Application of Conjoint Analysis to Preferences of Low *

Kaiser Foundation Research Institute, Oakland CA

Investigators

Abstract

DESCRIPTION: The proposed study uses conjoint analysis (CA) to elicit preferences of a sample of HMO low back pain patients for chiropractic and primary care visits that vary by attributes of process and health outcomes. Study subjects identified through administrative data will receive a mail questionnaire, the core of which is a set of 10- 12 hypothetical choices between visits that vary by specified attributes of outcome and process, including cost. Responses will be analyzed using probit regression based on random utility theory. Reliable and generalizable economic evaluations of therapies in complementary and alternative medicine (CAM) such as chiropractic are needed to inform resource allocation decisions of health insurers and payers. Since most benefits from CAM are quality of life improvements rather than life extension, appropriate utility estimates will be crucial. We hypothesize that assessments of "process"? utility will be especially important in the CAM context, based on considerable research suggesting that therapeutic benefits from CAM (e.g., reassurance, information) beyond direct health outcomes. CA may be potentially quite valuable in economic evaluations of CAM because in theory it can produce both utility and willingness to pay (WTP) estimates that incorporate process as well as final outcome. The specific aims of this study are to (1) conduct a feasibility study of conjoint analysis as a method of eliciting preferences over chiropractic and conventional treatment of acute low back pain among a population of approximately 400 adult members of Kaiser Permanente Northwest diagnosed with acute low back pain, focusing on assessments of validity and reliability, including respondent burden and internal consistency; (2) produce preliminary utility estimates that incorporate both process and outcome attributes, for eventual use in cost-utility analyses of low back pain treatment within the HMO context; and (3) produce preliminary WTP estimates for process attributes of low back pain treatment for eventual use in cost-benefit analyses within the HMO context. This study will generate useful preliminary data that informs the future development of benefit assessment methods that are appropriate for economic evaluations of chiropractic and that are sensitive to both process and outcome. We expect that lessons from this study will readily generalize to future economic evaluations of other therapies in CAM as well as in conventional medicine, especially those delivered in the managed care context.