Economic Protocol for Mental Health Intervention

1

Family-based psychosocial intervention for returning child soldiers in Central African

Republic: a randomised control trial with economic evaluation

Background

The 2007 Paris Principles(1) defines “a child associated with an armed force or armed group”

(CAFFAG) as “any person below 18 years of age who is or who has been recruited or used by

an armed force or armed group in any capacity, including but not limited to children, boys,

and girls used as fighters, cooks, porters, messengers, spies, or for sexual purposes. It does

not only refer to a child who is taking or has taken a direct part in hostilities.” These

CAAFAG, also known as child soldiers(2) , witness and experience a range of traumatising

events in combat, causing child soldiers who have been released to have very low emotional

well-being: child soldiers suffer from a variety of mental illnesses including posttraumatic

stress disorder, depression, and anxiety(3). This state of health makes returning home a very

disorienting experience for these soldiers and it is extremely important for their families and

communities to accept and support them through this time of change. However, the

children's homes are usually within conflict zones, and their families and community

members have also been traumatised and disoriented. Lack of education and preparedness,

fear, spirituality, and recognised behavioural differences in the child (e.g. hostility, anger,

unable to handle situations without violence(4)) create stigmatising attitudes and behaviours

amongst family members and the community(5) hindering successful reintegration of these

children. Research has shown that child soldiers who are supported by their family upon

their return are more likely to be socially reintegrated and have higher emotional well-being

than those stigmatised and rejected, and, lack of social support has been shown to increase

risk for PTSD, anxiety, and depression(6). Also, it's been reported(4) that PTSD in returned

child soldiers are vulnerable to viewing the world as "good versus bad," encouraging violent

2

behaviour and reactions outside of fighting and in their communities, potentially leading to

increased contact with criminal justice system. Humanitarian aids have guidelines(1) to

support their staff in the reintegration and reunification process, but are consistently met

with the challenges listed above. A family-based psychosocial intervention could support

psychosocial adjustment(7), decrease the risk of mental illness and encourage reintegration

of child soldiers cross-culturally, but no policy is currently in place to allow automatic use of

such a tool.

Policy context, target audience, perspective

A recent peace deal in the Central African Republic (CAR) has enabled the release of

hundreds of children since May of 2015(8), and along with retracing these children to their

homes, UNICEF, with support from public sector partners (i.e. the U.N.), has been working

feverishly to promote reintegration. To help facilitate successful reintegration while using

the research base above, this study would like to explore the usefulness and cost of a

family-based psychosocial intervention for recently reunited child soldiers and families.

From the perspective of services provided by the public sector of humanitarian aids and the

criminal justice system, and with consideration of funding from the charity UNICEF, is a

caregiver-child family-based psychosocial intervention program for returning child soldiers

preferable to the current, non-existent family-based psychosocial intervention program?

Aims and objectives

This economic evaluation will examine the effects and cost-effectiveness of a caregiver-child

family-based psychosocial intervention program to strengthen families and support positive

reintegration of returning child soldiers. A randomised control trial design will be used to

compare the effectiveness and cost-effectiveness of this intervention with that of routine

3

practice, which currently stands as non-existent. The evaluation will compare the monetary

costs and effects of the 4-week intervention with the effects of the control (no intervention)

on the child soldier’s emotional well-being, rate of social integration, and amount of

interaction with the criminal justice system.

Trial design

A two-armed, clustered, pragmatic randomised control trial (RCT) will be performed. After

receiving approval from UNICEF, this study will recruit through the matches made in the

tracing process. Consent will be obtained from both the child and caregiver. Participants will

be groups of similar demographics (village, gender, and age group) and we will use a

computer generated sequence to randomly assign groups to the intervention or control(9).

To reduce selection bias, the person recruiting participants will be concealed from

treatment allocation and the person allocating the group assignment will have no previous

interaction with the participants.

Setting

The setting will take place in communities throughout CAR. The initial release of over 300

child soldiers happened in the town of Bambari in 2015(8). We will begin recruitment with

the matched families of those children.

Participants

A participant involves one child soldier and his or her caregiver. This study will include

children that fit within the definition of a child soldier given by the 2007 Paris Principles,

referenced above. Children will have served at least one month in armed forces(6) abducted

by the armed forces. Accounting for any possible intended disagreement during

4

intervention, we will exclude children who had volunteered to be in armed forces will be

excluded. The caregiver is defined as having full responsibility/guardianship of the child and

will only be included as such. A final exclusion factor for both caregiver and child will be the

result of any psychosis symptoms present or show severe substance abuse based on results

of an initial psychological screening.

