When Customization Costs Lives: The EMR That Kept Growing
A real-world example of Sunk cost fallacy in action
Context
A regional hospital system embarked on a major electronic medical record (EMR) implementation to replace a patchwork of legacy systems. Leadership authorized extensive customizations to match long-standing local workflows rather than adopting the vendor's standard configuration.
Situation
After 18 months and $3.0M spent, the project was behind schedule and clinical staff reported rising usability problems. The vendor recommended reverting to their standard workflow templates (estimated one-time cost $300k and three months to stabilize), but the project steering committee instead approved further custom work to "finish what was started."
The bias in action
Decision-makers framed choices around how much had already been invested instead of comparing future costs and benefits, repeatedly approving change requests to justify prior spending. Project champions argued that abandoning custom features would waste two years of work, even though the customizations were causing training delays and higher error rates. The team ignored objective stop/go criteria (schedule slips, user-acceptance scores, and cost-per-month-to-complete) and treated sunk costs as a reason to continue. As a result, incremental investments were approved without a fresh ROI analysis.
Outcome
The implementation timeline slipped another 18 months while an additional $1.2M was spent on bespoke features. Clinicians faced prolonged training and workflow disruptions; inpatient scheduling errors and order-entry delays rose. The organization ultimately accepted the vendor's standard configuration after cumulative spending far exceeded initial projections, and morale among clinicians and IT staff eroded.