What providers should know about patients who post healthcare reviews online

When dissatisfied patients slam provider entities in online reviews, they tend to focus on administrative frustrations and thwarted requests that they, the patients, felt were reasonable to make. These complainants also tend to use a lot of words. 

By contrast, satisfied healthcare consumers write quick, list-like rundowns of things that pleased them. Their online reviews tend to recognize the competence and caring of providers and their staffs. They also tend to give shout-outs for good communication and care coordination. 

The patterns are brought to light in a study conducted by multidisciplinary researchers at the University of Pennsylvania. JAMA Network Open published the team’s scientific report Aug. 1. 

PhD student Neil Sehgal, Anish Agarwal, MD, MPH, and colleagues analyzed 1.1 million consumer reviews posted to a single social-media platform—Yelp—over a seven-year period. The reviews covered more than 138,000 provider organizations in 20 categories—everything from hospitals and urgent care clinics to speech therapy offices and dialysis centers. 

The researchers looked at star ratings (1 star = lowest, 5 stars = highest) and examined correlations between those scores and patients’ open-ended review narratives. 

Balanced study sample 

As it happened, the study set was fairly evenly divided along the happy vs. unhappy continuum. Almost half the reviews, 46.3%, were overall negative while just more than half, 50.1%, were positive. 

Among the team’s key findings: 

• Negative patient experiences frequently centered on quality of communication and administrative issues. Negative feedback centered on unmet expectations. 
   
• Positive reviews emphasized supportive staff interactions. 
   
• The word “not” was most frequently correlated with negative ratings, whereas “and” was most commonly correlated with positive ratings. 
   
• Among 200 topics on which patients commented, the strongest negative correlations involved payment issues and poor treatment.
   
• The strongest positive correlations involved kindness and anxiety relief. 

In their discussion, Sehgal and co-authors take up the question of how providers might use the findings to improve their patient-satisfaction scores. 

“Incorporating real-time online-review data into existing quality-improvement frameworks,” they remark, “could help clinicians, administrators and policymakers identify emerging concerns, monitor patient sentiment and tailor interventions that enhance patient-centered care across diverse health care settings.” 

And has it all over not 

The analysis of patients’ free-text feedback went further. Using machine learning and natural language processing, Sehgal and colleagues looked at linguistic features, topic modeling and n-gram associations. 

They reiterate that the word “not” was the most strongly correlated n-gram with negative reviews. 

For example, a 1-star review stated, “Despite calling numerous times, I was never given a quote so I had no idea what to expect. On arriving at the hospital on the day of the procedure, I was told that I could not check in since my insurance was not verified.”

As for the word “and” being most closely correlated with positive reviews, the authors give as an example a 5-star review that described “a great experience [I had] during a procedure. Everyone was very attentive, nice and helpful.”

‘Enablers of clinical excellence’ 

Drilling deeper into the potential applicability of their findings to real-world settings, Sehgal et al. advise providers against seeing improvements in patient communications as standalone goals. Instead, they suggest, such betterments should be viewed as “enablers of clinical excellence.” 

“When patients are informed and respected, they can be more likely to follow treatment plans and report positive outcomes,” the researchers write. “Conversely, administrative inefficiencies may serve as early warning signals of deeper structural issues that could indirectly compromise clinical outcomes.” 

The team calls for future research to investigate, in depth, the interplay between these factors and others. 

Such research, they add, might examine “how experiences captured in online reviews relate to care quality metrics such as diagnostic accuracy, treatment adherence or hospital readmissions.”

The study is available in full for free.