HEALTHCARE AUTOMATION — WORKFLOW ACCEPTANCE AND TIME-STUDY RECORD
Version: 2026-10-14
Companion guide: https://www.veroscribe.com/blog/healthcare-automation-guide

Use synthetic identifiers for rehearsal. Keep real identifiers and confidential
evidence in approved clinic storage, not in a public form or shared demonstration.
The completed example uses fictional records and assumed costs.

1. SCOPE AND CONFIGURATION
Clinic/workflow reference:
Administrator / clinician / privacy / technical owners:
Primary exception owner / absence backup:
Paid queue-review capacity and coverage hours:
Product, version, environment, interface, channel:
Template version, eligibility rule and preference rule:
Source-of-truth record and stable identifier:
Trigger and dispatch timing:
Allowed reads / writes / sends:
Explicitly prohibited actions:
Completion evidence (accepted / delivered / read / acted on are separate):
Deduplication key and operation-ID reconciliation route:
Booking-change/dispatch coordination and later-change correction route:
Fallback, pause mechanism and restart/backlog reconciliation:
Date / local time zone / configuration evidence references:

2. COHORT AND MEASUREMENT PLAN — COMPLETE BEFORE COLLECTING DATA
Baseline dates / pilot dates / comparable weekdays:
Proposed example: 10 clinic days per phase. Choose and justify your own duration.
Eligibility / exclusions and reasons:
Screened / excluded / eligible events in EACH phase:
Unique patients (report separately from appointment events):
Staffing, visit mix, channel, language/accessibility coverage:
Concurrent workflow changes / observer and measurement method:
Follow-up cutoff for outstanding work:
Missing measurements and how they will be reported:
Primary denominator: all eligible events, including failed/unknown attempts.
Secondary denominators: channel-specific attempts and successful completions.
Choose the study size for the outcome being measured.

3. PER-EVENT OFFLINE RECORD — COPY A BLANK ROW FOR EACH EVENT
Event reference / phase / eligibility decision:
Source record version / current preference checked at:
Trigger / queue / dispatch / accepted / delivered / acknowledgment timestamps:
Expected outcome / actual outcome:
Operation ID / retry count / deduplication result:
Reception active minutes:
Clinician active minutes:
Technical support active minutes:
Other role and active minutes:
Wait duration (separate; do not add to active effort):
After-hours active minutes (subset of role totals, not added twice):
Failed attempts / rework already included in role totals:
Defect or exception / owner / backup / next review deadline:
Patient-access consequence / complaint / manual contact work:
Unresolved work at cutoff:
Private evidence references / observer / date:
Remediation / recheck / approval:

Timing rules:
- Count each person's active effort, including simultaneous work by two people.
- Do not double-count overlapping timers for the same person.
- Keep elapsed waiting separate from active effort.
- Report setup/training separately from recurring review and monitoring.
- Unknown or missing time is not zero. Preserve failed and canceled attempts.

4. ACCEPTANCE CASES — RECORD ACTUAL FINDINGS, NOT ONLY PASS/FAIL
For EACH case complete:
Configuration/date / input fixture / expected result / actual finding /
Pass-Fail-Not run / evidence references / defect / remediation owner /
retest result / approver and scope / remaining conditions.

HA-01 Normal path
Input: one eligible appointment, verified destination, approved template.
Expected: one send; matching appointment, template, operation and state history.
Evidence: source snapshot plus queue, provider and clinic-state records.

HA-02 Stale booking and race boundary
Input: reschedule/cancel before dispatch; separately change during dispatch.
Expected: suppress obsolete pre-dispatch versions; show dispatch coordination.
Changes after dispatch use the separately approved correction process.
Evidence: versioned appointment timeline and message audit events.

HA-03 Identity/preference conflict
Input: ambiguous patient match; separately revoke preference after queuing.
Expected: no send to disputed destination; held state and staff task.
Evidence: source/preference snapshots, no-send evidence and assigned task.

