Find the administrative cause of discharge delays

A discharge is clinically authorised at 10:00. Billing finishes at 11:10, but a payer query remains open. The dashboard reports a discharge delay without showing which team can act next.

This analysis separates administrative timestamps from the clinical discharge decision. The proposed queue helps explain unfinished billing, documents and payer coordination; it does not authorise clinical action or make its listed tasks conditions of care or departure. Local sequencing and exceptions remain inputs from the provider’s responsible owners.

Define the clocks before measuring them

In the illustrative case, the interval from clinical authorisation to billing readiness is 70 minutes. That is not automatically “billing delay”: some work may depend on legitimate late information, and the team needs the event history to understand it.

The example supports a 40-minute interval before final billing input and 30 minutes from that input to bill readiness. It cannot establish the total departure interval because that event is still missing. Do not substitute the dashboard refresh time as the departure time.

Decide which tasks can happen in parallel

Operations teams can often identify administrative preparation that does not require the final clinical decision, but the permitted tasks and sequencing need local approval. A preliminary bill review and a document check may have different dependencies; copying one generic sequence across all cases can create unnecessary waits.

Model each task with a case reference, owner, dependency, status, evidence and timestamp. Keep sensitive clinical content in its authorised system. The task queue should carry the minimum reference needed for the operational handover.

Bahmni's feature list identifies inpatient and billing functions. An integration still needs to establish which actual event in the deployed systems means clinically authorised, bill ready or physically departed; module names do not settle those definitions.

The diagram is a proposed coordination view. It does not make every listed task a legal or clinical condition of departure.

Handle revisions without resetting the history

A revised bill may reopen a review task. Preserve the earlier bill-ready event and the reason for reopening, then record when the revised task completes. Otherwise the system can make the elapsed time look shorter by erasing the first attempt.

Keep a payer acknowledgement distinct from a substantive decision in the task record. This guide supplies no insurer response-hour requirement or legal escalation rule. The responsible payer and legal owners must provide any such operating requirement from current applicable material.

Run a coordination pilot with limited authority

Use the first pilot to read agreed source events and organise administrative tasks for one ward. It should not write clinical approval, eligibility or release decisions. Ask system owners to demonstrate permissions and timestamp meanings before relying on interval reports.

Test missing final inputs, a revised bill, a repeated discharge event, a payer query and an unavailable interface. During an outage, use the approved coordination list and reconcile manually completed tasks before replaying messages. Stop the pilot if task status is being mistaken for clinical authority.

Report median and p90 intervals for a clearly defined completed cohort, stating the calculation method. List the number and age of still-open cases separately so long waits do not disappear from the report. Compare similar service groups and disclose exclusions; an average across unlike cases can hide the operational problem.

Keep the longest open case in view

Use an illustrative reporting snapshot with five completed administrative intervals of 30, 40, 50, 70 and 180 minutes. Their median is 50 minutes. Under the explicitly chosen nearest-rank method, p90 is the fifth value, 180 minutes: the rank is the ceiling of 0.9 × 5. A sixth case has been open for 300 minutes at the snapshot. Report that open case separately; it has no completed interval yet, but excluding it from view would make the queue look healthier than it is.

These numbers demonstrate calculation and cohort handling, not a performance target. With only five completions, p90 is unstable and should not be used to claim an improvement. Keep service scope, start/end definitions and sample size beside the result.

Have the discharge coordinator review the event history with billing and the clinical owner. The accepted report should explain both completed and open work without turning a dashboard status into permission for clinical action or a condition on departure.

Next actions

1. Agree clinical, administrative and departure event definitions with responsible owners.
2. Trace a small synthetic set containing revised bills and unresolved payer queries.
3. Review task authority, fallback and open-case reporting before a bounded ward pilot.

Topics: hospital discharge process turnaround time · hospital discharge turnaround · discharge billing workflow · cashless discharge administration