Industry-Specific Sources of Big Data Research Opportunities and Challenges: The Agency for Healthcare Research and Quality Data Bases and Software Tools

For researchers interested in evaluating industry-wide or “macro-level” organizational behavior issues, obtaining data can be much more challenging, than for “micro-level” researchers who can rely on obtaining the... [ view full abstract ]

For researchers interested in evaluating industry-wide or “macro-level” organizational behavior issues, obtaining data can be much more challenging, than for “micro-level” researchers who can rely on obtaining the cooperation of a few organizations, surveying several employees from each organization. In this era of “Big Data,” researchers more than ever before need to take advantage of existing publicly available data bases to help test important macro-level research questions. This study focuses on one set of such data bases from the hospital industry, illustrating both the opportunities and challenges in seeking to perform research over several years.
In today’s world of Big Data, there is perhaps no industry which has made available huge data bases to researchers than the healthcare industry. One of the primary sources of these data bases is the Agency for Healthcare Research (AHRQ) which is part of the U.S. Department of Health & Human Services. Through the Healthcare Cost and Utilization Project (HCUP, http://www.hcup-us.ahrq.gov/), researchers can acquire millions of healthcare claims each containing hundreds of data elements. These data bases include hospital in-patient and out-patient claims. Beyond providing access to these data bases, HCUP has also developed several software tools which enable researchers to assess various aspects of quality, including prevention quality indicators (PQIs), inpatient quality indicators (IQIs), patient safety indicators (PSIs), and pediatric quality indicators (PDIs).
Over the last decade, the combination of such data bases and such tools have provided an excellent opportunity for these researchers to test models assessing the relationship between organizational structure and quality outcomes. Unfortunately, the healthcare industry is a dynamic one in which change is a constant. As a result, the process of coding claims, delivering care, and administering a program as broad as HCUP has resulted in many updates of the Quality Indicator tools. For example since 2007, there have been at least 10 updates of the Patient Safety Indicator tool, some with only minor programming changes and some with the addition and exclusion of indicators. According to the administrators at HCUP the version of the tool used should always be the most recent, but for purposes of consistency it is more difficult for researchers to continually change tools. This study will assess the impact of this dynamic situation by comparing the results produced from the original 2007 tool with the most recent 2015 version, using a single year of claims. Results will inform researchers in the future when dealing with similar data base issues.