PDSA (PlanâDoâStudyâAct) is often cited but rarely executed with enough rigor to survive real-world community services delivery. The common failure is running a âpilotâ that is too big, too vague, and impossible to verify. This article shows how to run small, safe tests with fast feedback loops, then scale changes with evidence. It ties practical PDSA to Continuous Improvement Cycles and to capability expectations in Competency Frameworks. The goal is an improvement method that front-line teams can use weekly, and that funders and governance bodies can audit without guesswork.
What makes PDSA ârealâ versus performative
A real PDSA cycle has a narrow test, a defined prediction, a short time window, and an explicit decision rule about what happens next. A performative cycle is âwe tried a new formâ with no baseline, no prediction, and no verification. In community services, where staff rotate and settings vary, rigor matters: if you canât prove a change is being used correctly, you canât claim it reduced risk.
Two oversight expectations you should design for
Expectation 1: Changes must be safe-by-design and rights-aware. Many community services changes affect safety and rights (supervision levels, restrictive practices, medication workflows, escalation thresholds). Oversight bodies expect providers to manage change riskâtesting in a controlled way, documenting mitigations, and ensuring changes donât create unintended harm or inequity for people served.
Expectation 2: Evidence of scaling discipline. Funders and system partners often see pilots that never translate into standard practice. A credible provider can show a scaling pathway: what was tested, what evidence supported scaling, what training/briefing occurred, and what audits confirm sustained adherence after rollout.
Plan: define the change so it can be tested in days, not months
Start with a problem statement linked to a failure mode (e.g., âmedication discrepancies are discovered late during evening shiftsâ). Then define a single change that is observable (e.g., âshift-start MAR reconciliation with a sign-off and an escalation ruleâ). Write a prediction: âIf we add reconciliation and escalation, discrepancies will be detected earlier and medication near misses will reduce within 30 days.â
Pick 2â3 measures: one outcome (near misses), one process (reconciliation completion), and one balancing measure (shift overtime minutes, staff satisfaction, or on-call call volume). Set a short test window: 7â14 days is often enough for a first cycle.
Do: run a safe pilot with clear boundaries
Limit the pilot to a single site, a single shift band, or a defined cohort (e.g., high-risk medication clients). Identify who is involved and how they are briefed. Provide the minimum tools needed: a checklist, a template, or a dashboard view. Make it easy for staff to comply under pressure.
Most importantly, define pilot safety rules. For example: if reconciliation identifies a discrepancy, the escalation pathway must be used immediately; if staffing drops below a threshold, the pilot pauses; if a rights impact is identified, leadership reviews before continuing. These boundaries protect people served and protect the credibility of the improvement method.
Study: collect feedback that reflects reality, not optimism
Study is not a meeting where people say it âfelt better.â It is a structured review of your measures plus direct operational feedback across roles. Gather information from at least: frontline staff who used the change, a supervisor who verified it, and (where relevant) a clinical lead or quality lead who can judge safety implications.
Use short questions: What slowed you down? What workarounds appeared? What information was missing? Did the change create new failure modes? In community services, the most valuable learning is often about information flow and handoffsânot the control itself.
Act: decide to scale, modify, or stopâbased on evidence
Act requires an explicit decision rule. Examples: scale if process compliance is above 85% and outcome trend improves; modify if compliance is below 85% due to tool problems; stop if balancing measures worsen materially (e.g., overtime spikes) or if safety concerns emerge. Document the decision and the reason so governance can follow the logic.
Operational examples that meet the 4-part development gate
Operational example 1: Rapid PDSA to improve escalation for deterioration
What happens in day-to-day delivery. The pilot introduces a simple âdeterioration trigger cardâ used at the end of each visit for a defined high-risk cohort (recent discharge, complex comorbidity, frequent ED use). Staff record one of three triggers (new confusion, breathing change, medication side-effect concern) and follow a scripted escalation pathway to an on-call clinician or supervisor. Supervisors review triggers daily and confirm that escalation documentation is complete and timely.
Why the practice exists (failure mode it addresses). Deterioration is often missed because early signs are documented inconsistently and escalations depend on individual judgment under time pressure. The trigger card exists to standardize recognition and ensure information moves to decision-makers quickly.
What goes wrong if it is absent. Staff document subtle changes but donât escalate; supervisors discover issues late; the person deteriorates overnight or over a weekend; avoidable ED use increases. The failure presents as âsuddenâ crisis, even though weak signals were present in routine notes.
What observable outcome it produces. Evidence includes increased timely escalations (a positive signal), clearer escalation decision trails, reduced late-night crisis calls, and a reduction in avoidable ED transfers for the cohort over 30â60 days.
Operational example 2: PDSA for supervision consistency in a multi-site program
What happens in day-to-day delivery. The pilot tests a standardized supervision template and cadence for one site: every direct supervisor must complete two structured supervision sessions per staff member per month, with one including an observed practice check. The template prompts review of incidents, near misses, documentation quality, and a single competency focus. The program manager audits five supervision records weekly and provides immediate feedback.
Why the practice exists (failure mode it addresses). Supervision drifts when workload is high, becoming reactive and inconsistent across managers. The template exists to standardize what âgood supervisionâ contains and to ensure it drives practice quality, not only wellbeing check-ins.
What goes wrong if it is absent. Staff development becomes uneven; weak practice persists; managers rely on informal coaching that is not documented; turnover and burnout rise; and quality issues surface late because no consistent forum exists to detect drift.
What observable outcome it produces. Improved supervision completion rates, better documentation quality, fewer repeat incidents linked to the same practice gaps, and stronger audit trails for workforce assurance.
Operational example 3: PDSA to reduce medication omissions during peak workload
What happens in day-to-day delivery. The pilot shifts one medication administration window by 30â60 minutes to reduce clustering and adds a âcritical medsâ double-check for a small set of medications. Staff use a brief sign-off at the time of administration. The supervisor reviews omissions daily and checks whether omissions correlate with staffing, scheduling, or tool failures.
Why the practice exists (failure mode it addresses). Omissions often occur when too many tasks collide in the same time window and staff improvise. The change exists to redesign workload and add a targeted control where harm potential is highest.
What goes wrong if it is absent. Staff skip steps under pressure, document later, and near misses go unreported. Over time, the service sees repeat omissions, clinical destabilization for some individuals, and increased scrutiny from funders and families.
What observable outcome it produces. Reduced omission rate for the pilot cohort, stable or reduced overtime, improved double-check compliance, and fewer after-hours clinical escalations related to missed doses.
Scaling without drift: what to standardize
When you scale, standardize the minimum necessary: the definition of the change, the tool/checklist/template, the training or briefing method, and the verification approach (audit/observation frequency). Keep local adaptation limited to workflow details that donât change the controlâs intent. Then embed it into routine governance: monthly review of compliance and outcomes, and quarterly board visibility for high-risk controls.