Recognition of inherent biases in administrative data

[摘要] Administrative data are valuable tools for primary care research. Using routinely collecteddata avoids the burden of data collection for bothresearchers and research subjects, making many studiesmore feasible and pragmatic. Additionally, using administrative data can help avoid biases inherent in surveyresearch, including recall bias, response bias, socialacceptability bias, and sampling bias.1 Administrativebilling and dispensing data can be used to exploreservices provided in the health care system from bothprovider and user perspectives. Understanding the healthof the population and how the population uses thehealth care system provides essential information forhealth system planning and policy development.While whole population studies avoid some biases,they introduce others. This does not negate the valueof administrative data research, but it is imperative thatwe understand these biases and implement appropriate responses that consider the importance of inclusion,diversity, equity, and accessibility in research.

[学科分类] 卫生学