Looking for a restaurant recommendation? A hair salon? A tailor?

Online review sites such as Yelp have become the go-to places to research so many kinds of businesses.

But what about hospitals? Can you click your way to the best hospital for you?

A new study out of Indiana University suggests that when it comes to patient experience, crowdsourced review websites are a decent barometer. But although Yelp may be able to tell you how friendly the staff was or how responsive the billing office was, these reviews won't do much to help you gauge other important qualities, such as patient safety and quality of care.

Still, patients are increasingly turning to social media and crowdsourced reviews to inform their decisions about where to go for care. That — as well as increased attention by Medicare to patient surveys —  is changing how hospitals factor patient opinion into their business models.

Hospitals are hiring patient experience officers and building up social media teams to use such platforms to promote their brand, fend off negative publicity, and learn from feedback they may not be receiving directly through more formal patient surveys.

"The social sentiment can be extremely powerful when it comes to health-care decision-making, when it comes to health-care reputation," said Dwight McBee, the chief experience officer at Temple Health. "There's a lot of power in word of mouth, and we see that playing out in the online environment."

‘There’s no great tool’

Researchers at Indiana University aimed to compare hospital ratings from Yelp, Facebook, and Google against the U.S. Centers for Medicare and Medicaid Services' Hospital Compare, a federal data-driven evaluation of hospital performance based on self-reported metrics and patient surveys.

Review sites, such as Yelp, are known for their hot and cold entries from people who  had either a great or terrible experience, not so much for middling reports. That could make them unreliable.

"We typically think when people write about other kinds of businesses, it's because they had an extreme experience — either they loved it or they hated it," said Victoria Perez, an assistant professor at Indiana University's School of Public and Environmental Affairs, who co-authored the study.

But researchers found that hospitals' crowdsourced rating lined up pretty well with their patient experience rating in Hospital Compare, Perez said.

The highest-ranked hospitals on crowdsourced sites were also the highest ranked in Hospital Compare's patient experience ratings 50 percent to 60 percent of the time, according to the study published in the peer-reviewed journal Health Services Research.

The federal patient survey asks questions about how well medical staff communicated with patients, whether they adequately explained medication and recovery instructions, and how quickly staff responded to requests for help — all experiences often described in Yelp reviews, too.

But researchers also found that 40 percent of the time, the hospitals with the highest crowdsourced ratings had among the lowest scores for patient safety and quality in Hospital Compare, such as pneumonia death rate and the number of patients who returned to the hospital within 30 days.

"When we think about what is a good hospital or a bad hospital, it's easy to get distracted by amenities that make your experience better," Perez said. "People may think those things are correlated, but we find, in fact, they're not correlated at all."

Hospital Compare may be a better source for information about important features such as infection rates, but the tool can be cumbersome for users who aren't sure how to navigate the dozens of metrics measured.

"There's no great tool. The lesson in all this is there's difficulty with all these reviews and comparison tools. There isn't a one-size-fits-all, but it's important for patients to do their homework," said Michael Consuelos, senior vice president of clinical integration for the Hospital and Healthsystem Association of Pennsylvania.

Using data from   Hospital Compare, the association developed its own database of Pennsylvania hospitals, called Care in Pennsylvania, where patients can compare hospital performance in specific areas, such as cardiac care or obstetrics

Hospitals use Yelp, too

Still, the trend of social media and online reviewing shows no sign of ending.

Main Line Health's social media team is constantly scanning such sites as Twitter and Facebook, as well as Yelp and other review sites, for patient comments. They respond to patients as quickly as they can — even if only to say "we're looking into it" — and relay comments to the relevant care units.

The real-time responses ensure that patients know that  their concerns are being taken seriously. More important, listening to patients is crucial to safety and quality, said Barbara Wadsworth, senior vice president of patient services and chief nursing officer at Main Line.

"You can have the best doctor or best nurse, but if they're not kind to the patient and the patient doesn't feel comfortable asking questions," they may not get the care they need, Wadsworth said.

Penn Medicine decided to "get ahead" of online review websites by integrating the reviews it collects from third-party patient surveys into doctor's profiles, said Craig Loundas, an associate vice president who leads patient experience.

A review committee weeds out comments deemed "slanderous," that call into question a doctor's clinical competency and could damage a reputation, so patients aren't getting a completely unfiltered view. This type of comment accounts for about 4 percent of  165,000 comments gathered. Any serious concerns are investigated internally, even if the comment isn't posted, Loundas said.

Online reviews often capture feedback that hospitals won't get from the scripted patient surveys they ask some people to complete, said Raina Merchant, director of Penn Medicine for Digital Health, who has studied how health systems can learn from Yelp.

For example, Yelp reviews are often written by the person who brought a friend or relative to the hospital. Caregivers are easy to overlook in a patient experience survey but could be crucial in deciding where a family goes for care, such as if an adult child is bringing a parent to the hospital.

"What these more crowdsourced sites add is additional, sometimes more nuanced, information that's also important," Merchant said. "We don't have a standardized way for asking about the views of the support people who come with you to the hospital."

Earlier this year, Temple Health launched a new "social listening" campaign that monitors 60 social media platforms and websites where patients may be talking about their experience at Temple. The exercise has produced some helpful lessons for staff, such as how much patients value even fleeting interactions with medical and administrative personnel.

"We've leaned into this new change and found it extremely beneficial," McBee said. "Those who don't choose to embrace this leave a lot of valuable feedback from their patients on the table."

• • •

The online tool  allows you to compare up to three hospitals. Charts compare local hospitals' ratings with state and national averages.

  • Hospital Compare evaluates hospitals on dozens of metrics and combines ratings in smaller categories into an overall five-star rating. So don't look at just the overall star rating — hospitals with an overall high star rating may not score as high in individual service areas.
  • Focus on metrics that are most relevant to your needs. For example, if you are pregnant, look specifically at a hospital's rating for early deliveries that weren't medically necessary. The tool also offers details on death rates, infection rates, and unplanned return visits for specific procedures, such as a knee replacement or heart attack.
  • Hospital Compare's "Timely and Effective Care" section includes metrics on emergency departments that could be useful if you live near multiple hospitals and want to establish a family emergency plan. Here you can see each hospitals' average ER wait times and patients who left without being seen, compared with state and national averages.