Every year, clinicians around the world collect continuing education credits the way frequent flyers collect miles. A lecture here, an online module there, a weekend course in a nice hotel. The credits are logged, the licence is renewed, and the system records that learning has taken place.
What the system does not record is whether anyone’s practice changed on Monday morning.
That is not a small gap. It is the whole point. Continuing education exists to improve care. Yet almost every accreditation system measures inputs: hours, credits, attendance. The one thing that matters, a change in what clinicians actually do, is almost never measured at all.
What the evidence actually says
The good news is that continuing education does work, a little. The best-known summary is a Cochrane review of educational meetings led by Louise Forsetlund, published in 2009. It pooled 81 randomised trials involving more than 11,000 health professionals and found a median absolute improvement in care of 6%. The effect was larger when attendance was higher and when meetings combined interactive and didactic formats. Purely didactic lectures were less likely to change practice.
A separate Cochrane review of audit and feedback, led by Noah Ivers in 2012, examined 140 studies. Showing clinicians data on their own performance compared with a standard produced a median 4.3% absolute improvement in compliance with recommended practice. Again, modest. Again, highly variable, from almost nothing to 16% at the upper quartile.
Two things stand out from these findings. First, the average effect of education is real but small, which should temper the enthusiasm of anyone selling a course. Second, the variation is enormous. Some programmes change practice substantially; many do almost nothing. The difference lies in design, not in the number of credits awarded.
What a 6% effect means at scale
A median 6% absolute improvement can sound trivial. It is not, if it applies across a whole workforce. Imagine a region with a thousand general practitioners, each seeing a few hundred patients a year with a condition where recommended care is delivered 60% of the time. Lifting that to 66% means tens of thousands of additional patients a year receiving the recommended treatment, from a single well-designed programme. Very few drugs deliver an absolute change of that size across a population so cheaply.
The problem is not that the effect is small. It is that we pay the same for programmes that produce 16% and programmes that produce nothing, and have no way of telling them apart, because nobody measures the outcome. A market in which quality is invisible to the buyer drifts toward whatever is cheapest and most convenient. That is precisely what has happened. The clinician choosing between a rigorous twelve-week audit cycle and a two-hour webinar faces the same credit either way, so the webinar wins, and the provider who built the audit cycle eventually stops offering it. Good programmes are not driven out because they fail. They are driven out because nobody can see that they work.
The famous 17 years is a warning, not a fact
Discussions of practice change often cite the claim that it takes 17 years for research to reach routine care. It is worth knowing where that number comes from. In 2011, Zoë Slote Morris, Steven Wooding and Jonathan Grant reviewed the literature on time lags in translational research in the Journal of the Royal Society of Medicine, under the pointed title “The answer is 17 years, what is the question.” They found only 23 relevant papers, few of them comparable, using different measures and different starting points.
In other words, we do not really know how long it takes for evidence to change practice, because we rarely measure practice. The famous figure is less a finding than a symptom of the same blind spot that affects continuing education.
Why the system rewards the wrong thing
The incentives behind continuing education are almost perfectly misaligned with its purpose.
Clinicians need credits to keep their licence, and they are busy. The rational choice is the cheapest, quickest source of credits, not the most useful one. Providers compete on convenience, price and venue, because that is what buyers reward. Commercial sponsorship funds part of the market in many countries, and sponsors have their own interests in what gets taught. Regulators count credits because counting is easy and auditing practice change is hard.
Nobody in this chain is acting badly. They are all responding sensibly to what is measured. The problem is that what is measured is attendance.
What education that changes practice looks like
The programmes that produce the larger effects in the research tend to share a few features:
- They start from a gap in the learner’s own practice. Audit data showing that a clinician’s prescribing, referral or screening rates differ from peers creates a reason to learn that no brochure can.
- They are interactive. Case discussion, practice with feedback and small groups outperform lectures.
- They are repeated. Single sessions fade. Follow-up, reminders and a second round of data sustain change.
- They address the system, not just the individual. A clinician who learns a new protocol but works in a clinic whose order sets, templates and staffing do not support it will slip back.
The last point is the most neglected. A great deal of what gets labelled “education failure” is a clinician returning from a course to an environment designed around the old way of working.
