The effect of direct-to-consumer genetic tests on health care utilization and patient welfare

Presenter: Aslam H. Anis, University of British Columbia

Abstract

Background: The popularity and awareness of genetics has increased since the completion of the mapping of the human genome. Direct-to-consumer (DTC) genetic tests that provide consumers with a personalized assessment of their genetic risk for a list of diseases and conditions are now being offered over the internet.

Objective: To evaluate the impact of personal information obtained from DTC genetic test results on health care utilization and consumer welfare.

Methods: A web survey using conjoint analysis was conducted on a panel of Canadian adults. Mimicking the results provided by online DTC genetic tests, hypothetical scenarios were created comprising disease type, personal risk (relative to the average population), and quality of the research evidence on which the personal risk estimate was based. In each of 12 scenarios shown to participants, 4 diseases were selected from a group of 8. For each disease, average population risk, a personal risk level from 5 levels ranging from much lower than average to much higher than average (represented with absolute numbers) and a research quality level (low/medium/high) were selected. Respondents were asked to state whether or not they would (1)schedule an additional doctor’s visit, (2)change to a healthier lifestyle and (3)become more worried, based on each given risk profile. In the analysis, personal risk profiles were classified into one of 4 groups (very high/high/moderate/low), according to the proportions of personal risk levels that were higher than, equal to or lower than the average level. Diseases were grouped into preventable and non-preventable. Age, gender, self-reported general health and education level were modeled as covariates. Logistic regression models were fitted for each outcome.

Results: A total of 316 respondents (47% male, median age=52 yr, IQR:41-65) completed the survey. More respondents reported their health status as very good (33%), good (44%), or fair (11%) vs. excellent (8%) or poor (4%) and the proportion of respondents with high school, college and university education was 36%, 34% and 30% respectively. Of the 3782 scenarios included in the analysis, distribution of personal risk profile was 17% (very high), 16% (high), 48% (moderate) and 18% (low). Results from the analysis suggest that higher levels of personal risk within a scenario were associated with a higher likelihood of making an additional appointment with a doctor, adopting a healthier lifestyle, and feeling more anxious and worried. Respondents with poorer health were more likely to make an unscheduled doctor’s appointment. Females, older respondents and those with family histories of the diseases were also more likely to adopt a healthier lifestyle. Males and those with at least a university education were less likely to be worried. No significant difference was found by whether or not a disease is preventable or the quality of the evidence.

Conclusion: Our findings showed that personal genetic information increases demand for health care in the short run (increased physician visits) and there is some evidence that lifestyle modifications would also be made. When including the effect of anxiety, the overall impact on consumer welfare remains to be further investigated.