Program-analysis prose from counts you already have

What you’re making

A draft of the program-level story: what you offered, who came, what you measured, what changed, what you will change. Laid out to your own headings, with every figure traceable.

This is not the post-activity outcomes report. That recipe is one activity. This one is the year, or the accreditation term. The failure mode is the same: promotional language and numbers that cannot be found in the export.

What to have open

Activity counts, learner interactions, credit totals, commercial-support yes/no, outcome-measurement ticks, and whatever else you actually report. Last year’s analysis. Notes on what you already know went wrong (an RSS that drifted, a joint-providership that ate the calendar, a commendation criterion you are aiming at).

No learner-level rows. No faculty COI spreadsheets. Aggregates only.

The prompt

I am drafting a program analysis for an accredited CME/CE program. I am the author of record. Do not draft yet.

COUNTS I HOLD (do not add to these):
[paste: activities, formats, physician interactions, other learner interactions, credits, commercial support, whatever I have]

OUTCOMES WE ACTUALLY MEASURED:
[paste: % of activities at knowledge / competence / performance / patient health, and whether those were mostly subjective or objective]

WHAT I ALREADY KNOW WENT WELL OR BADLY:
[paste notes]

PRIOR ANALYSIS, FOR VOICE:
[paste, or "none"]

Tell me:
1. What story these counts actually support, in four sentences, including the unflattering parts.
2. Where I might be tempted to claim a trend the numbers do not show (including any year-over-year claim I cannot make because I only gave one year).
3. Which commendation or core-criteria stories these data could support, and which they could not. If you are not sure of current ACCME wording, say so and do not quote a criterion.

Use only my figures.

Draft the analysis using these headings:

[paste your headings, or use:]
1. Scope of the program this period
2. Learners and formats
3. Independence and commercial support
4. What we measured, and at which level
5. What the measurements actually showed
6. Problems we already know about
7. Changes for the next period

Requirements:
- Match the sample's sentence length and formality.
- Every number is one I gave you. Do not round. Do not invent a prior-year comparison.
- Do not write "the program improved patient care" unless I supplied patient-level results.
- Subjective competence stays labelled as self-report where that is what I measured.
- Mark gaps as [NEEDS:].

SAMPLE:
[paste]

COUNTS AND NOTES, UNCHANGED:
[paste again]

How to check it

Run the number audit. Then:

Trend pass. If you only supplied one year, delete every “increased,” “declined,” and “continued.”

Criterion pass. If the draft quotes an ACCME criterion, open the current requirements and compare the words. Models mix versions.

RSS pass. If regularly scheduled series dominate your counts, the narrative should say so. A draft that talks only like a conference program is describing some other office.

Where it goes wrong

It writes the brochure. “Robust,” “impactful,” “across the continuum.” Your self-study is not the annual report to the board.

It fills PARS for you. Learner counts, hours, and income figures must come from your export. One transposed denominator is a finding in a survey.

It promotes you to commendation. Forty percent of applicants in one published cohort reached it. The draft will assume you did. You say whether you did.

Program-analysis prose from counts you already have · bettermethod.ai/recipes/program-analysis/ · Last reviewed August 2026
These tools change often. Check the review date before you rely on anything product-specific in here, and check the guidance against your own accreditor and organization policies.