Hospital Online Reviews: What They Can and Cannot Tell You

Reviews are not noise

For years, many hospitals treated online reviews as complaints from a loud few. The research says something more useful. Public reviews carry a real signal about patient experience. They also carry distortions that are easy to miss. A hospital that understands both can learn from what patients write without being misled by it.

We read reviews through two mirrors. What an institution says about itself is one. What people live is the other. The image is the distance between them. Reviews are one of the few places where the second mirror is written down in public, which is why they deserve a careful reading rather than a defensive one.

What the research found

The most useful studies compare review sites with formal surveys, which use sampling, standard questions and statistical adjustment. In the United States the formal instrument is HCAHPS, the national survey of hospital patient experience. In England it is the NHS inpatient survey.

Reviews track surveys, imperfectly. Bardach and colleagues examined 3,796 US hospitals that reported HCAHPS scores. Only 962, about one in four, had Yelp ratings at all. Among hospitals with more than five Yelp ratings, the share of high Yelp ratings correlated with the HCAHPS share at 0.49. That is a moderate relationship, not a match.

The pattern holds in another health system. Greaves and colleagues studied 9,997 ratings posted on the NHS Choices website for 146 acute hospital trusts in England during 2009 and 2010. The ratings correlated with national survey measures, with coefficients from 0.31 to 0.49. Cleanliness showed the closest agreement.

Some link to outcomes appears. In both studies, better online ratings went with lower mortality or readmission rates. The correlations were weak, and the authors present them as associations, not causes. In the English data, these links were at least as strong as those for the formal survey.

Reviews cover ground surveys do not. Ranard and colleagues used language analysis on 16,862 Yelp reviews of 1,352 US hospitals. They found 12 domains of experience that HCAHPS does not measure, including the cost of the visit, insurance and billing, compassion of staff, and care of family members. In a separate close reading of 244 online reviews, Bardach's team found that only 57 percent mentioned any HCAHPS domain. Perceptions of safety appeared in 41 percent, and billing or unexpected costs in 11.5 percent. Friends and family wrote 38 percent of those reviews, a view that patient surveys do not capture.

Where reviews mislead

Each of these studies names its limits. The limits matter more than the correlations.

• Self-selection. People choose to write. Those with a striking experience, good or bad, are more likely to. In the Ranard data, 31 percent of reviews gave one star and 33 percent gave five.
• Small counts. The median hospital in that study had four reviews. Four voices cannot describe a hospital that treats thousands. Bardach's team limited its main comparison to hospitals with more than five ratings, and Ranard's team found agreement with HCAHPS grew stronger as the minimum number of reviews rose.
• No denominator. A review shows that something happened to someone. It does not show how often. Ten complaints about billing mean one thing among a hundred patients and another among a hundred thousand.
• Fake and incentivised reviews. Not every review is genuine. Ranard's team excluded hospitals with a high share of reviews flagged by Yelp's own filters. In 2024 the US Federal Trade Commission adopted a rule that bans fake reviews, paying for reviews of a set sentiment, undisclosed insider reviews and review suppression.
• Not representative. Reviewers are not a sample of patients. Their views cannot be read as the opinion of a city, a country or a community. They are a record of what some people chose to say in public.

None of this makes reviews worthless. It makes them one source among several, read with their weaknesses in view.

What reviews add

Formal surveys are built for comparison. They ask the same questions everywhere, on a fixed schedule. That is their strength. It is also what they cannot see past.

Patient experience online reviews fill three gaps. First, topics: billing, parking, waiting beside a relative, the tone of a phone call. Surveys rarely ask. Patients raise them anyway. Second, timing: a review can appear days after a visit, while survey results often arrive months later. Third, public image: reviews are what prospective patients actually read. Whatever their accuracy, they shape how the institution is seen.

That last point is often missed. A survey tells a hospital how its patients rated it. Reviews show how the hospital appears to everyone else. Hospital reputation management that starts from star averages ends in arguments about averages. Start from what people describe.

A method for reading them

Good patient reviews analysis is mostly discipline. Six habits make the difference.

• Read themes, not stars. An average rating hides its reasons. Group what people write by subject: communication, waiting, cleanliness, billing, safety. A worsening theme tells you where to look. A falling average does not.
• Keep the source attached. Every finding should link back to the public posts it came from. A claim without its source cannot be checked, and a claim that cannot be checked should not drive a decision.
• Compare what you publish with what people report. Set the hospital's own statements beside public mentions. If the website promises short waits and reviews describe long ones, that gap is the finding.
• Track change over time. One month of reviews is an anecdote. The same theme rising across a year is a signal. Measure against your own history before comparing with anyone else.
• Study patterns, never people. Analyse topics across many posts. Do not profile, track or pursue individual reviewers. The aim is to understand the institution, not to monitor its critics.
• Respond in public with care. Thank the writer, acknowledge the concern and offer a private channel. Never confirm that someone was a patient or mention their care. In 2022 the US Department of Health and Human Services settled with a dental practice that had disclosed patient information in replies to reviews.

"Providers cannot disclose protected health information of their patients when responding to negative online reviews." Melanie Fontes Rainer, Director, HHS Office for Civil Rights

Read this way, reviews complement surveys and clinical data. They do not replace them, and alone they prove little. When the evidence is thin, the honest answer is that it is thin.

How Mirror approaches this

Mirror reads what an institution publishes and what people say about it in public, then sets the two side by side. Each gap it shows carries its public source, so anyone can open the original and judge it. It reads themes across many posts rather than following individuals, and it marks where the evidence is too thin to support a conclusion.

Sources

1. Yelp Reviews Of Hospital Care Can Supplement And Inform Traditional Surveys Of The Patient Experience Of Care (Ranard BL, Werner RM, et al.), Health Affairs 35(4):697–705. Health Affairs, 2016.
2. The relationship between commercial website ratings and traditional hospital performance measures in the USA (Bardach NS, Asteria-Peñaloza R, Boscardin WJ, Dudley RA), BMJ Quality & Safety 22(3):194–202. BMJ Quality & Safety, 2013.
3. Associations between internet-based patient ratings and conventional surveys of patient experience in the English NHS: an observational study (Greaves F, Pape UJ, King D, et al.), BMJ Quality & Safety 21:600–605. BMJ Quality & Safety, 2012.
4. From the closest observers of patient care: a thematic analysis of online narrative reviews of hospitals (Bardach NS, Lyndon A, et al.), BMJ Quality & Safety 25(11):889–897. BMJ Quality & Safety, 2016.
5. FTC Issues Final Rule on Consumer Reviews and Testimonials. Morgan Lewis, 2024.
6. HHS Civil Rights Office Enters Settlement with Dental Practice over Disclosures of Patients’ Protected Health Information. U.S. Department of Health and Human Services, 2022.

Comments

Your nameYour comment

Comments appear after our team reads them. Please do not include patients' or colleagues' personal details.

Post comment