Conjoint analysis (healthcare)

See also: Conjoint analysis (in marketing), Rule Developing Experimentation

Why conjoint in healthcare market research?

Pharmaceutical manufacturers need deeper and deeper market information they can rely on to make the right decisions and to identify the most promising market opportunities^[1][6] . They can obtain great benefits from understanding physicians’ prescription preferences and opinions, as these are the key people in prescribing medical treatments. Consequently, conjoint analysis research projects might represent a remarkable help for pharmaceutical companies during the development of new drugs, and if properly conducted they might even allow the estimation of product sale and market share^[7]. Conjoint analysis^{[4] [5]} is a multivariate statistical technique based on the study of the joint effects on consumers of the elements that compose a product or service; it allows eliciting the relative importance of such elements. Therefore, by an additive model, it allows the estimation of the total utility of different profiles or combinations of attributes, and consequently, it allows the identification of the optimal configuration for a new or existing product or service. This technique was first developed in the early 1970s by Green and Rao^[3] . Since then, it has received increasing academic and private sector attention.

Traditional conjoint models

One of the most important statistical problems concerning conjoint studies lies in the design. The choice of the design depends strongly on the characteristics of the study, such as the number of variables involved, the potential presence of interaction effects among the variables, the ability and the motivation of the target respondents to the conjoint survey, the statistical ability of the researcher, and last, but not least, the availability of software. For the time being, healthcare researchers have been used most of the widespread conjoint collection models (i.e., full profile, partial/incomplete profile), designs (i.e., full factorial designs, fractional factorial designs, resolution III designs, etc), and evaluations approaches (i.e., ranking, rating, choice from a set, paired comparison) available in the statistical literature.

Prescription-based conjoint model

These models are adequate when respondents’ evaluations apply to products or services to be used by themselves (FMCG, Durable goods, financial products, etc). However, for some projects, respondents have to provide evaluations for products or services to be used by a group of people, that is the case of physicians with respect to their patients, when they are asked to assess the preference for a new or existing drug or medical treatment. The healthcare researcher wants to assess what physicians will do for a number of patients. More precisely, he/she intends to elicit the preference for a new treatment and to estimate in the most precise way the related preference share that eventually will lead to the estimation of the market share for such a treatment. Consequently, a prescription-based model^[2] seems to be an appropriate tool for this research situation. Basically, it consists in exposing physicians to several treatments profiles at-a-time, where all current treatments appear next to the new treatment (conjoint task). Physicians are then asked to assign some points to each treatment in the task. More precisely, in this exercise they are asked to think of the next 100 patients with the disease of concern and to record the number of prescriptions that they would prescribe for each of the treatments outlined in the task presented. Consequently, all treatment assignments in the task add up to 100 points or more. More than 100 points are allowed for projects involving treatments that could be co-prescribed. This framework is especially appropriate for a deep understanding of the relationships between the current medical treatments for a particular disease and different definitions for the product to be launched.

See also

• Conjoint analysis (in marketing)
• new product development
• marketing research
• quantitative marketing research

External sources

^[1] Chakraborty, G., R., Ettenson, & G., Gaeth. 1994. How consumers choose health insurance. “Journal of Health Care Marketing”, 14, 21–23.
^[2] Furlan R., Corradetti R. (2006), “Aspects of Experimental Design in the Prescription-Based Conjoint Analysis Model”, Proceedings of ENBIS 6, 2006, Wroclaw, Poland.
^[3] Green P. E., Rao V. R. (1971) Conjoint measurement for quantifying judgemental data, “Journal of Marketing Research”, 8, 355-63.
^[4] Green P. E., Srinivasan V. (1978) Conjoint analysis in consumer research: issue and outlook, “Journal of Consumer Research”, 5, 103-123.
^[5] Gustaffson A., Herrmann A., and Huber F. (2001) Conjoint analysis as an instrument of market research practice, in: “Conjoint Measurement: Methods and Applications”, Gustaffson A., Herrmann A., & Huber F. (Eds.), Berlin: Springer, 5-46.
^[6] Scottish Office Department of Health. 1992. “The patient’s charter: what users think”. Tech. rept. HMSO, Edinburgh.
^[7] Sculpher, M., Bryan, S., Fry, P., DeWinter, P., Payne, H., & Emberton, M. 2004. Patients’ preferences for the management of non-metastatic prostate cancer: discrete choice experiment. “British Medical Journal”, 328, 382–385.