Health RegData
Why research the healthcare industry?
The healthcare industry is typically highly regulated in the national and subnational jurisdictions that the QuantGov team has analyzed. Healthcare regulations define everything from the incredibly complex web of insurance regulations to the highly specific approval processes for all manner of drugs and medical devices. Without a doubt, these regulatory restrictions have a major effect on the healthcare industry. Many organizations and individuals - such as the healthcare policy team and the QuantGov team at the Mercatus Center - are intent on studying the relationship between healthcare regulations and outcomes such as the supply of medical devices and rate of innovation in the healthcare industry.
Currently, the QuantGov team has two data series surrounding the healthcare industry. The first is the identification of U.S Federal healthcare regulations. In this data series, researchers trained a custom machine learning algorithm to identify relevant healthcare regulations. The output is a probability that a piece of regulatory text is related to healthcare. Data is currently available for each title and part combination in the Code of Federal Regulations for the year 1970-2019. Additional metadata surrounding the identified documents include restriction counts, word counts, complexity metrics, and industry classification.
The second data series is data from the identification and quantification of U.S. State healthcare regulations. Over 24,000 labeled training documents were used to create a custom algorithm that identifies healthcare regulations at the state level. All states quantified in the State RegData 2.0 project were examined and almost all of these states have two separate years of data. This data series also has additional metadata including restriction counts, word counts, complexity metrics, industry classification, and a probability that a regulation pertains to the healthcare industry.
For both of the data series, federal and state, the custom machine learning algorithms are intended to strictly identify regulations that pertain to the provision of healthcare, of healthcare professionals, drugs for human use, medical products, and healthcare research. However, the current iterations of both algorithms do pick up on some regulations that discuss very similar topics including occupational health and safety, the environment, and social assistance. These regulations typically affect health outcomes, but not the provision of healthcare services.