Intervention

This intervention will be run similarly to that of a pilot study(10) recently run on children and

their caregivers living in conflict zones. Measures will be collected pre-intervention, post-

intervention (4 weeks), and at a 3-month follow-up. This is a group-based psychosocial

intervention. It consists of 2-hour blocks, three times a week for four weeks. It will consist of

seven chapters of training including: 1) Psychoeducation, 2) Relaxation Techniques(11), 3)

creative problem-solving techniques to tackle big family issues that may arise, 4)

interpersonal communication, 5) community-based conflict resolution methods, 6) effective

parenting and child contribution, and 7) a summary of learned practices. A trained

teacher/group leader will run each group.

According to established good practice(12) measures and materials not currently in CAR’s

official languages, French or Sangho, will go through a process of translation and adaptation

before commencing the study, ensuring coherency with language and local cultural

definitions. Accounting for the illiteracy rates in CAR(13), it’s possible some participants may

not be able to read. Therefore, research assistants will be prepared to read questions from

the measures in the official languages.

Sample size

5

Data on mental illness interventions and reintegration processes for child soldiers in CAR is

scarce, and general family-based interventions are yet to be verified in similar populations

around the world(7). However, the pilot study, which we are basing our intervention from,

worked with a conflict affected population and although not child soldiers, these children

witnessed traumatizing events such as torture and massacre. If we use this as a comparison,

we can adapt the means of both groups from the depression/anxiety outcome from this

study in a power sample calculation(14). The calculation predicted a sample size of 116 to be

significant to the true population for this study.

Identification, measurement and valuation of outcomes

The primary clinical outcome for this study will be lower internalising symptomology in the

intervention arm. Secondary outcomes will include less conduct problems, higher pro-social

behaviour, more social integration, and less interaction with the criminal justice system for

those in the intervention arm. The African Youth Psychosocial Assessment Instrument

(AYPA)(15) is an African based, validated questionnaire(16), and will be used to measure

internalising symptoms, conduct problems, and pro-social behaviour. Due to cultural

differences, past research has relied on creating a local measure for community

reintegration(6),therefore our study will create a local 11-question measure for integration.

Interaction with the criminal justice system will be measured through self-report and

interview from both the child and guardian, and short interviews will be conducted with

local community members. The World Health Organization Quality Of Life-Spirituality,

Religiousness and Personal Beliefs Field-Test Instrument (WHOQOL-SRPB)(17) is based off of

(and used in tandem with) the original WHOQOL-100(18) but includes a domain to analyse

the religious and spiritual influences on cultures' perspectives. This will be used to collect a

6

generic quality of life measure to be used in for the statistical analysis of this economic

evaluation.

Identification, measurement, and valuation of resources

This table represents all identified resources with methods for measuring and valuing those

resources. Literature supporting the chosen methods and available resources are

referenced within.

Type of Resource

Identified

Method for Measuring Resource Valuation of Resource

Staff:

Teachers/Group leaders

for 8 session intervention

- Documented hours of direct

contact with the group

- Records and diaries for any

indirect time spent on

materials

- Reference to current

UNICEF salaries for staff

members. Wages/salary

per teacher will be

divided by hours of

contact worked per

week to give a final

value of cost per hour.

- Per literature(19), client-

related time will be

calculated using the

ratio of direct and

indirect

Training of staff:

Trainers, facility,

materials

- Document hours of

contact the trainer spent

teaching staff.

- Hours spent at facility for

training

- Unit prices for materials

- Reference to

organization

supplying the

intervention

materials and trainer

for unit prices(20).

