Advance care planning- How can it go wrong?

Dr Toby Dinnen

GP Academic fellow at Cardiff University and Palliative care speciality doctor Bridgend



Dr Dinnen summarises a recently published paper on Advance Care Planning pitfalls.


Over the last two years I have had the opportunity to carry out research looking at the patient safety implications of advance care planning (ACP). ACP is a process of enabling patients to discuss and record goals and preferences for future healthcare with family, loved-ones and healthcare providers. With an ageing population plus increasing multi-morbidity and frailty, ACP is of growing relevance and importance. As medical technology evolves there are ever-increasing offers of interventions, with some of these promising to extend life. There are cultural and medico-legal pressures to offer such interventions, but this is not always automatically the right thing to do for the patient. The risks of iatrogenic harm and burden of futile treatments must be weighed against the likely benefits. When ACP works well it is a powerful intervention which enables individuals to maintain a degree of autonomy after they lose mental capacity: their priorities for future care can be shared. Less is known about how ACP impacts on patient safety in a ‘real world’ context. The aim of this project, written up as a paper for BMJ Supportive and Palliative Care,  was to look at patient safety incident reports, to understand how, why and if ACP has contributed to any patient harm, and to show how we can improve ACP in future.

As part of the patient safety research team at Cardiff University we had access to a huge database containing all the patient safety incident reports from England and Wales from 2005-2015. We identified 70 reports where an issue with ACP had caused a patient safety incident. Each one a valuable ‘snapshot’ providing insights into errors in the provision of ACP, and how this affects patients and their families. Working with colleagues experienced in patient safety and qualitative research, we use mixed methods approaches to extract as much information from the identified reports as possible.

We grouped the reports according to which stage of the ACP process the incident occurred (i.e. during creation, communication or application of ACP). In around a quarter of reports a lack of any discernible ACP caused the actual incident. In 40% of reports there was an issue with communication around the advance care planning discussions and decisions. In over a third of reports even when ACP had been completed and communicated, it was subsequently not adhered to, resulting in a patient safety incident. Several underlying themes contributed to this including staff lacking knowledge or confidence in ACP. This study demonstrates that even if uptake is improved and we had a faultless system for recording and disseminating this information to the all the relevant healthcare providers there would still be issues with how ACP is used to guide decision-making, potentially causing harm to patients.

So is ACP safe?

This study only looks at reports when something relating to ACP triggered a patient safety concern. There were 70 cases over a 10 year period for all of England and Wales that we were able to find. This is not an indication of the proportion of time ACP goes wrong, rather a reflection on what types of incidents may occur more generally.

Interestingly, in a significant proportion of these incidents, it was a lack of advance care planning that led to a patient safety incident, rather than itself ACP causing harm directly. In most cases either ACP wasn’t completed or insufficiencies in communication or application of ACP was at fault.

Conclusion: ACP is not unsafe, but this study has demonstrated incidents when lack of ACP in general or insufficient ACP had implications for patient safety. Getting this important communication wrong can cause harm to patients and relatives at a particularly vulnerable time. Improving the systems involved and tackling the underlying human factors is of course challenging, but hopefully this study provides some guidance on how best to approach this.

(Visited 1,073 times, 1 visits today)