Chapter 3 76 Cost-utility The number and duration of outpatient treatment contacts and inpatient days will be calculated in cost in Euro’s. They are related with changes in quality of life measured with the EQ-5D (39). Cost-utility can be calculated by transforming scores on the EQ-5D into so called QALYs (49). To calculate cost-utility, only the cost of health care consumption is included, not social costs such as rent, benefits, etc. Blinding Thirteen of the above scales are self-report scales, and two (IMRS-clinician version and SES) will be rated by non-blinded clinicians who will not be involved in the IMR training. Granholm’s Goals Template will be rated by the blinded research assistants in an interview with the clinician. At the second and third measurements, the measurement procedure will be singleblinded: the research assistants will not know whether a client is participating in the experimental condition or the control condition; this procedure will preclude experimenter’s bias. To ensure this, clients and clinicians alike will be instructed before the interview not to inform the interviewer in any way about the client’s condition. At the end of each interview, interviewers will answer the question whether they have found out which condition the clients are in. Sample size On the basis of the effect sizes of the studies by Hasson-Ohayon et al. (2007), Levitt et al. (2009) and Färdig et al. (2011) (16–18), we anticipate a medium effect size of 0.40 on the primary outcome variable (self-rated IMR scale, Mueser et al. 2004, 2005 (23, 24)). On the basis of the power analyses with three times of measurement (mixed models), equal allocation to the experimental and control groups, a power of .80, alpha at 0.05, and an effect size of .40, it will be necessary to randomise 148 clients: 74 to the experimental condition and 74 to the CAU group (50, 51). Randomisation Because IMR is a relatively long-term care programme—in our pilot study at least 8 months at one session per week—the drop-out from treatment may be relatively high. Other studies had 29 % (17), 15 % (18); 51 % (20), and 72 % (15). Due to the 50 % drop-out rate expected from treatment in the experimental condition (a pessimistic estimate), we have chosen to allocate more clients to the experimental condition (IMR) than to the CAU group (proportions of 3:2). We will randomise 200 clients (on the basis of a conservative calculation): 120 to the experimental condition and 80 to the CAU group.
RkJQdWJsaXNoZXIy MjY0ODMw