- Divide the annual

7

overhead cost of the

facility by hours

spent for unit price

per hour

Intervention materials:

“Relaxation Technique”

scripts, cinema clips,

extra supplies (pencils,

paper, etc.), tablet for

cinema clips

- Count for number of

scripts needed for each

participant/group (some

may be shared), purchase

of one set of cinema clips,

final participant

involvement for number of

pencils/paper, and one

tablet

- Trauma-Focused CBT

website for price of

scripts(11)

- Provider’s listed

price for cinema

clips

- Local store unit

prices for extra

supplies

- Unit price for tablet

at local supplier

Facility:

Space for therapy

- Document number of

hours spent at facility

during intervention

- Account for any extra

furniture that may need to

be purchased

- Annual overhead

and capital price for

facility divided by

amount of hours

spent in the space

for unit price per

hour(19)

- Local supplier for

unit costs for extra

furniture

Criminal Justice System:

Contact

- Self-report from child

- Interview/report from

guardian

- Interview/report from

community members

- Price/reports of

costs(19) from

criminal justice

system (hourly wage

of officer, any time

spent in cell)

8

- Consequence

valuation(19): Cost of

property loss and

lost production time

if crime was a

robbery from local

store

Method of economic evaluation

This study must accommodate for a more generic form of comparison between treatments

(intervention arm vs. control), therefore a cost-utility analysis will be performed. Quality of

life years (QALY) measured through the EQ-5D(21) is the most preferable incremental unit of

health gain used in order to quantitatively compare benefits produced by the interventions

with monetary costs of the interventions. Unfortunately, the validated EQ-5D is not

applicable to this complex population, and therefore a strict conversion to the well-known

QALY is impossible. However, the WHOQOL-SPRB collects and measures the respondents

perceived quality of life, and higher scores represent a higher quality of life for each of the

five main domains: physical, psychological, level of independence, social relationships, and

environment. So, after calculating mean scores for both arms, we will use the algorithm

provided in the measure manual, then convert scores to a 0-100 rating, divide each score by

100, and produce scores (associated with time of intervention and standardised to a year)

that fit onto a QALY-like cardinal scale of 0 (bad health) to 1 (best health)(22) This will allow

for an incremental unit of intervention effects (or benefits) to compare with monetary costs

in order to statistically quantify a monetary amount for combined cost-effectiveness.

Statistical/sensitivity analysis

9

The study will use an incremental cost effectiveness ratio (ICER)(23) to define a monetary

amount for each QALY-like unit of benefit gained as represented in the following equation:

Cost of the intervention (Ca) [MINUS] cost of control (Cb) (this case, $0)

[DIVIDED BY]

Difference in the psychological domain scores for the intervention (Ea) [MINUS] Difference in

the psychological domain scores for the control (Eb)

An estimation process of non-parametric bootstrapping will be used in order to report an

accurate confidence interval (CI) of 95%(24) to represent any uncertainty in the cost-

effectiveness results computed by the ICER. We chose non-parametric bootstrapping as it is

widely applied in cost-effectiveness analyses, to allow for computational resampling of the

original data to develop an “empirical estimate of the sampling distribution of the ICER.”(24)

Presentation of results

Although we offer a unit of cost per benefit gained, there is no current threshold to denote

a willingness-to-pay for a QALY-like unit offered by the government in CAR, creating

uncertainty of how this cost-effectiveness result will place on a larger scale. UNICEF has a

working budget (fluctuating with donations) for humanitarian aid in CAR(25), but this is for all

causes, and with the influx of released child soldiers to support, it is difficult to define a

worthy threshold. We hope this cost-utility analysis will be useful for decision-makers as the

situation progresses. Variability of individual Reponses to treatment must be accounted for

when preparing the decision(24). Also, the parameters offered above (with a standardised life

year) is based upon the assumption that the treatment will have lasting effects outside of

the time observed. It is unknown when the conflict will desist, and this brings further

10

uncertainty to how well the benefits of the treatment will intervene with on-going stress.

Time-horizons should also be considered. It is uncertain what environment and priority

changes this conflict zone might bring, and the possibility of a sudden change in available

funds and staff due to reallocation for an emergency (i.e. mass bombing requiring mass

population first aid) could affect the intervention process.

Placing the results in context

As mentioned above, UNICEF may not be able to afford this intervention due to limited

funds currently allocated for expenses in CAR, lack of strict funds for child soldier

reintegration, fluctuating amounts dependent on donations, and uncertainty of future

disasters that may absorb funds and staff being used for child soldier reintegration. These

results are also not generalizable to the international population of child soldiers because of

the varying response to intervention being related to cultural standards (i.e. Nepali soldiers'

reintegration success also depends on their caste(6)).

11

References

1. Fund UNICsE. Paris principles. New York, NY; 2007.

2. Kohrt BA, Yang M, Rai S, Bhardwaj A, Tol WA, Jordans MJ. Recruitment of child

soldiers in Nepal: Mental health status and risk factors for voluntary participation of youth

in armed groups. Peace Confl. 2016;22(3):208-16.

