Problem-Solving Framework: Define, Analyze, and Improve

Define the problem in observable terms

A problem-solving framework helps a team move from a concerning outcome to a testable explanation and an effective change. Start by describing what is happening, to whom, where and when. “Service quality is poor” is a judgment; “one in five appointments started more than 30 minutes late last month” is a measurable observation if the records are reliable. Verify the denominator and whether the pattern differs by site or time of day.

Clarify why the problem matters and who experiences the cost. A delay can affect patients, staff, transport schedules and other services differently. Consult people doing the work and those receiving it. Their accounts can expose a step missing from management reports, although a small number of complaints does not establish prevalence. Separate evidence from assumptions at the outset.

Set the boundary of the team’s authority. A local clinic may change its appointment reminders but not a national procurement rule. Identify constraints that are truly fixed and those that are habits open to review. A solution outside the team’s control might still require escalation, but the action plan should say who can decide it.

Diagnose a mechanism before choosing a fix

Map the current process from request to outcome. Look for handoffs, queues, information gaps, rework and variation. An average delay may conceal a few extreme cases, or an apparent staff shortage may actually be a scheduling mismatch. Use a simple timeline or process map with actual cases. Do not treat the first plausible cause as the root cause without checking it against the data.

Compare competing explanations. If appointments run late, possibilities include late arrivals, overbooking, unavailable rooms, long consultations or delayed test results. Each predicts a different pattern. Measure start times and bottlenecks for representative days, then ask which explanation fits. A cause-and-effect diagram can organize hypotheses, but drawing one does not validate its branches.

Watch for unintended incentives. A target to reduce average wait time could lead a team to exclude complex cases or redefine when the clock starts. Pair an improvement measure with a balancing measure, such as complaints, missed care or staff overtime. State the baseline and data quality before announcing a target.

Compare options and choose a test

Generate options that address the supported mechanism. If overbooking drives a predictable midday queue, a revised schedule may help; adding reminders will not fix it. If late arrivals are the major issue, consider whether transport, communication or booking rules are involved. Estimate effort, resources, equity and risks for each option. Avoid assuming the most visible or expensive intervention is the best.

Select a change small enough to test safely. Define its owner, location, start date and expected effect. A two-week pilot of a revised booking interval might measure median and long-tail waits, completion rates and staff overtime. Document what else changes during the pilot so the result can be interpreted. A single good day is not proof of a lasting improvement.

Plan for people who might be disadvantaged. A digital-only booking change could reduce administrative time while excluding users with limited connectivity. A faster queue may prioritize short cases and leave complex ones waiting. Ask affected groups about the proposal and monitor outcomes by relevant categories where privacy and sample size permit.

Learn from implementation

Collect data consistently before and after the change. A run chart can show whether a pattern shifted beyond normal variation, but interpretation depends on the time window and concurrent events. If a new staff member joined during the pilot, note that rather than crediting the schedule alone. Combine measured outcomes with observations of how people actually used the new process.

If the result is worse, investigate. The idea may be wrong, implementation may be incomplete or the measure may miss an important benefit. Do not expand a weak intervention simply because a budget has already been spent. Revise, retest or stop, with a documented decision. If it works locally, test whether the necessary resources and conditions exist elsewhere before scaling.

Assign responsibility for maintaining the improvement. A new checklist can decay when staff change unless it fits the workflow and someone reviews exceptions. Update training, instructions and system settings where needed. Measure whether gains persist after the initial attention fades. Sustainability is part of the solution, not an afterthought.

Communicate a bounded conclusion

Use an explicit decision rule at each stage. Before collecting more data, ask whether an uncertainty would change the choice; if it would not, further measurement may delay action without improving it. Before scaling a pilot, decide what improvement and what balancing result are acceptable. A threshold set after seeing favorable data is less convincing than one recorded in advance. Review the rule if assumptions change, but explain why it changed.

Report the original problem, evidence for the likely mechanism, options considered, intervention tested, result and remaining uncertainty. A team may reasonably conclude that a scheduling change reduced delays during the pilot while acknowledging that a longer period or another site is needed to judge durability. Avoid a success story that erases people who did not benefit.

Problem solving is iterative. A clear definition guides evidence collection; evidence guides a proportionate test; observed results guide revision. The framework is useful when it keeps the team honest about causes and makes it possible to distinguish real improvement from a plausible-sounding plan.

For an apparently complex problem, separate what can be improved locally from what needs a wider system decision. Staff can document a recurring supply gap and reduce waste, but they may not have authority to change a purchasing contract. Escalation with specific evidence and a proposed decision is stronger than telling the local team to “work harder.” The distinction also prevents a local success metric from masking an upstream problem.