Professional Practice, Ethics, Safety and Risk Management in Healthcare
This domain tests judgement: how a biomedical engineer behaves when a device, a colleague or a commercial pressure creates a patient safety conflict. It combines professional ethics, incident investigation methods and formal risk tools.
Ethical framework
- Four principles: beneficence, non-maleficence, autonomy (consent and information) and justice (fair allocation of scarce equipment).
- Professional duties include competence within scope, honesty in reporting, declaring conflicts of interest, confidentiality, and escalating concerns about unsafe practice, including to a regulator if internal channels fail.
- Public safety overrides commercial or reputational convenience.
Risk management tools
- FMEA assigns severity, occurrence and detection scores, with RPN = S × O × D. RPN prioritises but does not decide: a high-severity item with a low RPN still demands attention.
- Root cause analysis uses structured timelines, the five-whys or fishbone, and distinguishes active failures from latent conditions. The Swiss cheese model explains how multiple defences must align to fail before harm occurs.
- A just culture separates system error from reckless behaviour, which is what makes reporting viable.
- Risk control hierarchy: inherently safe design, then protective measures, then information and training. Training alone is the weakest control and should not be the primary answer.
Incident reporting
- Device adverse events are reported to the manufacturer and to the authority; in Hong Kong this is the Medical Device Division's incident reporting channel under the MDACS framework. Reporting is not an admission of fault, and the engineer usually contributes the technical narrative.
Worked example
A syringe pump has an occlusion that occasionally fails to alarm. FMEA scores S = 8, O = 3, D = 4, giving RPN 96. Adding an independent downstream pressure sensor changes occurrence to 1 and detection to 2, giving RPN 16. Because severity remains 8, the team still cannot accept residual risk without documented justification, so they add a startup self-test and clear labelling. This shows the severity floor argument.
Traps that cost marks
- Treating an RPN threshold as a pass/fail gate and ignoring severity.
- Accepting retraining as the sole corrective action for a design-induced error.
- Blaming an individual when latent system conditions caused the failure.
- Failing to declare a supplier relationship during an equipment evaluation.