Which services should a practice add capacity for?

The waiting list is the most available number in the building and the least suitable one for this decision. It counts patients who have already been routed to the practice — by a referring clinician, by an existing panel, by a payer network that lists you. A specialty with no referral relationship does not appear on it as unmet demand; it appears as nothing at all. Growing against the list therefore does something quite specific: it relieves whichever part of the existing intake is most congested, which is not the same as serving whichever part of the local population is least served.

The number that would settle it is a denominator, and it exists. Two federal series publish employment and establishment counts by county and by industry code, and either of them can put a practice's own volume over a local population of employers, workers or competing establishments. Neither is a demand estimate. Both are structural, both are free, and both count something narrower than their names suggest — which is why using them well means knowing what they exclude before quoting them.

County Business Patterns counts paid employees during the pay period that includes the twelfth of March, at whatever the establishment's primary industry is. That means it is a single-week snapshot rather than an annual average, that a scheduler working inside a hospital-owned practice is counted under the parent establishment's code, and that since 2018 the published cell carries deliberately infused noise, flagged per cell in the release. It is a good denominator and a poor headline, and the difference matters when a hire is being justified with it.

The Quarterly Census of Employment and Wages counts something else again: jobs covered by unemployment insurance, not people. A clinician holding two covered posts is counted twice. A proprietor drawing no covered wage is not counted at all — which in a small independent practice can mean the owner-clinician, the single most expensive unit of capacity in the building, is structurally invisible in the series being used to price adding another one.

The commitment side is what makes this decision unusually unforgiving. Recruitment, credentialing and payer enrolment for a new clinician run months ahead of the first billable visit, so the estimate that justified the hire cannot be corrected inside the period it was made for. By the time the demand assumption is testable, the salary has been running for two or three quarters. That asymmetry is the argument for spending an afternoon on a denominator rather than a week on a waiting-list report.

There is also a quieter failure worth naming. New capacity is filled through the same intake that is already handling enquiries, and if that intake drops a meaningful share of them, the added capacity does not fill at the rate the business case assumed — it fills at the rate the phones and forms permit. Practices routinely commission a clinical hire and an intake problem in the same quarter without connecting them.