The settlement advice arrives, the amount is lower than billed, and someone in accounts writes the difference off to "policy terms." That phrase is doing an enormous amount of work, and most of it is hiding money.
The three buckets
Predicted and accepted. Deductions you knew were coming and agree with — a room-rent proportionate deduction on a case where consent was recorded, a non-payable consumables list, a sub-limit on a named procedure. These are correct. They are not losses; they are the contract working.
Illustrative of the split, not of any hospital’s claims. The three blocks are drawn equal because the proportions vary by hospital, payer and case.
Copay. A share the patient owes by policy design. This is routinely counted as "lost," and it is not lost — it is collectable, from the patient, and if your discharge process does not collect it then that is a process problem, not a payer problem.
Unexplained. Everything else. A line reduced without a stated reason, a tariff applied that does not match the contract on file, a charge disallowed that was pre-authorised. This bucket is your query list, and it is usually much smaller than the total shortfall — which is exactly why it gets lost inside it.
Why the split has to be automatic
A settlement advice for a moderately complex admission runs to dozens of lines. Reconciling those by hand against the billed items, the pre-authorisation, and the payer's tariff is several hours of work per claim. At any volume, nobody does it. The write-off is not a decision — it is what happens when reconciliation is too expensive.
Parsed line by line and matched back against what was predicted, the same statement produces a short list of items that genuinely do not reconcile. That list is worth a phone call. The full shortfall never was.
What to do with the aggregate
One unexplained deduction is a query. The same unexplained deduction across forty claims is a contract conversation, and a much better one, because you are no longer arguing about a single patient. Aggregating the payer's own stated reasons tells you which conversation to have and how much it is worth.
This is also where recurring denial patterns become rules. If a payer consistently disallows a particular combination, that is knowledge worth encoding — proposed to a human with the evidence attached, never applied automatically, because a rule that fires wrongly is worse than no rule.
Drapto parses settlement advices line by line and separates predicted deductions from unexplained ones. See the reconciliation layer →
Written for Indian practices
Every figure here uses Indian policy mechanics, Indian payers and Indian clinical vocabulary. Nothing on this blog is translated from a US or UK playbook.