Operated function · credentialing
Nobody is at fault. Eleven verifications went out to eleven organisations, each of which answers on its own schedule, and none of them is being chased individually because chasing eleven things across forty candidates is a full-time job nobody has.
A single file needs verification from many separate places: licensing boards, training programmes, previous employers, certification bodies, insurers, references. Each is a different organisation with a different process, a different channel, and no obligation to be quick.
The requests go out. After that, each one is a thread nobody owns. There is no single view showing that seven of eleven have come back, which two have been outstanding longest, and which one was never actually sent because the address bounced.
Chasing is where the time is recovered and it is the work that does not happen, because chasing eleven items across forty candidates is hundreds of small acts a week and it competes with processing new files, which feels more productive.
The candidate is in the dark. They were hired, they are being paid or waiting to be, and they cannot tell whether their file is nearly done or has not moved in a month. The good ones take other offers during this window, which is the most expensive way to lose a hire.
Then expiry arrives. Licences and certifications lapse on a schedule that is entirely knowable in advance, and a lapse discovered after the fact means a professional who was working without a current credential — which is a compliance event rather than an administrative one.
A file is complete only when every third-party verification returns, and each request is an independent thread with no owner, no age and no chase schedule — so file duration is set by the slowest unchased item.
That is why credentialing duration is so insensitive to staffing. Adding processors speeds up the parts you control, which are a minority of the elapsed time, and changes nothing about eleven letters sitting in eleven organisations.
Treating each verification as a tracked item with an age and a scheduled chase is the whole intervention. It is unglamorous, it is high volume, and it is exactly the kind of bounded repetitive work that is worth delegating — hundreds of small acts, each individually trivial, collectively the difference between six weeks and three.
Delivery confirmation is the cheapest win inside it. A request that never arrived is indistinguishable from a slow verifier until somebody checks, and it is frequently a meaningful share of the outstanding items.
Expiry is the same mechanism run forward rather than backward. Every credential has a known end date, and re-verification started early enough is routine rather than urgent — a lapse is only ever a failure to look ahead.
And the verifier-level pattern is the structural fix: when one board or one programme delays most files, that is a relationship or a channel to change once, rather than a delay to absorb per candidate.
Elapsed time from file open to complete — measured by days against your own baseline, split into our processing and waiting on each verifier.
Verifications with an age and a chase history — measured by items individually tracked, against a baseline where the file had a status and the items did not.
Requests that were never delivered — measured by the count identified by delivery confirmation, against a baseline where non-delivery looked like slowness.
Days a professional is employed and unable to work — measured by elapsed days between start date and credentialed status, which is the direct economic measure.
Credentials that lapse before renewal — measured by expiries reached without a current renewal, against your own prior period.
Which verifiers actually delay files — measured by median response time by verifier, which points at a channel or relationship fix rather than a per-file chase.
any credentialing, privileging or appointment decision, and no assessment of whether a professional is competent or suitable. Nothing here evaluates a candidate. It obtains, tracks and records verifications — the decision belongs to your credentialing committee or authority and it is not a decision a vendor should be near.
File and status live in your credentialing system of record. Nothing migrates, and no second record of whether somebody is credentialed exists — two answers to that question is a patient-safety and liability problem at once.
Verifiers are reached through whatever channel each one requires, which in this domain still frequently means post and fax. Where a verifier offers no electronic route, that is stated as a constraint on duration rather than engineered around.
The verification requirement set and the verifier directory are yours: written, versioned, owned. The directory in particular becomes valuable, because it records who actually responds and how quickly.
A credentialing file contains a person’s complete professional history and sometimes adverse findings. Access is scoped to the operation and does not extend to material the verification set does not require.
Verification is obtained and recorded; it is never characterised. Whether a returned verification is satisfactory, whether a gap in history is concerning, and whether an adverse finding is disqualifying are judgements for your committee, and anything of that kind routes rather than being summarised.
Every request and receipt carries a record: what was requested, from whom, when, how, what came back, and when. In this domain that record is the evidence of primary-source verification itself, which is what an accreditation review examines.
Operational access is not permission to train. Professional history data does not become material improving anything serving another organisation.
The credentialing committee owns the requirement set and the decision, and both need stating in writing. The distinction that makes delegation appropriate is that obtaining a verification and deciding what it means are different acts.
Whoever answers to your accreditor should confirm that the record produced satisfies primary-source verification evidence requirements in your jurisdiction and under your accreditation standard. That is a compliance question with a specific answer and it belongs at scoping.
Where an obligation attaches through an accreditor, a licensing body or a data class, it is marked applicability-gated rather than presented as standing.
Your current open files — read-only, nothing requested and nothing chased — with every outstanding verification aged individually and grouped by verifier.
The observation phase produces a view most credentialing operations have never had: not how long files have been open, but how long each individual verification has been outstanding, and which verifiers are responsible for most of the delay.
It usually finds two things immediately. A number of requests were never delivered and are waiting on nothing, and a small number of verifiers account for most of the elapsed time — which is a channel or relationship fix rather than a staffing one.
If you continue, the first delegation is delivery confirmation and scheduled chasing on the existing open files. That is the highest-return step, it makes no decision of any kind, and it is measurable within a month.
The decision is not outsourced and should not be — it stays with your committee or credentialing authority, and nothing here evaluates a candidate, characterises a verification or judges an adverse finding. What is delegated is obtaining and tracking: issuing requests, confirming delivery, ageing each item and chasing on schedule. Those are administrative acts, and separating them from the decision is what makes the delegation appropriate at all.
Per request, largely true. Across a book, less so: some of what looks like slow verifiers is requests that never arrived, and unchased items sit longer than chased ones at almost every organisation. The verifier-level breakdown also tells you which relationships are worth a direct conversation — a single slow verifier accounting for most of your delay is a specific, addressable problem rather than an ambient one.
Most hold the file and its status well. The question that separates them is whether they age each individual outstanding verification, confirm delivery, and chase on a schedule without somebody remembering. If yours does all three, this is a duplicate. The observation phase settles it by ageing your current outstanding items, which takes read access and no change to anything.
A report is a list somebody has to look at, and lapses are almost always found afterwards rather than prevented — which means the report existed and the looking did not. The distinction here is that renewal starts early enough to be routine, by rule rather than by review. If your lapse count over the last two years is zero, the report is being read and this part is not your problem.