Perspective and Costing in Cost-Effectiveness Analysis, 1974-2018.

OBJECTIVE

Our objective was to examine perspective and costing approaches used in cost-effectiveness analyses (CEAs) and the distribution of reported incremental cost-effectiveness ratios (ICERs).

METHODS

We analyzed the Tufts Medical Center's CEA and Global Health CEA registries, containing 6907 cost-per-quality-adjusted-life-year (QALY) and 698 cost-per-disability-adjusted-life-year (DALY) studies published through 2018. We examined how often published CEAs included non-health consequences and their impact on ICERs. We also reviewed 45 country-specific guidelines to examine recommended analytic perspectives.

RESULTS

Study authors often mis-specified or did not clearly state the perspective used. After re-classification by registry reviewers, a healthcare sector or payer perspective was most prevalent (74%). CEAs rarely included unrelated medical costs and impacts on non-healthcare sectors. The most common non-health consequence included was productivity loss in the cost-per-QALY studies (12%) and patient transportation in the cost-per-DALY studies (21%). Of 19,946 cost-per-QALY ratios, the median ICER was $US26,000/QALY (interquartile range [IQR] 2900-110,000), and 18% were cost saving and QALY increasing. Of 5572 cost-per-DALY ratios, the median ICER was $US430/DALY (IQR 67-3400), and 8% were cost saving and DALY averting. Based on 16 cost-per-QALY studies (2017-2018) reporting 68 ICERs from both the healthcare sector and societal perspectives, the median ICER from a societal perspective ($US22,710/QALY [IQR 11,991-49,603]) was more favorable than from a healthcare sector perspective ($US30,402/QALY [IQR 10,486-77,179]). Most governmental guidelines (67%) recommended either a healthcare sector or a payer perspective.

CONCLUSION

Researchers should justify and be transparent about their choice of perspective and costing approaches. The use of the impact inventory and reporting of disaggregate outcomes can reduce inconsistencies and confusion.