Skip to content

Patient Satisfaction Scores

January 12, 2011

After the hospital at which I used to work began using Press Ganey last year, I researched the issues about validity and statistical methods Press Ganey used that didn’t seem right to me.

After reviewing Press Ganey documents, things still didn’t seem right.

For example, one Press Ganey document (.pdf file) trying to explain away small sample sizes shows that with a sample size of 25, the standard error approximates 20. During an interview with a Press Ganey researcher, I was told that Press Ganey seeks to have a standard error of 2 or less. When the sample size increases to 50, the standard error approximates 14 – still seven times more than Press Ganey’s goal. A sample size of 100 only decreases standard error to roughly 10. Statistically significant?

Another Press Ganey document (.pdf file) provides scripts for hospital administrators to use when physicians object to Press Ganey’s statistical methods. Tell the doctors that you’ll put them in touch with the Ph.Ds at Press Ganey headquarters. “Rarely does a physician pursue the matter further when offered to speak to someone knowledgeable in statistics,” the documents says.

I did speak to Press Ganey statisticians. The answers they had to my questions really didn’t cut it.

When things still didn’t make sense, I interviewed some statisticians at St. Louis University and then co-published a couple of articles about patient satisfaction and Press Ganey in EP Monthly, the magazine where I have been a contributing author for more than 10 years. Links below.

Seven Things You May Not Know About Press Ganey Statistics

Are Press Ganey Statistics Reliable?

Didn’t seem to have much of an effect upon the misuse of the statistics.

I also engaged in a point-counterpoint with another physician about whether patient satisfaction scores are useful. A heavily edited version of that article appears on EP Monthly here. I submitted the original version of the article to Kevin, MD and he graciously agreed to post it on his blog.

I have a couple of other articles about Press Ganey to write in the next few months. I also received some raw data about a survey that asked doctors and patients about patient satisfaction surveys. I plan to publish that study later this year.

Will post all the articles here when I finish them.

6 Comments leave one →
  1. Steve Bauer permalink
    July 25, 2012 1:50 pm

    Dr Sullivan

    Thanks for the link. I think what is happening is that Press Ganey and the other survey companies are using an inappropriate definition of Population. They are using population as the number of surveys sent out and ‘n’ as the number of returns. Population should refer to the population as whole, meaning every patient that comes to the department. Example, we use NRC as our survey tool, and our n out of 7000 patient visits a month is around 30-40. 0.4% of patients. Ridiculous to nth degree.

  2. August 11, 2012 1:48 pm

    When I was at a hospital that used Press Ganey, I don’t recall them using “population” in the results, only the calculations from the responses returned. If they did include the population, it would highlight the fact that their data are invalid which would essentially invalidate a large portion of their business model.
    I’m not sure if NRC excludes certain patient populations, but Press Ganey has done so (including admitted patients, return patients, and patients who leave AMA) which further decreases the ability to make generalizations about care to the entire emergency department.

  3. Julieta Gabiola permalink
    October 14, 2012 10:10 pm

    I enjoyed reading your Pt satisfaction scores write up. Thank you.

    Julieta Gabiola, MD

  4. Karen Scott permalink
    September 13, 2013 10:23 am

    I appreciated your article about the statistics behind Press Ganey. Thanks for your research

  5. Dave Bryant permalink
    January 30, 2014 2:28 am

    PG only stands behind the results if a certain number of surveys are reached. Why then are they able to report scores when that number is not reached? Shouldn’t they just report “quantity not sufficient.” By reporting scores that are not valid, if it negatively affects me (which it has), would that be libel? Could we have a class action suit to stop PG from reporting scores when not enough surveys are returned?


  1. My Experience with a Healthcare Survey; or, From Nagging Doubts to Growing Skepticism

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: