in Swedish high-school students. Acta Psychiatr Scand, 99(5), 324-331.
Osby, U., Brandt, L., Correia, N., Ekbom, A., & Sparen, P. (2001). Excess mortality in bipolar and unipolar disorder in Sweden. Arch Gen Psychiatry, 58(9), 844-850.
Rasing, S. P. A., Creemers, D. H. M., Janssens, J. M. A. M., & Scholte, R. H. J. (2017). Depression and Anxiety Prevention Based on Cognitive Behavioral Therapy for At-Risk Adolescents: A Meta-Analytic Review. Frontiers in Psy-chology, 8, 1066. Retrieved from https://www.frontiersin.org/article/10.3389/fpsyg.2017.01066.
Reich, W., Herjanic, B., Welner, Z., & Gandhy, P. R. (1982). Development of a structured psychiatric interview for children: agreement on diagnosis comparing child and parent interviews. J Abnorm Child Psychol, 10(3), 325-336.
Reich, W., Herjanic, B., Welner, Z., & Gandhy, P. R. (1982). Development of a structured psychiatric interview for children: Agreement on diagnosis compar-ing child and parent interviews. Journal of Abnormal Child Psychology, 10(3), 325-336. Retrieved from https://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=1983-03615-001&site=ehost-live. doi:10.1007/BF00912325
Reiner, R. C., Jr., Olsen, H. E., Ikeda, C. T., Echko, M. M., Ballestreros, K. E., Manguerra, H., . . . Kassebaum, N. J. (2019). Diseases, Injuries, and Risk Fac-tors in Child and Adolescent Health, 1990 to 2017: Findings From the Global Burden of Diseases, Injuries, and Risk Factors 2017 Study. JAMA Pediatrics, 173(6), e190337. doi:10.1001/jamapediatrics.2019.0337
Rodríguez, M. R., Nuevo, R., Chatterji, S., & Ayuso-Mateos, J. L. (2012). Defini-tions and factors associated with subthreshold depressive conditions: a systemat-ic review. BMC Psychiatry, 12, 181. doi:10.1186/1471-244x-12-181
SCB. (2019). Statistics Sweden (A government agency for public statistics). https://www.scb.se/en/finding-statistics/statistics-by-subject-area/population/population-composition/population-statistics/pong/tables-and-graphs/yearly-statistics--the-whole-country/life-expectancy/
Schoenbach, V. J., Kaplan, B. H., Grimson, R. C., & Wagner, E. H. (1982). Use of a symptom scale to study the prevalence of a depressive syndrome in young ado-lescents. Am J Epidemiol, 116(5), 791-800.
Shea, B. J., Grimshaw, J. M., Wells, G. A., Boers, M., Andersson, N., Hamel, C., . . . Bouter, L. M. (2007). Development of AMSTAR: a measurement tool to as-sess the methodological quality of systematic reviews. BMC Med Res Methodol, 7, 10. doi:10.1186/1471-2288-7-10
Shemilt, I., Thomas, J., & Morciano, M. (2010). "A web-based tool for adjusting costs to a specific target currency and price year". Evidence & Policy, 6(no. 1), 51 - 59.
83
Shiroiwa, T., Sung, Y. K., Fukuda, T., Lang, H. C., Bae, S. C., & Tsutani, K. (2010). International survey on willingness-to-pay (WTP) for one additional QALY gained: what is the threshold of cost effectiveness? Health Econ, 19(4), 422-437. doi:10.1002/hec.1481
Skokauskas, N., Lavelle, T. A., Munir, K., Sampaio, F., Nystrand, C., McCrone, P., . . . Belfer, M. (2018). The cost of child and adolescent mental health services. Lancet Psychiatry, 5(4), 299-300. doi:10.1016/s2215-0366(18)30089-0
Sobocki, P., Ekman, M., Agren, H., Krakau, I., Runeson, B., Martensson, B., & Jonsson, B. (2007). Resource use and costs associated with patients treated for depression in primary care. Eur J Health Econ, 8(1), 67-76. doi:10.1007/s10198-006-0008-3
Sobocki, P., Jonsson, B., Angst, J., & Rehnberg, C. (2006). Cost of depression in Europe. J Ment Health Policy Econ, 9(2), 87-98.
Sobocki, P., Lekander, I., Borgstrom, F., Strom, O., & Runeson, B. (2007). The economic burden of depression in Sweden from 1997 to 2005. Eur Psychiatry, 22(3), 146-152. doi:10.1016/j.eurpsy.2006.10.006
Ssegonja, R., Nystrand, C., Feldman, I., Sarkadi, A., Langenskiold, S., & Jonsson, U. (2018). Indicated preventive interventions for depression in children and ado-lescents: A meta-analysis and meta-regression. Prev Med, 118, 7-15. doi:10.1016/j.ypmed.2018.09.021
Statistics Sweden. (2019). Consumer Price Index (CPI). Retrieved from https://www.scb.se/hitta-statistik/statistik-efter-amne/priser-och-konsumtion/konsumentprisindex/konsumentprisindex-kpi/pong/tabell-och-diagram/konsumentprisindex-kpi/kpi-faststallda-tal-1980100/
Stice, E., Shaw, H., Bohon, C., Marti, C. N., & Rohde, P. (2009). A meta-analytic review of depression prevention programs for children and adolescents: factors that predict magnitude of intervention effects. In: American Psychological As-sociation.
