Let me preface this information with two important facts:
- What I'm about to show you has a serious sample size issue.
- What I'm about to show you has a serious data structure issue
But it's still interesting. A friend of PCC (you know who you are) asked me whether or not we can judge the efficacy of the flu shot (and the flu shot vs. flumist) using our client data. That's an intriguing question and, ultimately, the answer is no. Why not? Because we have no understanding how much of the flu work our clients do represents all the flu shots given to their patients. And we have no idea how accurately our clients diagnose and then record the flu among their patients.
Still, I never fear inaccuracy in either my numerator or denominator. Igor called into four fairly large practices and we can up with this data:
Flu Rates Among Immunized PCC Clients
|
% of Flu Diagnoses for Kids with Flu Shot |
% of Flu Diagnoses for Kids with FluMist |
Client 1
|
.2% |
0% |
Client 2
|
2.9% |
3.2% |
Client 3
|
.2% |
N/A |
Client 4
|
.1% |
N/A |
I'm sure you can see my concerns with the data - this raises more questions than it answers. So, I turn it to you, fair readers. What percentage of your patients had the flu last year? Is there a difference among those who used FluMist and those who didn't? What do your rates look like in general - are they 2/10s of a percent like some clients, or 10x that amount like some others? Obviously, you can share your data with me anonymously. :-)
Comments
0 Comments