Taking Calculated Risks: What Do Patients Want?

Posted: 13 Jun 2007

See all articles by F. Reed Johnson

F. Reed Johnson

Research Triangle Institute

George Van Houtven

Research Triangle Institute - Center for Economics Research (CER)

Abstract

This session derives a conceptual framework for obtaining valid measures of risk-benefit tradeoff preferences, evaluates alternative methods of measuring such preferences, and reports maximum acceptable risk (MAR) estimates for several therapeutic areas. Stated-choice (SC) or conjoint methods offer a preference-elicitation format that can simulate realistic choices, allows nonlinearity in preference relations, and facilitates estimating MARs for a range of therapeutic benefits. Resulting preference estimates provide a full set of marginal rates of substitution for indirect utility function arguments such as cost, treatment efficacy, symptom duration, side effects, convenience and process factors, and adverse-event risks. Thus money equivalents or willingness to pay, time-tradeoff equivalents, and risk equivalents or MARs can be derived from the same underlying preference structure.

We report MARs for treatments in several therapeutic areas, including an inflammatory bowel disease, vasomotor symptoms related to menopause, and multiple sclerosis. We compare patients' willingness to accept adverse-event risks with clinical measures of actual therapeutic risks for the treatments. In each case, the survey instrument employs a standard risk-information treatment including pretested visual aids, choice-format preference-elicitation tasks, a two-stage question format with a forced-choice question followed by an opt-out or prefer-neither question, a D-efficient experimental design, multiple internal validity tests, and mixed-logit, hierarchical Bayes parameter estimation. As expected, patients' willingness to accept adverse-event risks depends systematically on the level of therapeutic benefit offered. MAR estimates for various therapeutic benefits range from nearly zero to mortality risks as high as 5%.

We compare our MAR estimates with published standard-gamble (SG) utilities. We find that SG utilities generally are larger than comparable MAR estimates. The probabilistic SG task provides less information for quantifying patients' level of risk aversion over a range of possible outcomes than SP tradeoff tasks. Moreover, the SG task is unlike actual treatment decisions and may not provide a valid basis for identifying clinically relevant net benefit thresholds.

In addition to concerns about potential legal liabilities associated with severe adverse events, decision makers may doubt whether patients are sufficiently knowledgeable to make risk-benefit judgments. Nevertheless, most patients in our samples provided coherent, consistent, and internally valid tradeoff information. The resulting MAR estimates are greater than actual risks for many clinically relevant outcomes. These results are consistent with anecdotal reports of patient protests when a drug is withdrawn from the market as a result of detecting serious adverse events in follow-up studies.

Effective risk management requires that physicians and patients adhere to strategies designed to reduce risks. Such strategies are likely to be less effective if patients' preferences are inconsistent with the risk-management strategy and are likely to result in patient dissatisfaction and potential conflicts between regulatory authorities and patient advocacy groups. Thus better understanding of the circumstances under which well-informed and consented patients are willing to accept adverse-event risks may help in designing more effective risk-management programs and increase the net benefits obtained from new health technologies.

Keywords: risk-benefit analysis

Suggested Citation

Johnson, F. Reed and van Houtven, George, Taking Calculated Risks: What Do Patients Want?. iHEA 2007 6th World Congress: Explorations in Health Economics Paper, Available at SSRN: https://ssrn.com/abstract=993573

F. Reed Johnson (Contact Author)

Research Triangle Institute ( email )

RTI-HS
P.O. Box 12194
Research Triangle Park, NC 27709
United States
919-541-5958 (Phone)
919-541-7222 (Fax)

George Van Houtven

Research Triangle Institute - Center for Economics Research (CER) ( email )

3040 Cornwallis Road
Research Triangle Park, NC 27709-2194
United States

Do you have negative results from your research you’d like to share?

Paper statistics

Abstract Views
554
PlumX Metrics