CME Accreditation Software Problems: The 6 That Surface at Reaccreditation, and How to Avoid Them

Why does disclosure keep getting scoped as a checkbox?

Because it looks like one. Somebody discloses or they do not, so a form field seems adequate, and every packaged platform obliges. What the ACCME Standards for Integrity and Independence actually require is four distinct states with four distinct owners: collect financial relationships from everyone in a position to control content, decide which of those relationships are relevant to the content, mitigate the relevant ones, and disclose to learners before the activity.

Real cases show why one flag cannot carry that. A planner discloses a consulting relationship with a company whose product is out of scope, so it is collected but not relevant, and someone made that judgement on a date. A speaker discloses a relationship that is relevant, so a named reviewer performs a peer content review with a recorded outcome before the slides are used. Store one boolean and you have thrown away the entire audit trail while appearing compliant.

The fix is to make disclosure an object rather than an attribute. It carries a date, a company, the person's role, the scope decision, the mitigation action, the reviewer and the outcome, and it attaches to a person and an activity rather than to a person alone. Then the pre activity gate becomes automatic: an activity cannot open for registration while any content controlling participant has an unresolved relevant relationship. Coordinators stop chasing forms because the system chases them, and the reaccreditation sample assembles itself because the evidence was captured as the work happened.

What goes wrong when transcripts and attendance history migrate?

Learner transcripts are the least glamorous and highest risk part of any CME migration, because a clinician renewing a licence is relying on that history and will not accept an explanation. The failure is rarely a lost record. It is a silently changed one.

Three patterns recur. Credit values that were expressed differently in the outgoing system, for example half credits recorded as decimals in one place and as separate quarter records in another, which sum correctly in aggregate and wrongly per activity. Duplicate learner identities, because the same physician exists under a personal email from 2018, an institutional account, and a name change, and merging them badly either doubles a transcript or deletes half of it. And activities whose credit type was never recorded because the old system only issued one kind, so the migration has to infer it, which is a guess dressed as data.

The only safe approach is to reconcile before you cut over, not after. Recompute totals per learner in both systems and compare, hold the differences as a working queue, and resolve identity merges with a human review rather than a matching rule. Publish a learner facing transcript in parallel for several weeks so clinicians find the errors while the old system is still available. Expect the exercise to surface historic problems that predate the project, and agree in advance who decides what the correct answer is.

Why do CPE Monitor, MOC and single sign on integrations break after launch?

Because each one has a clock and an identifier that the education workflow does not naturally hold. Pharmacy credit reported through CPE Monitor requires the learner's NABP e-Profile identifier and date of birth, and must be submitted inside a defined window that most providers treat as sixty days. Maintenance of certification registration requires activity level attributes plus learner identifiers that the physician supplies once and should never be asked for again. Neither of those is a field on an activity. They are eligibility preconditions that must be satisfied before a learner is allowed to claim.

Systems that treat credit types as extra numbers on an activity therefore break in a specific and predictable way. A learner claims, the certificate issues, the downstream submission fails silently because an identifier was missing, and nobody discovers it until a pharmacist cannot find the credit in their own record months later. By then the window has closed.

Build the submissions as a visible queue with retries and an owner, not as a fire and forget call. Prompt for each profession's identifier before the claim is permitted rather than after, and show the learner the state of their own submission. Single sign on against a health system identity provider is the other common surprise: it rarely costs much engineering, and it routinely costs four to eight weeks of institutional review, so start it in week one and treat it as a dependency rather than a task.

What happens when joint providership and commercial support are not covered?

When you are the accredited provider for a specialty society's annual meeting, you carry the accreditation responsibility for content you did not create, with a partner who has their own registration system and their own sponsors. You need their planning documents, their disclosures, their attendance data, and a clean line between educational grants and exhibit or advertising revenue, because those are treated and reported differently.

