The Merged Hospital: Emergency Changes and Login Delays
Two hospitals merged their IT departments 14 months ago. One side used a weekly change advisory board and a configuration management database; the other used autonomous DevOps teams and infrastructure as code. The merged department now runs a cloud-based electronic health record for 6,000 clinical staff across 11 sites. Symptoms: the configuration management database has 2,300 duplicate records; 19 emergency changes in the last quarter bypassed the change advisory board; nurses report 40-90 second login delays at shift change; and the EHR supplier changed an API without notice, breaking medication alerts for six hours. The CIO wants to 'standardise on one process model and buy a monitoring tool.' The nursing director says clinicians are losing trust. A supplier contract renewal is due in 90 days. The merger had no dedicated budget for process integration.
1. Which of the four dimensions is most implicated in the 19 emergency changes bypassing the change advisory board? Justify your answer with evidence from the setup.
Read the worked answer
Value streams and processes is most implicated, with organizations and people close behind. The merger joined two incompatible change models: a weekly CAB and autonomous DevOps. The resulting process has no agreed path for normal change, so teams route work around it as emergency change. Nineteen bypasses in a quarter is a systemic signal, not 19 individual failures. The evidence includes the 2,300 duplicate CMDB records, which mean the CAB cannot assess impact reliably, and the weekly cadence or six-week approval that cannot match a 30-day supplier release cycle. Organizations and people matter because the two cultures have not been integrated and no shared change authority has been agreed. Information and technology is also affected by poor configuration data, and partners and suppliers by the API change. But the bypass is primarily a process design problem: if the normal path is too slow or unclear, emergency change becomes the default.
2. Recommend actions across at least three of the four dimensions to reduce the 40-90 second login delays at shift change.
Read the worked answer
Organizations and people: create a joint clinical-IT group with named ownership for the login service, and agree who can authorise identity changes at shift change. Information and technology: measure the end-to-end login path, not just server uptime; fix the identity provider and directory sync; clean the duplicate CMDB records so dependencies are known. Value streams and processes: map the login service value stream from nurse request to resolution, remove unnecessary approval steps, and define a standard change for routine identity fixes. Partners and suppliers: add API change notice and performance requirements to the EHR supplier contract before renewal.
Start with the dimension that explains most of the delay - likely technology and process - but do not treat it as a monitoring problem. The CIO's proposed tool would only make the delay visible. A cross-dimensional plan is needed because clinical shift patterns, supplier behaviour and configuration data all shape the experience.
3. The CIO wants to 'standardise on one process model and buy a monitoring tool.' Why is this unlikely to restore clinician trust?
Read the worked answer
It addresses only part of the system. Standardising on one process model may reduce variation, but if it is the weekly CAB model it will not fit autonomous DevOps or a 30-day supplier cycle; if it is the DevOps model it may not satisfy the audit and control needs of a hospital. Either way, it ignores organizations and people: clinicians have experienced six hours of broken medication alerts and slow logins, and trust is rebuilt through visible reliability, not a process document. Buying monitoring improves detection but does not fix the identity provider, the duplicate CMDB or the supplier's API change behaviour. Think and work holistically says the four dimensions interact; a technology tool cannot compensate for process and relationship failures. A better approach is to co-design the change model with clinical and engineering representatives, define risk-based change authorities, and set a measurable reliability target for login and medication alerts. Trust follows evidence over time.