CCA (AHIMA): Two Discharge Counts and One Annual Report

Answer: Use 4,120 - the DTD count - as the inpatient discharge figure, after confirming the four December 31 cases are dated by date of occurrence.

Reconciliation of the 85:

• 52 observation encounters. Not inpatient discharges at all. Observation is an outpatient status; including it inflates the denominator and depresses length of stay. Exclude.
• 21 patients still inpatient at year-end. They have not been discharged, so they cannot be a discharge. They belong in the midnight census and, prospectively, in next year's discharges. Exclude.
• 8 duplicate same-day cases. A census-interface defect in the billing system created second records. These do not appear in a clean DTD feed; verify against the medical record numbers rather than accepting either report on faith. Exclude from anything drawn from patient accounting.
• 4 December 31 discharges dated January 1 by billing. A timing convention, not a real difference. They are discharges of this reporting year and stay in the 4,120.

Conclusion: 4,120 is the defensible inpatient discharge count. The billing figure is a revenue-cycle artifact and should never be used as a statistic.

Caution: do not simply adopt the DTD number without spot-checking. Verify a sample of the 8 duplicates and the 4 date-rolled cases against source documents, because the direction of the error matters if the DTD feed turns out to be the flawed one.

Formula: ALOS = total inpatient days (for the period) ÷ total inpatient discharges (including deaths) for the same period; both drawn from the same source system and the same patient population.

Applying it here: numerator = inpatient days only; denominator = 4,120. Observation days never enter the numerator, and observation encounters never enter the denominator.

Consequence of the error: observation stays are typically far shorter than inpatient stays. Adding 52 observation cases to the denominator while leaving their days out of the numerator produces a mathematically lower ALOS - a figure that looks like an efficiency improvement and is actually a definition error. If a board uses that number to justify bed capacity or a payer uses it in a contract, the hospital is making decisions on a number that is not measuring what it claims.

Direction of bias matters: mismatched numerator and denominator is the classic error. A denominator inflated with cases whose days are missing always biases ALOS downward; a numerator inflated with observation days biases it downward even more.

For the readmission measure: use the same 4,120-based population and the same inclusion/exclusion logic, or the readmission rate will not be comparable to the ALOS trend or to any benchmark.

1. Publish a data dictionary. Write down, for each statistic, the definition, the inclusion and exclusion rules, the source system, the date basis (discharge date vs. billing date), and the owner. "Inpatient discharge" must explicitly exclude observation, still-inpatients, and transfers within the facility that are not discharges.
2. Designate one system of record for inpatient discharges - the DTD module - and require all other reporting to derive from it rather than from patient accounting.
3. Fix the census interface defect that duplicated the 8 same-day cases; a monthly automated reconciliation between DTD and billing, with an itemized exception list, catches this while it is still small.
4. Reconcile monthly, not annually. An 85-case variance discovered in month 12 is an audit finding; discovered in month 1 it is a ticket.
5. Document the corrections and decide on restatement of prior periods before publishing the three-year trend, and footnote any definition change so the board does not read a definitional shift as a change in performance.
6. Validate against an external benchmark where one exists, since internal consistency does not prove the definition matches what other hospitals report.