HA-04 Uncertain completion
Input: lose response after provider acceptance.
Expected: reconcile the existing operation, no blind duplicate retry.
Evidence: operation ID, reconciliation response, total sends and clinic state.

HA-05 Delivery failure during owner absence
Input: failed delivery; primary owner unavailable.
Expected: backup task/deadline, booking preserved, contact not falsely confirmed.
Evidence: failed event, coverage assignment, booking and contact state.

HA-06 Outage and replay
Input: stop worker, change booking, replay queued events twice.
Expected: preserve backlog, reread current source, prevent duplicates,
and reconcile queued actions with work completed manually.
Evidence: before/after backlog, source versions and all send/task records.

5. COMPLETED FICTIONAL EXCEPTION ROW
Event: E-104 / appointment A-104 / pilot / eligible.
Configuration: synthetic reminder system, template R-1, local clinic time.
2026-10-14 09:00: version 7 queued; appointment Oct 15 at 14:00.
09:02: authorized staff move booking to 15:00; source version 8.
09:03: version 7 suppressed; version 8/preference verified.
09:04: operation M-208 accepted, then delivery failure recorded.
09:05: exception assigned to reception lead; covering receptionist backup;
review by 11:00. Appointment remains booked at 15:00.
10:10: backup attempts approved manual contact; no answer.
11:00: lead reviews; contact unconfirmed. Lead owns the next approved manual
contact attempt at 15:00. If unsuccessful, retain open for the covering
receptionist's 16:30 pre-closing review. No automatic booking cancellation.
Effort cutoff for this row: 11:00; later attempts remain outstanding.
Reception active effort: 3 minutes (includes manual call and lead review).
Clinician active effort: 0 minutes. Technical active effort: 0.5 minute.
Total active effort: 3.5 person-minutes. Wait is not added to this figure.
Expected: no stale/duplicate send; failed state accurate; owner and backup present.
Actual fictional finding: those states preserved; patient contact unresolved.
Evidence references: SYN-A104-v8, SYN-M208, SYN-E104 (invented teaching labels).
Acceptance: fictional administrative path passes; patient contact not complete.
Example approver: fictional practice-manager role.
Recheck triggers: template, interface, preference rules or coverage change.

6. FICTIONAL MATCHED-PERIOD CALCULATION
10 baseline days and 10 pilot days; 200 eligible events per phase.
Role                 Baseline min   Pilot recurring min   Assumed USD/hour
Reception            600            320                   30
Clinician             40             50                  120
Technical support      0             90                   60
TOTAL                 640            460

Recurring change: 640 - 460 = 180 minutes, or 0.9 minute per eligible event.
Baseline 3.2 versus pilot 2.3 minutes/event; difference 54 seconds/event.
One-time setup: 180 minutes. Initial pilot total: 460 + 180 = 640 minutes.
Reception capacity value: 280 / 60 * 30 = USD140.
Added clinician effort: 10 / 60 * 120 = USD20.
Added technical effort: 90 / 60 * 60 = USD90.
Net labor-capacity value: 140 - 20 - 90 = USD30.
Allocated software for this SAME ten-day period: USD40.
Recurring net: USD30 - USD40 = -USD10.
Setup value: 180 / 60 * 60 = USD180. Initial net: -USD190.
Replace the assumed rates with the clinic's costs and supplier quote.

7. FINAL BOUNDED DECISION
Observed cohort / missing data / limitations:
Whole-team minutes, per-event denominator and setup:
Wrong recipients / duplicates / stale sends / unresolved exceptions:
Patient-access results / complaints / overtime / callbacks:
Costs, currency, period, included volume, overage and cancellation:
Decision: hold / limited pilot / continue / stop.
Approved configuration, channel, template, visit type and action:
Unresolved conditions / owner / funded remediation / due date:
Clinical, administrative, privacy and technical approvers / dates / scope:
Stop triggers / next review / evidence retained / manual fallback: