Six Weeks to CPOE/eMAR Activation: Fix, Phase, or Delay
Riverside Community Hospital (168 beds, MEDITECH Expanse) is six weeks from CPOE/eMAR activation. The readiness assessment shows 78% of medical staff have completed the two-hour training (target 90%); pharmacy reports 12% of test orders lack a dose or route because orderable defaults were copied from the legacy build; nursing has created 14 unit-specific cheat sheets that duplicate the medication administration workflow on paper; and the downtime procedure has never been tested. The CNO wants a full big-bang activation on the original date because travel-nurse contracts end September 30. The CMIO proposes delaying 90 days. The CFO estimates a delay costs about $180,000 in bridge staffing and a lost vendor incentive. An ICU pilot ran 10 days with three medication-administration near misses, all attributed to scanning workflow confusion, none reaching patients.
1. Which three readiness gaps carry the greatest risk of patient harm rather than project delay, and why?
Read the worked answer
Rank by whether the gap can silently produce a wrong dose or a missed dose, not by how visible it is on a status report.
1. Missing dose or route on 12% of orderables. This is a build defect, not a training problem. If a prescriber can sign an order without a dose, the nurse must interpret intent at administration time. That converts a design error into a bedside judgment call under time pressure, which is exactly how wrong-dose events occur. It affects every unit and cannot be fixed by super-users.
2. The 14 unit cheat sheets. These are workarounds that make paper the real record. Once staff maintain parallel documentation, the eMAR stops being the source of truth: doses can be charted on paper and not in the system, or charted twice. The ICU near misses point the same direction, since scanning confusion is what prompts paper fallback.
3. The untested downtime procedure. A go-live concentrates failure modes. If the system drops and no one has practiced paper ordering and back-entering, orders are lost or duplicated on recovery. This is low-frequency but high-severity.
The 78% training completion matters, but it is recoverable in six weeks and is visible. The other three are latent.
2. Recommend an activation approach and state the gate criteria you would require before any live order is entered.
Read the worked answer
Neither big-bang nor a blanket 90-day delay is justified by the evidence. A delay of 90 days is longer than the build fix requires, and big-bang would push a known orderable defect into production across 168 beds. I would recommend a phased activation with two units first (ICU plus one med-surg floor), and only after the gates below are met; if the gates fail, the date moves, and the $180,000 is treated as the cost of not harming patients.
Gates:
- 100% of high-risk orderables (insulin, opioids, anticoagulants, chemotherapy, parenteral nutrition) reviewed by pharmacy with dose and route mandatory at order entry; documented sign-off.
- A second ICU pilot of at least 10 days with zero unmitigated scanning errors and barcode compliance above 90%.
- All 14 cheat sheets withdrawn and replaced by a single approved quick-reference; unit-level variance routed through change control.
- Downtime drill executed end-to-end, including paper order sets and back-entry reconciliation, with results signed off by nursing and pharmacy.
- 90% training completion by role, with a hard cutoff for prescribers who have not completed it.
- A daily go/no-go huddle for the first 14 days with authority to pause the next wave.
The phased plan preserves the vendor incentive if wave one starts on time and protects the remaining units if it does not.
3. What would you put on a one-page board dashboard for the go/no-go decision, and what would each measure tell you?
Read the worked answer
The dashboard should separate readiness (can we start) from safety (should we continue), because boards tend to over-weight schedule.
Readiness measures:
- High-risk orderable defect count, open vs closed. This is the single best predictor of wrong-dose risk at go-live.
- Training completion by role, with prescribers shown separately. Aggregate percentages hide the physician gap that drives order quality.
- Downtime drill pass/fail. Binary is appropriate; partial credit is meaningless here.
Safety measures during and after activation:
- Medication administration near misses and errors, per 1,000 doses, plotted against the pre-go-live baseline. A rise in near misses with no harm is an early warning, not a success.
- Barcode scanning compliance by unit. Low compliance means workarounds are re-emerging.
- Override rate on any new interruptive alert, with the top three override reasons. High overrides in week one usually indicate a build problem, not clinician resistance.
- Paper-order volume during downtime and back-entry reconciliation discrepancies.
I would also show the schedule line explicitly, so the board sees the tradeoff rather than discovering it later. A dashboard that only reports percent complete invites a decision to proceed on optimism.