Stockings, E., Degenhardt, L., Dobbins, T., Lee, Y., Erskine, H., Whiteford, H., & Patton, G. (2016). Preventing depression and anxiety in young people: a review of the joint efficacy of universal, selective and indicated prevention. Psychol Med, 46(1), 11-26.
Swedish Council on Health Technology, A. (2004). SBU Systematic Review Sum-maries. In Treatment of Depression: A Systematic Review. Stockholm: Swedish Council on Health Technology Assessment (SBU)
Copyright (c) 2004 by the Swedish Council on Health Technology Assessment. Swedish Social Insurance Agency. (2016). Social insurance in figures 2016. Re-
trieved from https://www.forsakringskassan.se/wps/wcm/connect/d9a3498f-ea2a-40a7-a358-80722d13963a/socialforsakringen-i-siffror-2016-engelsk.pdf?MOD=AJPERES
Thapar, A., Collishaw, S., Pine, D. S., & Thapar, A. K. (2012). Depression in ado-lescence. Lancet, 379(9820), 1056-1067. doi:10.1016/s0140-6736(11)60871-4
Thompson, A. E., Anisimowicz, Y., Miedema, B., Hogg, W., Wodchis, W. P., & Aubrey-Bassler, K. (2016). The influence of gender and other patient character-istics on health care-seeking behaviour: a QUALICOPC study. BMC family practice, 17, 38-38. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/27036116
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4815064/. doi:10.1186/s12875-016-0440-0
84
Tindall, L., Mikocka-Walus, A., McMillan, D., Wright, B., Hewitt, C., & Gascoyne, S. (2017). Is behavioural activation effective in the treatment of depression in young people? A systematic review and meta-analysis. Psychol Psychother, 90(4), 770-796. doi:10.1111/papt.12121
Utvärdering, S. B. f. M. (2013). Utvärdering av metoder i hälso- och sjukvården - En handbok. Retrieved from
von Knorring, L., Åkerblad, A. C., Bengtsson, F., Carlsson, Å., & Ekselius, L. (2006). Cost of depression: effect of adherence and treatment response. Europe-an Psychiatry, 21(6), 349-354. Retrieved from http://dx.doi.org/10.1016/j.eurpsy.2006.04.005. doi:10.1016/j.eurpsy.2006.04.005
WHO. (2017). Current health expenditure (CHE) as percentage of gross domestic product (GDP) (%) Data by country Retrieved from http://apps.who.int/gho/data/node.main.GHEDCHEGDPSHA2011
Wulsin, L. R., Vaillant, G. E., & Wells, V. E. (1999). A systematic review of the mortality of depression. Psychosom Med, 61(1), 6-17.
Wykes, T., Haro, J. M., Belli, S. R., Obradors-Tarragó, C., Arango, C., Ayuso-Mateos, J. L., . . . Wittchen, H.-U. (2015). Mental health research priorities for Europe. The Lancet Psychiatry, 2(11), 1036-1042. Retrieved from http://www.sciencedirect.com/science/article/pii/S2215036615003326. doi:https://doi.org/10.1016/S2215-0366(15)00332-6
Yang, L., Zhou, X., Zhou, C., Zhang, Y., Pu, J., Liu, L., . . . Xie, P. (2017). Efficacy and Acceptability of Cognitive Behavioral Therapy for Depression in Children: A Systematic Review and Meta-analysis. Academic Pediatrics, 17(1), 9-16. Re-trieved from http://www.sciencedirect.com/science/article/pii/S1876285916304053. doi:https://doi.org/10.1016/j.acap.2016.08.002
Zechmeister, I., Kilian, R., McDaid, D., & group, M. (2008). Is it worth investing in mental health promotion and prevention of mental illness? A systematic review of the evidence from economic evaluations. BMC Public Health, 8, 20. doi:10.1186/1471-2458-8-20
Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1640
Editor: The Dean of the Faculty of Medicine
A doctoral dissertation from the Faculty of Medicine, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofMedicine. (Prior to January, 2005, the series was publishedunder the title “Comprehensive Summaries of UppsalaDissertations from the Faculty of Medicine”.)
Distribution: publications.uu.seurn:nbn:se:uu:diva-404547
ACTAUNIVERSITATIS
UPSALIENSISUPPSALA
2020
