Health Services Researchers' Use of the Dartmouth Atlas of Health Care: A Literature Review and Network Analysis 1996-2023.

The Dartmouth Atlas of Health Care published data on regional variations in healthcare for Medicare fee-for-service beneficiaries from 1996 to 2022. Despite its revolutionary impact on health services research and policymaking, to our knowledge, no comprehensive review has been undertaken to systematically study how the Dartmouth Atlas has been used by the research community. We identified all 8507 publications that cited the Dartmouth Atlas between 1996 and 2023. With this corpus, we conducted authorship network analyses and conducted a content analysis of a 5% random sample (n = 403). Network analyses showed a high degree of co-author clustering and little cross-community co-authorship. Content analysis showed that most publications had no author affiliation with Dartmouth (88%, 95% CI = 0.85-0.91), were quantitative observational studies (51%, 95% CI = 0.47-0.57), and cited the Dartmouth Atlas without using its data (71%, 95% CI = 0.67-0.76). Among the 29% publications (n = 116) that directly used Dartmouth Atlas data, most (76%, 95% CI = 0.67-0.83) used the crosswalk tool to define geographic units, and many (45%, 95% CI = 0.36-0.54) used the data as variables in hypothesis testing. Most studies did not acknowledge the data limitations of the Dartmouth Atlas when using its data, though more acknowledged similar limitations for their overall study design that also involved Medicare fee-for-service data. Our results demonstrated the wide impact of the Dartmouth Atlas and suggested potential opportunities for improving data users' understanding of the critical assumptions in producing the Dartmouth Atlas data.