A worked example: antibiotics in primary care
Consider a common target: antibiotic prescribing for respiratory infections that are usually viral. A traditional approach would offer a one-hour lecture on antimicrobial resistance, with credits. Most attendees already know antibiotics do not treat colds. Their prescribing is driven by time pressure, patient expectation and uncertainty, none of which a lecture changes.
A practice-change approach starts differently. Each clinician receives their own prescribing rate for these conditions, alongside the rates of anonymised peers. The session is spent on the hard cases: the patient who insists, the child with a fever on a Friday afternoon, the older adult with a cough that might be something more. The clinic agrees a small system change, such as a delayed-prescription option or a leaflet for patients. Three months later the data comes back. Those who changed get to see it; those who did not see that too.
The second approach is more work for everyone. It needs prescribing data, a facilitator and a follow-up cycle. It also has a far better chance of moving the number that matters. In most accreditation systems, it would earn roughly the same credit as the lecture.
Credits for change
The fix is not to abolish credits. It is to attach more of them to evidence of change.
Some accreditation bodies already offer credit for performance improvement activities, in which clinicians review their own data, make a change and re-measure. Expanding that model, and weighting it more heavily than passive attendance, would shift the market. Providers would compete on helping clinicians improve measurable outcomes, not on hotel quality.
There is a cost. Measuring practice requires data, and many clinicians, especially outside large systems, do not have easy access to their own performance data. Building that access is the real infrastructure project of continuing education, and it has barely begun in most countries.
The pieces exist. Electronic health records, prescribing databases and registries already capture much of what clinicians do. What is usually missing is the step that turns those records into a clear, fair report a clinician can act on: adjusted for case mix, compared with meaningful peers, delivered regularly and privately rather than as a league table. Clinicians rightly distrust crude comparisons that ignore how sick their patients are. A feedback report that seems unfair is dismissed, and then it teaches nothing.
This is where the next few years could make a real difference. Better data infrastructure, and analytic tools that make case-mix adjustment routine, could make personalised practice feedback far cheaper to produce than it is today. If that happens, the argument that measuring practice change is too hard will weaken considerably. The question then will be whether accreditation systems move with it, or keep counting hours because counting hours is what they have always done.
The case for the old model
It would be unfair to pretend that time in a lecture hall has no value. Continuing education also keeps clinicians exposed to new developments, connects them with peers outside their own institution and, for many, is one of the few protected spaces for reflection in a crowded working life. Not everything worth learning shows up in an audit within three months. Understanding a new class of drug, or a shift in the evidence on screening, may pay off years later in a decision no audit will ever track.
There is also a real risk in tying credit tightly to measurable change. What gets measured is what is easily measured: prescribing rates, screening uptake, documentation. Harder things, such as diagnostic reasoning, communication and judgement in uncertain cases, could be squeezed out simply because they resist metrics. A system that only credits measurable change could narrow learning to the countable.
That argument partly holds, and it points to balance rather than replacement. Some credit should remain for exposure and learning that cannot be tied to a number. But the balance has swung so far toward attendance that almost all credit currently comes that way. Shifting a meaningful share toward demonstrated change would not destroy the value of the lecture. It would stop the lecture being the only thing the system pays for.
Who could change the market
Regulators and licensing bodies set the rules, and the most powerful thing they can do is weight credit toward improvement activity and require some of it in every renewal cycle. Several systems already point this way. Revalidation of doctors in the UK, for instance, asks for evidence of quality improvement activity alongside continuing learning. Making such elements weightier and more common is an incremental change, not a revolution.
Employers hold the data. Hospitals and large practice groups can give clinicians routine access to their own performance compared with peers, which is the raw material of education that works. Many already collect it for management purposes and never share it with the people it describes.
Education providers will follow the money. If credit rewards change, the better providers will build programmes around audit, feedback and follow-up, and some will be good at it. Commercial sponsors will find less value in funding the passive lectures that currently carry their messages. And clinicians themselves can choose, even within the current system, the learning that starts from their own data. It is less convenient. It is also the only kind that reliably shows up on Monday morning.
The takeaway
Continuing education is one of the largest, least scrutinised investments in the health workforce. It works modestly on average and very well when designed well. The problem is that the system pays for sitting in the room, not for what happens afterwards. Until credits follow changes in practice, clinicians will keep collecting miles to nowhere.
What is the one course or learning experience that genuinely changed how you practise, and why did it stick? Share it with us in the comments.
