Cost-effectiveness of the Self-Help Plus Intervention for Adult Syrian Refugees Hosted in Turkey

Key Points Question Is the addition of the World Health Organization’s Self-Help Plus program—a group-based, guided, self-help psychological intervention—to enhanced usual care cost-effective compared with enhanced usual care alone for adult Syrian refugees or asylum seekers hosted in Turkey? Findings In this economic evaluation of 627 refugees or asylum seekers hosted in Turkey, the intervention group had a significantly better quality of life compared with the control group, at a cost of 6068 Turkish lira ($1147) per quality-adjusted life-year gained. Taking into account the stochastic uncertainty, the program had a 97.5% chance of being cost-effective. Meaning These findings suggest that Self-Help Plus is cost-effective as an intervention to prevent mental disorders in conflict-exposed refugees hosted in Turkey.


Estimation of quality adjusted life years by means of the EuroQol EQ-5D questionnaire:
The QALY estimates life years in complete health. This is estimated by multiplying the utility value corresponding to the participants' health state with the time lived in this health state. The utility value is ranging between 0 for the worst possible state of health (which means close to death) and 1 for the best possible state of health (which means full health). The time lived in a particular health state is defined by the questionnaire and depends on the frequency of measures. In our study a sixmonth time frame was used which means that the participant was asked to assess his or her health state for the last six months. To transform the measure into an annual health state, the utility value for the self-reported health state is multiplied by the fraction of the 12 months period which it represents. That means a utility value for a six-month period is multiplied by 0.5 to represent the QALY for a 12-month period. In case of repeated measures, the fractions of the estimated utility values will be summed up over the 12-month period 1 .

Statistical analysis of incremental costs utility
The task of health economics is to provide necessary information for an optimal allocation of health care resources [1].

Cost Intervention A -Cost Intervention B Δ Cost ICUR Utility Intervention A -Utility Intervention B Δ Utility ==
The ICUR is defined as the ratio between the cost differences and the utility differences of two that the alternative intervention is more expensive and less effective than the conventional or the non-intervention. If the ICUR is located at the upper right (north east) quadrant of the CEP the positive value indicates that the alternative intervention is more effective but also more expensive than the conventional or the non-intervention. If the ICUR is located at the lower left (south west) quadrant of the CEP the positive value indicates that the alternative intervention is less expensive but also less effective than the conventional or the non-intervention. If the ICUR is located at the lower right quadrant of the CEP the negative value indicates that the alternative intervention is less expensive and more effective than the conventional or the non-intervention. Obviously, a negative ICUR provides a much more clear decision basis than a positive one. If the ICUR is located at the upper left quadrant, the conventional or the non-intervention is more efficient than the alternative intervention and if the ICUR is located at the lower right, the alternative intervention is more efficient. Unfortunately, the ICUR mostly falls into the upper right quadrant. In this case an additional criterion is needed to decide whether an alternative treatment is efficient or not. This additional criterion is the maximal amount of money which one is willing to pay for an increase of the effect or the utility by one unit and is usually called maximum willingness to pay (MWTP).
The MWTP can be projected as a growth curve into the CEP and an alternative intervention will be assessed as efficient if the ICUR is located below this curve.
Estimation of the stochastic uncertainty of the ICUR is complicated by the fact that a ratio of differences between the study groups cannot be computed at the individual level but only at the level of the whole study sample. As a solution for this problem, bootstrapping techniques will be used to simulate the true sampling distribution of the ICUR 2, 3 . A further complication results from the fact that the ICUR (∆C/∆E) distribution is not defined for ∆E=0. As a consequence the confidence interval for the ICUR cannot be estimated based on parametric assumptions about the mean and the SD 2 . The 95% confidence interval estimated for the ICUR represents the fraction of 95 % of the bootstrap samples defined by cutting the highest 2.5 % and the lowest 2.5 % of the ICUR values simulated by the bootstrap resampling 1 .
The cost-effectiveness acceptability curve (CEAC) is the common technique to visualize this criterion 1,2 . At the horizontal axis the CEAC shows potential values for MWTO in an increasing order, the vertical axis shows the percentages of the estimated ICUR values which are located below the MWTP curve. Similar as the statistical confidence interval the CEAC indicates at which MWTP a particular percentage (eg 90%) of the estimated ICUR fall below the MWTP curve.