No packaged CME platform does this well, because it is half contract management and half fund accounting, so providers handle it in a shared drive plus email. The failure surfaces at year end when finance has booked a device manufacturer's payment identically to an exhibit fee and nobody can now say which it was. Reconstructing intent after the fact is not a reporting problem, it is an integrity problem, and it is the sort of thing that turns a routine review into a longer conversation.

Give the arrangement its own record: the agreement, the partner contacts, the responsibility split, a portal where the partner uploads planning documents and disclosures, an attendance import in whatever shape they can produce, and a funds ledger that tags each receipt as commercial support, exhibit, advertising or registration income at the moment it arrives. The letter of agreement, the acknowledgement to learners and the reporting then line up because they read from the same record instead of three people's memories.

Should you build custom or configure what you already own?

If you are a single specialty society or a small provider running enduring materials and a handful of live activities with AMA PRA Category 1 Credit only, buy. EthosCE and CloudCME are built for exactly that, they track requirement changes so you do not have to, and CloudCME in particular is genuinely strong at live activity check in, which is why so many academic centres use it. A custom build would be an expensive route to something you can licence this month.

HealthStream is worth naming separately because it is frequently misused here. It is excellent at hospital workforce compliance and competency training at scale, and it should stay in your stack next to a CME system rather than be asked to become one. The record an accredited provider needs is the disclosure and mitigation trail, the multi credit awarding rules and the provider reporting obligations, none of which is what workforce compliance software is shaped around.

Build when two or more of the following hold. You award three or more credit types and failed claims are a weekly support event. You run regularly scheduled series across more than two hospitals and attendance arrives in three formats. You act as accredited provider for outside partners more than a couple of times a year. Your mitigation workflow cannot be expressed in the configuration and therefore lives in a shared drive. The threshold is not activity count, it is the number of distinct rule sets your staff are holding together by hand.

How do hidden costs get into the quote?

In our delivery experience a first release covering activity planning with disclosure and mitigation, regularly scheduled series with offline tolerant attendance, and single credit claiming with certificates runs $60,000 to $130,000 in 12 to 18 weeks. A full platform adding the multi credit eligibility engine with CPE Monitor and maintenance of certification submission, joint providership with a partner portal, commercial support ledger and continuous PARS validation runs $150,000 to $350,000 phased over 6 to 12 months.

The overruns come from four places, and all four are scopeable in advance. Credit types, because each downstream registry is its own integration with its own identifiers, windows and failure modes, so a quote written for two and delivered for four will move. Sites, because each hospital running a series brings its own rooms, badge infrastructure and local habits, and the fifth site is not a copy of the first.

Then transcript migration, which is quoted as a data task and delivered as a reconciliation project. And institutional review time for single sign on and information security assessment, which is calendar rather than engineering but sits on the critical path all the same. Ask for these as four named line items. A developer who folds them into a general contingency has not delivered in this category.

What separates a build that works from one that fails here?

Ask them to model disclosure on the first call, before anything is signed. If they attach a boolean to a user, the conversation is over. The right answer is a dated, company scoped, role scoped relationship with a relevance decision and a mitigation record attached to an activity, and a developer who has done this will describe it without prompting.

Ask how they would capture attendance in a basement auditorium at seven in the morning, with no signal, for a clinician who will not install anything. Offline first capture, badge readers where the buildings already have them, and photographed sign in sheets extracted with a human confirming ambiguous names should all come up. If the answer is a QR code, they have solved the easy case and left you the year end reconciliation.

Ask what happens when a submission to a downstream registry fails. A visible queue with retries and an owner is the correct answer. Silence is the wrong one, and silence is the default in most implementations.

Then ask about series templates, because a coordinator configuring forty six weekly sessions once rather than forty six times is where the operational saving actually lands. And settle ownership before kickoff: repository, cloud accounts and the right to bring in another firm. At Digital Heroes the client owns the code from the first commit. The system holds credit history that clinicians renew licences against, so access to it must never depend on a vendor relationship.