Be careful how you ask, ‘Were you satisfied?’

Wording, rating scales influence survey results

At its most basic, gauging patient satisfaction is easy. You ask patients how satisfied they were with their care, and you calculate ratings based on percentages (80% very satisfied, etc.).

But if you want to use patient surveys to understand your flaws and make improvements, you must be sure that both your wording and your rating system provide you with meaningful answers.

"In my judgment, the reason to do surveys is to retain business and build business," says Jerry Brown, senior vice president of operations for Women’s Health Partners, a Nashville, TN-based physician practice management company. "It needs to be a survey that results in capturing actionable information.

"You need to determine, ‘What are the areas you need to be improving upon? What are the areas of service you’re not providing that your patients/customers are seeking?’" he says.

Brown is currently working with market research consultant James R. Toth to combine several existing surveys and to tailor them toward the the surveys used by the health plans and the National Committee for Quality Assurance (NCQA) in Washington, DC.

While Brown intends to tout his success in achieving high patient satisfaction, he also wants to show that he truly listens to patients’ concerns.

When fashioning a new survey his bottom line is, "Don’t just tell me I’m OK. Tell me how I can become better. In the survey we’re working with, we’re trying to find out not just that we rank on a certain point on the graph, but what can we do to move up the graph."

Here are some of the areas that Toth and Brown suggest you consider when drafting a patient survey:

Rating scale.

A four-point scale (excellent, good, fair, poor) forces the patient to rate the practice either positively or negatively, notes Toth, who is president of Edge Healthcare Research in Nashville. The practice, therefore, gets a clear answer to questions. "[The scale] doesn’t offer any middle ground," he says.

A five-point scale (excellent, very good, good, fair, poor) offers more information for those practices who expect positive ratings. "You use that when you expect to be highly rated and you want some additional delineation or specificity of the positive answers," he says.

Since the typical patient satisfaction score is 80% to 90% positive, many practices use the five-point scale, he says.

Both the NCQA patient survey and the new Consumer Assessment of Health Plans (CAHPS) survey makes occasional use of the 10-point scale in which patients choose a number from 0 (worst) to 10 (best). Toth advises against such wide scales.

"What’s the difference between a 2 and a 3? Nobody knows," remarks Toth. "When these 10-point scales are used, most responses are in the 7, 8, 9 range anyway."

Conversely, a yes/no question often provides too little information, he says. Consider the question, "When you called the practice, did you find our new automated answering machine easy to use?"

"What if I thought it was really bad? You don’t give me an option [to say that]. What if I thought it was OK, but I had a little bit of a problem at the end? Many times you want shades of the yes or no."


Make sure your questions are clear and connect directly with the answer choices, Toth advises. That means each question should include just one concept, he says.

For example, a survey may ask, "How would you rate the efficiency and friendliness of the receptionist?" Suppose the receptionist was efficient but not friendly? "How do you rate it? In the middle?" Toth asks.

Another common problem is the use of vague or overly general words. For example, Toth recalls one questionnaire that gave patients the chance to strongly agree, agree, etc., with statements such as "Everyone was professional."

"Who is everyone? That includes doctors, nurses, business staff," he says. "Now you’ve lumped them together, and you don’t allow me to specify."

Questions on the survey should have consistent wording and identical rating scales, says Toth. In other words, if one asks to rate "responsiveness of nursing staff," the next question also should begin with a noun, such as "waiting time in reception area."

Avoid biased wording

How you word questions also can inadvertently encourage positive or negative responses, he cautions. For example, one survey asked patients, "Did the physician spend as much time with you as you wished?"

"You’re going to get a lot of no’s. ‘No, I want more time.’" says Toth. "Maybe a better question would be reasonableness of the time the physician spent with you, with a rating of excellent, very good, good, fair, or poor."

Open-ended questions.

No matter how carefully you word the questions, the answers are always limited. So Brown likes to include some open-ended questions that give patients a chance to tell him how to improve the practice.

For example, the surveys have asked about services the patient received elsewhere that could have been provided by the practice. Prompted by responses, Women’s Health Partners has added new classes for expectant mothers and wellness programs.

One Women’s Health Partners practice also maintains a patient management committee of physicians, nurses, and other staff, which meets bimonthly. The committee responds to concerns or suggestions raised on surveys, as well as other issues that arise related to patient satisfaction and customer service, Brown says.

Brown considers the open-ended questions to be a core part of the patient survey. "We’re trying to structure our surveys to identify how we can always exceed [patient] expectations," he says.

[Editor’s note: For more information about wording and rating systems on surveys, contact James Toth, President, Edge Healthcare Research, 2301 Hillsboro Road, Suite 300, Nashville, TN 37212. Telephone: (615) 385-9315. Fax: (615) 385-5999.]