Designing and defending a paediatric implementation plan — OSCE

Station brief (8–10 minutes)

A paediatric emergency department has a college-endorsed sepsis guideline, yet only 52% of eligible children receive the first antibiotic within the recommended hour. The guideline has been emailed and posted on the intranet for a year with no measurable change. The head of department has asked you, the senior registrar, to present a proper implementation plan to the departmental quality and safety committee and defend your method. You will be questioned on how you diagnose the barriers, which frameworks you use and why, how you choose strategies, how you measure success, and how you protect equity and sustainability. Do not invent mandated adoption targets. [1] [3]

Tasks for the candidate

1. Reframe the problem as an implementation problem and explain why the email-and-intranet approach failed. [1] [9]
2. Describe how you would diagnose the barriers using the Cabana framework, giving an example from each tier, and assess the context using CFIR. [3] [4]
3. Explain how you would select implementation strategies matched to the barriers, naming the resource you would use and why you would build a parsimonious bundle. [8]
4. Describe the outcomes you would measure, distinguishing implementation outcomes from clinical outcomes, and state why fidelity is essential. [7]
5. Explain how you would protect equity and plan for sustainability from the outset. [5] [2]

Expected performance

Must hit. Names dissemination versus implementation as the reason the email approach failed; defines implementation science; uses Cabana to diagnose knowledge, attitude and behaviour/external barriers with a concrete example of each; names CFIR as a determinant framework for assessing context; maps barriers to strategies from the ERIC compilation and justifies a parsimonious bundle over an untargeted pile; distinguishes the implementation outcome (adoption, fidelity, sustainability) from the clinical outcome and states that fidelity separates a poor fit from poor delivery; commits to reporting reach by subgroup and to a sustainability plan with a named owner, data and policy levers. [1] [3] [7] [8]

Merit. Uses RE-AIM as the evaluation frame and explains each dimension; cites Morris on the seventeen-year lag to justify the discipline; cites Grimshaw on tailored strategies outperforming untargeted bundles; names a hybrid study design appropriate to implementation; distinguishes implementation science from quality improvement and clinical research; engages a family partner and a senior sponsor in the team; names Aarons on implementation leadership as a prognostic factor for sustainability. [2] [5] [9]

Fail. Accepts more dissemination (another email, a printed copy) as the solution; proposes a strategy pile-up without assessing barriers; measures only the clinical outcome and never adoption, fidelity or sustainability; reports only an aggregate average without subgroup disaggregation; or invents mandated adoption quotas. [9]

Sample candidate structure

"This is an implementation problem, not a knowledge problem. Emailing and posting the guideline is dissemination, and dissemination alone changes behaviour almost nothing — Morris showed evidence takes about seventeen years to reach routine practice even with good guidelines. My aim is to move the proportion of eligible children receiving the first antibiotic within one hour from 52% toward a target we set with the committee. I will use the knowledge-to-action cycle: identify the gap, adapt the bundle to this department, assess the barriers, choose and tailor strategies, deploy with fidelity, measure outcomes, and sustain. To diagnose why the guideline is not followed I will use Cabana — knowledge, attitudes, behaviour and external barriers — with structured staff interviews, and I will assess the context with CFIR, its five domains, because a copied intervention fails when context is ignored. I will map each barrier to a strategy from the ERIC compilation, building a parsimonious bundle of two or three matched strategies — likely an opinion-led escalation pathway, a sepsis trolley and visible decision support — because Grimshaw showed tailored strategies outperform untargeted piles. I will measure three layers: the clinical outcome, but critically the implementation outcomes from Proctor — adoption, fidelity, penetration, sustainability — because without fidelity I cannot tell a poor fit from poor delivery. I will evaluate with RE-AIM, and I will report reach by subgroup so an improving average cannot hide an untouched Indigenous or disadvantaged subgroup. Finally I will plan sustainability from the outset — a named owner, ongoing data, the bundle embedded into policy and orientation — because adoption decays the moment the champion rotates." [1] [8] [7] [9]