Electronic Health Record Go-Live Readiness Under Clinical Pressure

I recommend a conditional no-go or delay, not proceed. Readiness indicators are below any safe threshold: training completion is 62% against a clearly incomplete workforce; the medication-alert override rate is 41% versus a 15% target, which is a patient-safety signal; 12 configuration defects include two high-risk items without owners; and pilot support tickets are triple baseline. The communications pattern-monthly CIO email and one-time champion invitation-has not built the two-way engagement needed for a clinical rollout. The board announcement is a schedule constraint, not evidence of readiness. A three-week go-live would transfer known quality and safety risk to nurses and patients. I would recommend a four-week delay with explicit exit criteria: 90% role-based training completion, all high-risk defects closed and retested, override rate below 20% in a monitored pilot, named defect owners, daily floor support, and active nurse champions. If the CMO insists, require written risk acceptance from the CMO and CNO, but still escalate the patient-safety issues.

Root causes are not one training failure. Quality: defects were logged but not assigned, so the defect management process lacks ownership and severity-based exit criteria. Resource: training was likely generic, not role-based, and super-users were not resourced; the vendor’s position that training is the hospital’s responsibility leaves an unmanaged dependency. Communications: one-way monthly email does not support clinical change; champions were invited but not engaged, so peer influence is missing. Corrective actions: assign owners and due dates to all 12 defects, fix and retest high-risk alerts first; build role-based training for medication administration and document workarounds; recruit and pay/backfill nurse champions; hold daily 15-minute floor huddles; create a two-way escalation channel and a weekly readiness dashboard; and clarify in writing the vendor’s training and configuration responsibilities. The reasoning is to fix the system, the people, and the feedback loop together-training alone will not reduce overrides if alerts are poorly configured.

To the CMO and board, I would lead with patient safety and factual readiness, not blame. Message: “The project is not ready for the announced date. Current data show 62% training completion, a 41% alert override rate, and two unowned high-risk defects. Proceeding would ask nurses to use a system with known safety gaps. We recommend a four-week delay with a controlled recovery plan and measurable exit criteria. We need your decision on the revised date and on written risk acceptance if you choose to proceed.” I would avoid saying the board was wrong; instead, show the new information and the decision required. Provide a one-page dashboard and a clear ask: approve delay, approve backfill for training, and sponsor a message to clinical leaders. If the CMO still wants to proceed, document the advice and escalation to the CNO/quality committee. This preserves confidence because it demonstrates control, not concealment.