Continuous improvement examples: healthcare, manufacturing and administration

Healthcare operations: reduce referral clarification loops

A fictional outpatient referral office reviews 100 non-urgent referrals received during a defined week. Thirty require a clarification request before scheduling. The team defines a clarification loop as a return to the sender for missing administrative information; clinical review is outside the scope. The starting proportion is 30/100, or 30%.

A SIPOC fixes the boundary from administrative receipt to scheduling readiness. A Flowchart makes the return loop visible. The team uses the CTQC guide to define what “complete information” means and a Data Collection Plan to distinguish missing information from clinical exceptions.

It tests a clearer administrative checklist with one referring office. In a second fictional set of 100 comparable referrals, 18 require clarification: 18%, a 12 percentage-point reduction in that observed sample. This comparison alone does not establish causation. The team also checks exclusions, demand and administrative handling time; urgent pathways and clinical decisions remain under their existing controls.

If follow-up supports the change, Standard Work names the sequence and exception route. A Monitoring Plan assigns the weekly review and a Reaction Plan describes what to do if missing information rises again. No patient identifiers belong in the practice worksheet.

Manufacturing: investigate changeover delays

A fictional packaging line records twelve changeovers for one product family using the same start and stop definitions. The total is 360 minutes, averaging 30 minutes per changeover. A Time Study separates preparation, active adjustment and waiting; a Flowchart shows when tools and materials become available.

A Pareto summary finds that waiting for prepared materials accounts for 90 of the 360 recorded minutes. That is 25% of this observed time, not proof that material preparation explains every slow changeover. The team investigates several affected cases and tests staging materials before shutdown on four suitable changeovers.

The four fictional test changeovers total 104 minutes, averaging 26 minutes. The small test and different case mix limit the comparison. The team checks quality release, safe working arrangements and material errors rather than reporting faster changeovers as sufficient evidence of success.

Standard Work can capture a reviewed preparation sequence; FMEA can structure discussion of possible failures introduced by staging. The Control Plan assigns timing checks and exception handling. These templates do not authorize changes to machine safety procedures.

Administration: improve invoice request completeness

A fictional accounts team reviews 80 invoice requests, of which 20 are returned for missing purchase references. Its initial return proportion is 25%. It defines an eligible request, keeps duplicate submissions separate and records exceptions so the numerator and denominator mean the same thing in later reviews.

The team uses Root Cause Analysis to examine why references are absent. Interviews suggest that the request form does not explain where to find them; this is a hypothesis to test, not a judgment about staff performance. A revised instruction is piloted in one department with a clear exception path.

In 40 fictional pilot requests, six are returned: 15%. The apparent reduction is ten percentage points, but the smaller group and short period mean that sustained performance remains uncertain. The team also checks completion time, unresolved exceptions and whether work has simply moved to another role.

A Training and Handoff Checklist distinguishes explaining the new instruction from verifying that the receiving team can use it. A Process Audit checks the actual method against the agreed standard. The Action Plan records unresolved work and a review date.

Apply the pattern to your own process

Define one problem, make the work visible, agree a counting rule, test a plausible change and review both intended and unintended effects. Record the limits of the comparison as carefully as the favorable result. Add a sustaining plan only after the method has been checked.

Each linked tool guide provides two more worked examples and practical steps in CI Tools. Open a separate synthetic practice worksheet to try the interface; it does not replace existing work. Keep an editable file using Save to Computer and use approved non-sensitive data during the temporary beta.

Further learning

• IHI: Quality Improvement Essentials Toolkit
• ASQ: Cause analysis tools
• Lean Enterprise Institute: Standardized Work

Independent educational resources, not endorsements of CI Tools.