Northvale Health System: Resourcing a Clinical Platform Migration
Northvale Health System operates four hospitals and a network of outpatient clinics. Its electronic health record platform reaches vendor end-of-support in 20 months, and the replacement is a clinical transformation, not a software upgrade. IT has 210 staff, but 65% of their skills are in the legacy platform and its supporting interfaces. The CIO has identified 12 key-person dependencies, including the only two engineers who understand the radiology interface and a single analyst who maintains the medication formulary mapping. The board has approved $38 million, which funds the new platform but leaves little headroom for backfill or training. Clinical operations run 24/7, and radiology and laboratory uptime targets are 99.98%. A peer hospital's similar migration slipped 14 months after integration testing failed and clinicians lost confidence. The CEO wants a resource plan before the next board meeting.
1. Build the resource and skills case: what capacity gap will Northvale face, and which roles must be secured first? Justify using the constraints in the scenario.
Read the worked answer
Frame the gap in role-months, not headcount. With 20 months and 210 staff, roughly 137 people hold predominantly legacy skills. The migration needs roles the legacy estate did not require: integration and API architects, data migration and reconciliation leads, clinical informatics specialists, test and release managers, and change/release trainers for four hospitals.
Secure first - the single points of failure and the clinical bridge:
- The 12 key-person holders, starting with the two radiology interface engineers and the formulary analyst. Retention agreements, documented runbooks, and paired shadowing are cheaper than losing them mid-migration.
- Clinical informatics liaisons embedded in each hospital. Without clinician credibility, the peer hospital's failure mode repeats: testing passes technically, adoption collapses.
- A data migration lead with reconciliation experience, because medication and radiology data errors are patient-safety events, not defects.
Where the 137 legacy staff fit: many can be retrained into testing, service desk, and application support, but not all at once. A capacity model should show backfill cost per role and the point at which backfill exceeds the $38 million envelope. That is the conversation the CEO needs: either fund backfill or reduce scope and sequence go-live by hospital. Pretending the existing team can absorb migration on top of 24/7 support is the plan most likely to slip.
2. Evaluate whether Northvale should build the migration capability internally, use a systems integrator, or adopt a managed service for legacy support. Recommend a mix and state the conditions under which you would revisit it.
Read the worked answer
Use a mixed sourcing model, because the three needs have different economics.
- Systems integrator for the migration factory: integration build, data conversion tooling, and test automation. Northvale lacks these skills at scale, and hiring them permanently for a 20-month peak is poor value. The contract must include knowledge transfer, named Northvale counterparts, and exit/transition milestones, not just deliverables.
- Managed service for legacy run support: a 24-month bridge so internal staff can move to migration work. Keep clinical application ownership and change control inside Northvale; outsource commodity run and monitoring.
- Internal for clinical workflow design, change management, and safety cases. These require accountability that cannot be outsourced and are where the peer hospital failed.
Revisit conditions: if the integrator misses two consecutive gate reviews; if knowledge transfer metrics show Northvale staff are not being equipped to run the new platform; if the managed service's incident resolution falls below agreed clinical thresholds; or if the $38 million envelope forces scope reduction. Build an exit plan into each contract before signature, and have the CIO retain architectural decision rights. The test is not make-versus-buy in the abstract, but which capability Northvale must own on day one after go-live.
3. How should clinical uptime and patient safety be protected in the transition plan, and what governance body should own that risk?
Read the worked answer
Treat downtime as a clinical risk with a named clinical owner, not an IT service metric.
Protection should include:
- Staged go-live by hospital, never all four at once, with a pre-agreed rollback point and a decision authority who can halt.
- Parallel running for radiology and laboratory interfaces until reconciliation proves correct on live data for an agreed period. The 99.98% uptime target applies to the old and new estates simultaneously during cutover.
- Downtime procedures rehearsed with clinicians, including paper fallback and medication ordering, not just documented.
- A clinical safety case reviewed before each go-live, covering data integrity, alerting, and order sets.
- A freeze on non-essential changes during each cutover window.
Ownership: a clinical governance board chaired by the Chief Medical Information Officer or Chief Nursing Information Officer, with the CIO as a member, reporting to the board quality committee. IT can own technical readiness; only clinical leadership can accept patient-safety risk. Escalation should be explicit: the CMIO can stop a go-live, and that authority must be written down before pressure builds. The CIO's resource plan should show these activities as funded work, not as goodwill.