From needs assessment to learning objectives

What you’re making

Five draft objectives. Each one tied to a specific line of your needs assessment, each written with a measurable verb at the right level, ready for you to cut down to three and edit into your house wording.

Before you start, get clear on who is doing what. You supply the gap and the evidence. The tool supplies the wording. That is the only split that works.

A language model has never met your learners, read your chart review, or worked in your specialty. What it is genuinely good at is taking a gap you have already established and writing well-formed draft objectives faster than you can type them. Swap those roles around and you get fiction that reads beautifully.

What to have open

Your needs assessment material, in whatever state it’s in. Survey tabulations, the QI or chart review finding, the guideline change, last cycle’s evaluation comments, the specialty society statement. It doesn’t need to be tidy. It does need to be yours, with the numbers exactly as you have them.

Write your practice gap out as a sentence before you start, even a clumsy one. Current state, then desired state. If you can’t write it, the tool can’t rescue you.

The prompt

You are helping me draft learning objectives for an accredited continuing education activity. I am the author of record, not you. Draft candidates for me to edit.

PRACTICE GAP (current state vs. desired state):
[paste your gap statement]

EVIDENCE THAT THE GAP IS REAL:
[paste your needs assessment sources: survey results, chart review or QI findings, guideline changes, prior evaluation comments]

Use only the figures I have given you. Do not add, estimate, round, or supplement any statistic, prevalence figure, or citation. If you think something is missing, say so instead of filling it in.

TYPE OF NEED: [knowledge, competence, or performance. Pick one, or write "not sure"]
LEARNERS: [specialties, roles, practice setting, roughly how many]
FORMAT AND LENGTH: [for example, 60-minute live webinar, half-day workshop, enduring material]

Draft 5 candidate objectives. For each one:
- Write it as one sentence beginning "After this activity, participants will be able to..."
- Use a measurable verb that matches the type of need I named. Do not use "understand", "appreciate", "be aware of", "learn about", "gain knowledge of", or "be familiar with".
- Quote the specific part of my gap statement or evidence that it addresses.
- Name the highest outcome level it could plausibly be measured at, using Moore's outcomes levels, and say what the measurement would actually be.
- Flag it if it is not realistically achievable in the format and length I gave you.

Then answer two questions separately:
1. Which parts of my gap statement are not addressed by any of these objectives?
2. Which of these objectives are really the same objective written twice?

Those last two questions are where most of the value sits. Generating objectives is the easy part. Noticing that you’ve written the same objective three ways and left half the gap uncovered is the part that normally needs a second reviewer.

What good output looks like

Say your gap is this: internal medicine clinicians at our three community sites are documenting a heart failure medication review at 41% of post-discharge visits against a target of 90%, and they cite uncertainty about titration thresholds.

A usable draft objective looks like this:

After this activity, participants will be able to select an appropriate next-step titration for a patient with HFrEF who remains symptomatic on a starting dose of a guideline-directed agent.

Addresses: “cite uncertainty about titration thresholds.” Highest plausible level: Moore level 4, competence. Measurable with case vignette questions straight after the activity. Level 5, performance, would mean re-auditing documentation at 90 days. Feasible in 60 minutes: yes, if you limit it to two drug classes.

Three things make that usable. The verb describes something you could watch a person do. The link to your evidence is spelled out instead of assumed. And it tells you the honest ceiling on what you can measure rather than overselling it.

Output that just hands you five bulleted sentences has done about a fifth of the job.

How to check it

Where it goes wrong

It drifts down to knowledge. Models default to the level they can write most fluently, which is recall. You asked for competence and you will often get knowledge with a competence-sounding verb bolted on. Naming the type of need in the prompt helps a lot but doesn’t fix it completely, so check every time.

Too many objectives. It will happily hand you five strong-sounding objectives for a 45-minute webinar. The flag instruction catches some of that. You catch the rest.

Invented evidence. This is the one that can actually hurt you. Left to itself, a model will supply prevalence figures, guideline version numbers, and citations that are plausible, specific, nicely formatted, and wrong. The “use only the figures I have given you” line is not filler. It is doing real work, and you should still run the number pass.

It doesn’t know your house style. Whatever conventions your accreditation team applies to objective wording, the tool has no idea. Paste three of your previously approved objectives into the prompt as examples and the match improves straight away. Better still, keep them somewhere you can reuse. That’s covered in working from your own documents.