What a wellness program should report, and what it shouldn't claim
Session counts are easy to produce and nearly useless on their own. Here is the report I would want to read if I were approving the budget.
Every quarter I write a report for the company that hired us. It runs to about two pages. That isn't modesty, it's a rule: if the person who approved the budget can't read it in ten minutes, it won't be read, and a report that isn't read can't be argued with.
Here is what goes in it, and what I deliberately leave out.
Start with what the evidence actually says
Before I write a single number, I have to be honest about the standard. RAND's analysis of a large employer found that participation in lifestyle management programs was associated with lower health risks but not with lower healthcare cost or use. The eye-catching cost figures that circulated for years have not held up in later data.
That matters for how a report gets written. A program can say blood pressure came down in the group we tracked. It can't say it lowered your premiums. The second claim does not survive contact with an actuary.
If a vendor hands you a projected cost figure in quarter one, ask which assumption produced it and who has tested it. In my experience the answer is that it came from a PDF.
Three questions the report has to answer
What did we do? In enough detail to be checked. How many one-to-one sessions, on which days, with which teams. How many group sessions ran, and who came twice. This is the section that shows the program is real.
What changed? Using measures set before the quarter started, not chosen afterwards. If you decide in October that the thing you measured in June is the outcome, you will always find an outcome.
What are we changing next? One correction per quarter. If a report never contains one, it is marketing.
What I actually measure
Concrete things, in order of reliability.
Participation depth. Not sign-ups. How many people attended, how many of those attended four or more times, and how many new people started during the quarter. Forty people who came once is a different program from twenty-five who came weekly, even though both are "forty participants."
Adherence at 90 days. Of the plans written, how many are still being run. I ask. Plans decay quietly, and the honest number is usually humbling.
Strength and body composition. InBody scans where people opt in, plus a few movement benchmarks that need no equipment: a loaded carry, a squat, a hinge, a push-up standard. These move slowly, which is what makes them believable.
Self-reported behavior. Sleep, days trained, water, walks. Self-report is soft, but the same movement in the same direction across thirty people isn't noise.
Free text from surveys. The most useful line in any report I have written came from an employee who wrote "I stopped scheduling meetings at eleven on Tuesdays." That is what a program looks like when it is working.
What I leave out
Satisfaction scores on their own. "94 percent of attendees were satisfied" is a statement about a room, not about a person. I include satisfaction, but only next to behavior.
Aggregate percentages with no denominator. "Improved wellness by 12 percent" over an unnamed baseline isn't a finding.
Individual data. Body composition belongs to the employee. The employer sees trends and counts, never a name next to a number. I tell every participant that in their first session, because a scan that can reach their manager is a scan nobody fills in honestly.
How to read a report you didn't write
Four checks, ten minutes.
Is there a denominator? Every percentage should say out of how many people.
Is the time window stated? Participants over what period. A quarter and a year are not the same number, and vendors know it.
Are these the measures you agreed to? Compare against what was named before the quarter started, not what happens to look best now.
Is there a correction? If the report admits no problem, it was written to be forwarded, not read. The reports I trust most say something like "we lost the Thursday group session in July, here is what we did about it."
The format, if you want to copy it
Page one: the quarter in numbers, one table, and the correction we made. Page two: what the team said, verbatim, and what we are doing next quarter. Two pages, no stock photographs of people stretching.
I started writing reports this way because a client once told me his previous provider sent a forty-page deck he had never opened. He approved the renewal anyway, which told me the report didn't matter to him. It mattered to his boss, in the moment it was opened.
Write the report for that moment. Two pages, plain language, no unearned claims, and one honest number that moved.