A cost-effectiveness analysis of an in-hospital clinical pharmacist service

Objective: A randomised controlled study performed from 2007 to 2008 showed beneficial effects of a composite clinical pharmacist service as regards a simple health status instrument. The present study aimed to evaluate if the intervention was cost-effective when evaluated in a decision-theoretic model. Design: A piggyback cost-effectiveness analysis from the healthcare perspective. Setting: Two internal medicine wards at Sahlgrenska University Hospital, Goteborg, Sweden. Participants: Of 345 patients (61% women; median age: 82 years; 181 control and 164 intervention patients), 240 patients (62% women, 82 years; 124 control and 116 intervention patients) had EuroQol-5 dimensions (EQ-5D) utility scores at baseline and at 6-month follow-up. Outcome measures: Costs during a 6-month follow-up period in all patients and incremental cost-effectiveness ratio per quality-adjusted life-year (QALY) in patients with EQ-5D utility scores. Inpatient and outpatient care was extracted from the VEGA database. Drug costs were extracted from the Swedish Prescribed Drug Register. A probabilistic analysis was performed to characterise uncertainty in the cost-effectiveness model. Results: No significant difference in costs between the randomisation groups was found; the mean total costs per individual +/- SD, intervention costs included, were (sic)10 748 +/- 13 799 (intervention patients) and (sic)10 344 +/- 14 728 (control patients) (p=0.79). For patients in the cost-effectiveness analysis, the corresponding costs were (sic)10 912 +/- 13 999 and (sic)9290 +/- 12 885. Intervention patients gained an additional 0.0051 QALYs (unadjusted) and 0.0035 QALYs (adjusted for baseline EQ-5D utility score). These figures result in an incremental cost-effectiveness ratio of (sic)316 243 per unadjusted QALY and (sic)463 371 per adjusted QALY. The probabilistic uncertainty analysis revealed that, at a willingness-to-pay of (sic)50 000/QALY, the probability that the intervention was cost-effective was approximately 0.2. Conclusions: The present study reveals that an intervention designed like this one is probably not cost-effective. The study thus illustrates that the complexity of healthcare requires thorough health economics evaluations rather than simplistic interpretation of data.