L. T. Synan1,3, M. A. Eid1,4,5, C. R. Lamb1, S. L. Wong1,2 1Dartmouth-Hitchcock Medical Center,Department Of Surgery,Lebanon, NEW HAMPSHIRE, USA 2Geisel School of Medicine at Dartmouth,Department Of Surgery,Hanover, NEW HAMPSHIRE, USA 3Geisel School of Medicine at Dartmouth,Hanover, NEW HAMPSHIRE, USA 4White River Junction VAMC,VA Quality Scholars Program,White River Junction, VERMONT, USA 5White River Junction VAMC,VA Outcomes Group,White River Junction, VERMONT, USA
Introduction:
Ratings websites offer consumers the ability to compare products and services easily. Patient reviews/recommendations of hospitals are unsolicited and less structured compared to the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) surveys. Here, we evaluated ratings on 3 widely used consumer ratings websites, HCAHPS scores, and their correlation with patient safety in a national cohort of academic medical centers.
Methods:
We identified 356 US acute care teaching hospitals that sent 5,000-10,000 HCAHPS surveys in 2018, representing medium to large-sized centers. We collected patient satisfaction results for each hospital (based on the “Would you recommend” question) and post-surgical patient safety indicators (PSIs) from the publically available Hospital Compare website. We also collected hospital star ratings (1–5) from Facebook (n=281 hospitals; mean 375 reviews/hospital), Google (n=354; mean 326 reviews/hospital), and Yelp (n=295; mean 82 reviews/hospital). Differences in ratings were compared with paired t-tests. Concordance between ratings websites, HCHAPS scores, and a composite PSI were assessed with Pearson’s correlation coefficient.
Results:
Mean Facebook ratings (3.8, IQR 3.6-4.1) were higher than those found on Google (3.2, IQR 2.7-3.5) or Yelp (2.5, IQR 2.0-3.1), with statistically significant pairwise comparisons (p<0.001). Each exhibited positive correlation with the HCAHPS “Recommend Hospital” score (p<0.001). Yelp correlation strength (r = 0.54) was higher than Google (r = 0.51) and Facebook (r = 0.45). A linear regression model further demonstrated that Yelp (r2 = 0.29) was more strongly correlated with HCAHPS than Google (r2 = 0.26) or Facebook (r2 = 0.20) (Figure). No consumer-driven ratings platform exhibited correlation with the composite post-surgical PSI.
Conclusion:
There appears to be wide variation between platforms in consumer ratings of hospitals, and ratings on Facebook are significantly more favorable than on Google or Yelp. These are independently correlated with HCAHPS scores. While the ratings may be informed by different variables, these findings suggest that unstructured consumer reviews generally reflect similar directionality as HCAHPS satisfaction scores, particularly patients’ willingness to recommend the hospital to others. It is not clear why Yelp reviews are less favorable, but the difference across platforms may be important for users to consider. Further, the consumer ratings platforms are not correlated with the composite post-surgical PSI, so online ratings may not reflect the safety of surgical care received.
