Table 4.

Key messages and recommendations

Methodology and theory-building There is value in drawing on different perspectives and frameworks to explore the nature of problems before attempting to offer potential solutions.	Sharing findings from analysis of patient safety incident reports directly with stakeholders is an effective prompt for discussing gaps between official accounts and day-to-day experiences. Synthesis of complementary approaches (eg, the realist context–mechanism–outcome model with SEIPS) helps cross disciplinary boundaries and consider intersectionality between different perspectives.
Human factors issues Interventions can only be targeted at underlying mechanisms driving human factors issues when problems are studied in depth and in context.	As people experience different events, socially constructed learning in the form of sense-making, or meaning-making occurs leading to cycles of thought and behaviour that are refined and replicated according to experiences in future events. It is relatively rare that addressing knowledge gaps alone will make a difference in complex situations. Better integration of human-centred co-design principles and informal learning theory into future attempts at improvement are needed to increase the likelihood of success.
Safety in out-of-hours palliative care Problems are created, defined and constructed by people in ways that generate variable patient outcomes, experiential learning (desirable or otherwise) and consequences for future healthcare.	Optimal care is dependent on ‘interpersonal glue’: often mediated by trust, empowerment and ability to tell whether a situation demands a standardised, customised or flexible response. Optimal care and a holistic approach to safety in palliative care are seen to commonly require in-the-moment enacting of workaround strategies to manage risk in complex and adverse conditions.

SEIPS, Systems Engineering Initiative for Patient Safety.