18.3 Quality Improvement, Clinical Audit & Patient Safety

18.3 Quality Improvement, Clinical Audit & Patient Safety

Quick Summary: Quality is not an opinion about a facility; it is measured on defined dimensions, along the structure-process-outcome axis, using named tools. The examinable core is the difference between quality assurance and quality improvement, the Plan-Do-Study-Act cycle, the clinical audit loop (which must close with a re-audit), non-punitive incident reporting, and Maternal and Perinatal Death Surveillance and Response.

---

The Six Dimensions of Quality

A service can fail on any one of these while succeeding on the rest, and questions usually name a scenario and ask which dimension failed.

---

Donabedian: Structure, Process, Outcome

   STRUCTURE  ───────────►  PROCESS  ───────────►  OUTCOME
   What you HAVE            What you DO            What RESULTS

   • Staff numbers          • Adherence to         • Maternal mortality
     and qualifications       guidelines           • Case fatality rate
   • Buildings, water,      • Partograph use       • Complication rate
     electricity            • Hand hygiene         • Patient satisfaction
   • Equipment, drugs       • Correct triage       • Coverage achieved
   • Protocols, records     • Timely referral

Two rules for using it:

1. Good structure enables but does not guarantee good process, and good process makes good outcome more likely but never certain. A facility with every drug and machine can still deliver poor care if protocols are ignored.
2. When an outcome is poor, trace backwards. Was the process wrong, or was the structure incapable of supporting the process? A rise in postpartum haemorrhage deaths might be a process failure (active management of the third stage not practised) or a structure failure (no oxytocin in the cold chain). The corrective action is completely different, which is why the distinction is examined.

Process indicators are usually the most actionable at facility level, because outcomes like maternal mortality are rare events that move too slowly to guide monthly improvement, while "proportion of deliveries with a correctly completed partograph" moves immediately.

---

Quality Assurance Versus Quality Improvement

They are complements, not rivals. The Kenya Quality Model for Health (KQMH) is the national framework that brings both together, integrating standards, self-assessment, and continuous improvement, and it is supported by national quality standards and a facility assessment approach.

---

The Plan-Do-Study-Act Cycle

              ┌──────────────────┐
              │       PLAN       │  State the aim, the measure, and the
              │                  │  change to be tested
              └────────┬─────────┘
                       ▼
              ┌──────────────────┐
              │       DO         │  Test the change on a SMALL scale —
              │                  │  one clinician, one clinic day
              └────────┬─────────┘
                       ▼
              ┌──────────────────┐
              │      STUDY       │  Compare results against the prediction;
              │                  │  plot data over time
              └────────┬─────────┘
                       ▼
              ┌──────────────────┐
              │       ACT        │  Adopt, adapt, or abandon —
              │                  │  then run the next cycle
              └────────┬─────────┘
                       │
                       └────────► next PDSA cycle

The discipline that makes this work and that examiners test:

• Small and fast. Test on one clinician for one day, not across the whole facility for a quarter. A failed small test costs nothing; a failed facility-wide rollout costs credibility.
• Measure over time, not before-and-after. Plot the indicator on a run chart monthly so that a genuine shift can be distinguished from ordinary variation. A single before-and-after comparison cannot tell improvement from noise.
• Predict before you test. Writing down what you expect is what turns a change into a test.
• A quality improvement team should be multidisciplinary — clinician, nurse, records officer, pharmacy, and support staff — because the process being improved crosses all of them.

---

The Clinical Audit Cycle

Clinical audit measures actual practice against an explicit, agreed standard and then changes practice to close the gap.

1. Select a topic — high volume, high risk, or known problem area.
2. Set the standard from a national guideline, stated as a measurable criterion with a target (for example: "100% of women in labour have a partograph commenced at 5 cm dilatation").
3. Collect data on current practice against that criterion.
4. Compare and analyse — where is the gap and why?
5. Implement change — protocol, training, job aid, process redesign.
6. RE-AUDIT — measure again against the same criterion.

The loop must close. An audit that stops after implementing change has produced an opinion that things are better, not evidence. Re-audit is what distinguishes clinical audit from a survey, and a vignette describing a team that audited, retrained, and moved on to the next topic is describing an incomplete audit cycle.

Note also what audit is not: it is not research. Research asks what the right thing to do is and requires ethical approval; audit asks whether we are doing the thing already known to be right, and measures against an existing standard.

---

Incident Reporting and the Just Culture

Patient safety depends on knowing what went wrong, which depends on people telling you.

• An adverse event is unintended harm caused by health care rather than by the underlying disease.
• A near miss is an event that could have caused harm but did not, whether by chance or by interception. Near misses are more valuable than adverse events for learning, because they are far more common and carry no injury.
• A never event is a serious, largely preventable incident that should not occur where available preventive measures are in place — wrong-site surgery, a retained instrument, or a transfusion of incompatible blood.

Reporting must be non-punitive and systems-focused. A blame culture produces two effects, both lethal: staff stop reporting, and the organisation loses the information it needs to prevent recurrence. A "just culture" still distinguishes honest error and system failure, which are learned from, from recklessness and wilful violation, which remain accountable — but the default response to an error report is inquiry, not discipline.

Root Cause Analysis

A structured retrospective inquiry into a serious incident. It asks why repeatedly, driving past the immediate individual act to the system conditions that permitted it.

A nurse administers the wrong drug. Why? Two look-alike ampoules were stored side by side. Why? Shelving is arranged alphabetically with no separation of sound-alike names. Why? No storage protocol exists for look-alike, sound-alike medicines. System fix: physical separation and tall-man lettering — not "counsel the nurse."

The examinable point: a root cause analysis whose conclusion is that an individual should be more careful has not found a root cause.

---

Maternal and Perinatal Death Surveillance and Response

MPDSR is a national system and a statutory expectation, not an optional review.

• Every maternal death must be notified, and maternal deaths are treated as notifiable events requiring reporting within a defined short window.
• Every notified death is reviewed by a facility committee, which establishes the sequence of events, identifies avoidable factors across the three delays, and agrees actions.
• The three delays framework structures the review: delay in deciding to seek care, delay in reaching care, and delay in receiving adequate care once at the facility. Only the third is fully within facility control, but the facility review should identify all three.
• The output is a response — concrete actions with owners and dates — which is the "R" in MPDSR and the part most commonly omitted.
• Reviews are confidential and non-punitive, for exactly the reason incident reporting is: a punitive review produces concealed deaths rather than prevented ones.

---

Core Patient Safety Tools

• WHO Surgical Safety Checklist — three pause points: Sign In before induction of anaesthesia, Time Out before skin incision, and Sign Out before the patient leaves theatre. It is performed aloud as a team, because its function is shared verification, not form-filling.
• Patient identification — at least two identifiers before any drug, procedure, or specimen. Bed number is not an identifier.
• Hand hygiene — WHO's five moments: before touching a patient, before a clean or aseptic procedure, after body fluid exposure risk, after touching a patient, and after touching patient surroundings.
• Safe surgery and injection practice — one sterile syringe and needle per injection, per patient, every time.
• Medication safety — the classic rights of drug administration: right patient, right drug, right dose, right route, right time, with documentation; plus special handling of look-alike and sound-alike medicines and of concentrated electrolytes.