Print

Print


We also use Press Ganey, and our numbers are similar. Our ED sees about 6,000 per month, including 2,000 children. In a typical month we get 20-30 surveys back from the Peds ED and 30-50 from the general ED. Press Ganey insists that the data they generate from these surveys are valid. I think you would have a hard time finding an unbiased statistician who would agree with that position. 

In my opinion, there are two major flaws with the methodology. First, the number of responses is much too low to generate useful data. A survey that gets responses from only 1% of customers would have huge confidence intervals, rendering the data almost meaningless. Second, while the surveys are mailed in a random fashion, they are not returned in a random fashion. This all by itself undercuts any conclusions that might be drawn from the data. Each community has it's own patient biases. Non-random returns give us a glimpse into those biases, but I don't believe the responses give us an accurate picture of how the majority of our patients see us. BTW, our Peds ED scores consistently in the 70th percentile or above, so sour grapes has nothing to do with my opinion.

It seems curious to me that despite the points I just made, there is a very narrow spread in the Press Ganey scores. The difference between being in the first percentile and being in the 75th is only a few points. And yet, given the ridiculously small number of surveys that are returned, we should see large month-to-month variability in scores locally and a very wide range of scores nationally. Our scores locally rarely vary by more than 2 or 3 points. 

We did our own internal satisfaction survey based on the same questions asked in P-G. One of our secretaries made telephone calls to gather the data. We contacted 100 parents and had very few who refused to cooperate. Our satisfaction scores were much higher than those reported by P-G. Yes, I understand that gathering data by telephone introduces a whole new set of biases. However, we had a reasonable sample size and our information was gathered in a much more random fashion. I think it makes far more sense to use these data to make decisions about our operation than to use the random data generated by P-G. 

Jim Wilde MD




James A. Wilde MD, FAAP
Associate Professor of Pediatrics and Emergency Medicine
Director, Medical Student Research Elective, Dept of EM
Medical Director, Georgia United Against Antibiotic Resistant Disease (GUARD)
Department of Emergency Medicine
Medical College of Georgia
Augusta, GA 
[log in to unmask]
(706) 533-2925
Fax 706-721-7718


>>> Deborah Callanan <[log in to unmask]> 3/16/2009 6:53 AM >>>
Our hospital uses Press Ganey and is very concerned about scores.  We got 46 surveys in January when we had 7400 visits so the validity is certainly questionable.  It is very disheartening to get a raw score of 79-80 and have that translate to the first percentile.

--- On Sun, 3/15/09, Jay Fisher <[log in to unmask]> wrote:

> From: Jay Fisher <[log in to unmask]>
> Subject: Patient Satisfaction Scores
> To: [log in to unmask] 
> Date: Sunday, March 15, 2009, 7:50 PM
> Do you use patient satisfaction
> surveys in your PEM group?
> 
> Has your hospital administration suggested to your group
> that this is important to them?
> 
> 
> Thanks - Jay
> 
> 
> 
> Jay D. Fisher MD FAAP
> Director of Pediatric Emergency Services
> Emergency Physicians Medical Group
> University Medical Center, Las Vegas NV
> St. Rose Sienna Hospital, Henderson NV
> 
> For more information, send mail to [log in to unmask] 
> with the message: info PED-EM-L
> The URL for the PED-EM-L Web Page is:
>              
>    http://listserv.brown.edu/ped-em-l.html 
> 


      

For more information, send mail to [log in to unmask] with the message: info PED-EM-L
The URL for the PED-EM-L Web Page is:
                 http://listserv.brown.edu/ped-em-l.html

For more information, send mail to [log in to unmask] with the message: info PED-EM-L
The URL for the PED-EM-L Web Page is:
                 http://listserv.brown.edu/ped-em-l.html