3. Pine DS, Costello, J., and Masten, A. . Trauma, proximity, and developmental

psychopathology: the effects of war and terrorism on children. Neuropsychopharmacology.

2005;30:1781-92.

4. Schauer EaE, T. The psychological impact of child soldiering. In: Martz E, editor.

Trauma rehabilitation after war and conflict: Springer Science+Business Media, LLC; 2010. p.

311-60.

5. Betancourt TS, Agnew-Blais, J., Gilman, S.E., Williams, D.R., & Ellis, B.H. Past horrors,

present struggles: The role of stigma in the association between war experiences and

psychosocial adjustment among former child soldiers in Sierra Leone Social Science &

Medicine. 2010;70(1):17-26.

6. Adhikari RP, Kohrt, B.A., Luitel, N.P., Upadhaya, N., Gurung, D. & Jordans, M.J.D.

Protective and risk factors of psychosocial wellbeing related to the reintegration of former

child soldiers in Nepal. Intervention. 2014;12(3):367-78.

7. Brown FL, de Graaff, A.M., Annan, J., & Betancourt, T.S. Annual Research Review:

Breaking cycles of violence–a systematic review and common practice elementsanalysis of

psychosocial interventions for children andyouth affected by armed conflict. The Journal of

Child Psychology and Psychiatry. 2016.

8. UNICEF. News note: Hundreds of children released by armed groups in Central

African Republic 2015 [updated 14/05/2014. Available from:

https://www.unicef.org/media/media_81890.html.

9. Random.org. Sequence generator. 2017.

10. O'Callaghan P, Branham, L., Shannon, C., Betancourt, T.S., Dempster, M., &

McMullen, J. . A pilot study of a family focused, psychosocial intervention with war-exposed

youth at risk of attack and abduction in north-eastern Democratic Republic of Congo. Child

Abuse & Neglect. 2014;38(7):1197-207.

12

11. Carolina MUoS. TF-CBTWeb: a web-based learning course for trauma-focused

cognitive behavioral therapy 2005 [Available from: https://tfcbt.musc.edu/.

12. Prince MJ. Cross-cultural research methods and practice. In: Patel V, Minsa, Harry,

Cohen, Alex, and Prince, Martin, editor. Global mental health: principles and practice:

Oxford University Press; 2013.

13. BBC. Central African Republic country profile 2016 [updated 21/11/2016. Available

from: http://www.bbc.co.uk/news/world-africa-13150040.

14. LLC C. Sample size calculator. 2017.

15. Betancourt TS, Bass, J., Borisova, I., Neugebauer, R., Speelman, L., Onyango, G., &

Bolton, P. Assessing local instrument reliability and validity: a field-based example from

northern Uganda. Social Psychiatry and Psychiatric Epidemiology. 2009;44(8):685-92.

16. Betancourt TS, Yang, F., Bolton, P., & Normand, S.L. Developing an African youth

psychosocial assessment: An application of item response theory. International Journal of

Methods in Psychiatric Research 2014.

17. Organization WH. The World Health Organization Quality of Life 2012 [Available

from: http://www.who.int/mental_health/publications/whoqol/en/.

18. Organization WH. WHOQOL-SRPB Field-Test Instrument. 2002.

19. Shearer J, McCrone, P., and Romeo, R. Economic Evaluation of Mental Health

Interventions: A Guide to Costing Approaches. PharmacoEconomics. 2016;34(7):651-64.

20. Selvam S, G. Scaffoldings: training young people in christian life skills Nairobi2011

[Available from: http://www.sahayaselvam.org/2011/08/11/scaffoldings-training-young-

people-in-christian-lifeskills/.

21. Foundation ER. EQ-5D 2017 [Available from: http://www.euroqol.org/faqs/eq-5d-

5l.html.

22. Salize HJ, Jacke, C., Kief, S., Franz, M., & Mann, K. Treating alcoholism reduces

financial burden on care-givers and increases quality-adjusted life years. Addiction.

2013;108(1):62-70.

23. Fox-Rushby JaC, John. Economic evaluation: McGraw-Hill Education; 2007.

24. Drummond MF, Sculpher, M.J., Claxton, K., Stoddart, G.L., and Torrance, G.W..

Methods for the economic evaluation of health care programmes. 4 ed: OUP Oxford; 2015.

25. UNICEF. Central African Republic: country program document 2012-2016. 2011.

13