The director of our group was called to the administrative offices to explain why our Press Ganey scores had dropped eight percentage points. A slightly larger than normal proportion of patients rated us as “good” rather than “excellent” for the past couple of months. Now the hospital wants answers.
It wouldn’t be so bad if the hoops we had to jump through were rationally related to the care that we are providing. They aren’t. The things that are useful measures such as “quality of care” and “medical decisionmaking” are intangibles that can’t be measured and plugged into a spreadsheet. Try it. Describe what “quality care” is and then figure a way to quantify it.
Is quality care adhering to published guidelines? What if there aren’t any guidelines for your patient’s situation?
Does quality care amount to less complications than the other practitioners in your specialty?If so, then a large percentage of physicians will cherry-pick healthy patients who are less likely to suffer complications. What happens to doctors who care for the severely ill patients?
Maybe quality of care is equivalent to low cost. If we use that definition, then we’re going to be creating an incentive for doctors not to order “unnecessary tests” and not to find diseases. The old saying goes “if you don’t go fishing, you won’t catch any fish.”
It is nearly impossible to come up with a quantifiable definition of “quality care.”
So what happens? In some specialties, we allow our worth as physicians to be measured based on data that can be quantified: Our ability to make patients happy. When speaking specifically about emergency medicine, the measurements don’t start there, though. First, the system throws patients into situations that tend to make people mad or frustrated — in need of medical care and forced to wait, sometimes for an excessively long time, with a bunch of other people who are also in need of medical care — THEN we start measuring physician worth.
Sometimes patient happiness isn’t related in any way to the physician’s care, but the staff gets blamed anyway.
There are the creature comfort complaints like “the room was too cold” or “the food was horrible.” Patients may get blankets, but sometimes decreased satisfaction scores still carry over to the provider side of the survey.
Then there’s the “I saw my doctor the next day and he said that you should have given me antibiotics for my cold.” Great. The follow up doc is both a backstabber and an idiot. Doesn’t matter that the patient would have gotten better even if the doctor prescribed soap suds enemas because nothing is going to make a viral infection go away except time. Nevertheless, the physician providing medically appropriate care gets lower marks because of another doctor’s inappropriate medical treatment.
There are other examples, but you get the picture. The best similarity I can come up with is using a ruler to measure how cold it is outside. The instrument you’re using has little bearing on what you’re trying to measure.
Then I did some studying and found out additional information about patient satisfaction surveys in general.
To get an adequate sample size, for 1000 patients, you need about 280 respondents to have a 5% margin of error and you need 400 respondents to have a 1% margin of error. That’s between a 28% response rate and a 40% response rate for statistically valid data. Larger sample sizes need less response rates, but these numbers are just to give a general idea. Know what the response rate for a well-known patient satisfaction survey company is? Between 8% and 10%.
Then there’s the statistical term called “standard deviation.” The bell curve for any data set can vary. If 10% of people taking a test each got grades of 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100, then the bell curve would be very flat and wide like a sprawling hill. If 10% got grades of 45, 80% got grades of 50 and 10% got grades of 55, then the bell curve would be very steep and narrow like the Washington Monument. The steeper that the bell curve, the less variation in the data. Often patient satisfaction data has a very steep and narrow bell curve. Therefore a small change in the data from one facility – such as a few more people than usual rating you as “good” rather than as “excellent” – can have a profound and potentially misleading effect on where your facility falls on the bell curve.
So I’ve decided to create a survey of my own … about the surveys.
Please pass along the link to your friends and colleagues. I’m looking for input from patients, administrators, and health care professionals. The more input, the better the results. There are at most about 20 questions, so it shouldn’t take more than 5 minutes to complete.
The survey is at this link on www.esurveyspro.com
I’ll publish the updated results on this site weekly for the next few weeks.
By the way, please make sure that your answers are accurate since you’ll be asked different questions based upon what answers you give. I want to try to make the results as reliable as possible.