The HIPAA authorization form and effects on survey response rates, nonresponse Bias, and data quality A randomized community study

Background: There has been speculation that the Health Insurance Portability and Accountability Act (HIPAA) has muted participation in research but little direct evidence to substantiate those claims exists. Objective: To evaluate the effects of including a HIPAA authorization form (HAF) on multiple measures of survey performance. Research Design, Subjects: A community survey of Olmsted County, Minnesota residents conducted between September 2005 and April 2006. A total of 6939 cases were randomly assigned to 2 experimental conditions where half of the subjects received a 1-page HAF (n = 3469) and the other half did not (n = 3470). Main Outcome Measures: Response rate, nonresponse bias, and data quality. Results: At the end of data collection, a response rate of 39.8% was observed in the HAF condition and 55.0% in the No HAF condition (P < 0.0001). There was a negligible but statistically significant (P ≤ 0.001) over-representation of males in the No HAF condition but no difference in the amount of missing data between the 2 groups. However, reports of general health and the percentage of respondents indicating that they were nonsmokers were both significantly (P ≤ 0.01) lower in the No HAF condition than in the HAF condition. Conclusions: Inclusion of a minimally burdensome version of the HAF reduced survey response rates by up to 15 percentage points. This could have implications for a study's statistical power. There was little evidence that the form affected nonresponse bias or data quality.