A branch comparison is only meaningful once the branches are made comparable. Raw revenue per bed compares specialities, catchments and payer mixes, and attributes the result to management.
What has to be held constant
| Factor | Why it distorts | Adjustment |
|---|---|---|
| Case mix | A cardiac branch out-bills a paediatric one by construction | Compare within speciality, or index by case type |
| Payer mix | Scheme-heavy branches collect less per case, by contract | Report scheme and private separately |
| Catchment | Ability to pay differs by location | Compare trend against itself, not level against others |
| Capacity | A branch at ninety per cent occupancy cannot grow like one at sixty | Report per available bed-day, not per bed |
| Metric definition | “Revenue” means billed at one site and collected at another | One definition, computed centrally |
The last row is the common one
Groups usually discover, some way into a comparison exercise, that two branches were computing the headline number differently. Until the definition is central and identical, the comparison measures accounting convention.
Compare a branch to its own past
Trend against self removes case mix, catchment and payer mix at a stroke, because they are broadly constant within a branch over a quarter. It answers the question a group actually has — is this getting better — without pretending two different hospitals are the same hospital.
Questions we get asked
Is revenue per bed a fair way to compare hospital branches?
Rarely on its own. It compares speciality mix, payer mix and catchment as much as management, and a branch running near full occupancy cannot move the number the way an emptier one can. Revenue per available bed-day, reported separately for scheme and private volume, is closer to a like-for-like measure.
What is the most reliable cross-branch metric?
A branch’s own trend. Case mix, catchment and payer mix stay broadly constant within a branch across a quarter, so comparing a branch to its own previous period isolates what changed operationally, which is the question a group actually wants answered.
On the figures in this piece. Every rupee amount above is arithmetic on the assumptions stated beside it, not a market statistic. Where you need published Indian claim data with its source named, that sits on claim data. We publish no figure we cannot show the working for.