There’s been quiet a lot written in the UK recently about failures in an attitude to care within organisations, and how this is a major root cause of poor healthcare and avoidable death.

I was wondering about the links between caring, I guessed best expressed as “empathy”, and how patients or their relatives percieved their care, and also, the harder outcomes. Like survival. Was a person-person, “I recognise your pain”,  honest relationship with your doctor pleasant, and did it save lives?

For patients or their families [P] does physician empathy [I] affect patients satisfaction with care and mortality/morbidity [O]?

Well, it turns out there are systematic reviews on this subject. I dragged one in cancer care up. Using a FAST appraisal, the study team found 394 potential studies and cut this down to 39 using two reviewers, independently. They performed an appraisal of the risk of bias from the studies rather weakly, focusing on heterogeneity of study design and measure rather than patient/provider selection biases. