Quality improvement methods in child health — formative SAQs

SAQ 1 — Designing a paediatric QI project (10 marks)

A neonatal unit wants to reduce the proportion of infants receiving antibiotics beyond 48 hours when no infection is proven. The current rate is 40%. The registrar asks you how to design the project. [1] [8]

Questions

1. State the Model for Improvement and write a specific, measurable, time-bound aim for this problem. (3 marks) [1]
2. Describe the balanced set of measures you would track, defining outcome, process and balancing measures for this project. (4 marks) [1]
3. Outline how you would test a change idea using a PDSA cycle before any unit-wide rollout. (3 marks) [2] [8]

Model answer

Model and aim (3). The Model for Improvement asks three questions — what are we trying to accomplish, how will we know a change is an improvement, and what change can we test — then runs Plan-Do-Study-Act cycles. Aim: reduce the proportion of infants on our unit receiving antibiotics beyond 48 hours without a proven infection from 40% to 20% within 6 months. [1]

Balanced measures (4). Outcome: proportion of infants with antibiotics continued beyond 48 hours without proven infection. Process: proportion of antibiotic courses with a documented 48-hour stop-or-review decision. Balancing: proportion of infants with a subsequently proven infection who had antibiotics stopped at 48 hours (a re-start rate, to detect under-treatment), plus length of stay. [1]

PDSA test (3). Plan: test a 48-hour antibiotic stop-review checklist on one team, one shift, defining the data to collect. Do: run the test and record what happens. Study: compare the result to the prediction and gather staff and family feedback. Act: adopt, adapt or abandon — then scale to more shifts if successful before any unit-wide rollout. [2] [8]

SAQ 2 — Appraising the evidence and reading the data (10 marks)

A colleague reports a single-centre before-after study claiming a new protocol reduced central-line infections, and shows you a run chart with no clear baseline. The unit average improved but the rate for Indigenous infants did not. [8] [9] [10]

Questions

1. Why is a single before-after comparison a weak design for attributing improvement, and how does a time-series run chart address this? (3 marks) [8] [9]
2. State two standard run-chart rules for detecting special-cause variation and explain why you should not react to common-cause variation. (3 marks) [9]
3. Identify the reporting and equity problems in this project and state what should change before it is published or adopted. (4 marks) [10]

Model answer

Weak design (3). A single before-after gap cannot distinguish a true intervention effect from regression to the mean or a secular trend already underway. A time-series run chart with a clear baseline and a marked change point shows whether improvement coincided with the intervention and whether it sustained — far stronger evidence. [8] [9]

Run-chart rules (3). A shift is six or more consecutive points on one side of the median; a trend is five or more points all rising or all falling; a run is five or more consecutive identical values off the median; an astronomical point is an obvious outlier. Common-cause variation is the noise inherent in the process; chasing it by tweaking the system makes performance worse, not better — only special-cause signals warrant investigation. [9]

Reporting and equity (4). The project lacks a clear baseline and is reported without enough detail to appraise or reproduce; SQUIRE (Standards for QUalIty Improvement Reporting Excellence) exists to fix this, so the report should state the aim, measures, tests and sustainability. The equity problem is an aggregate success that hides a subgroup that did not improve — the data should be disaggregated by subgroup and equity made an explicit aim and measure before the project is claimed a success or spread. [10]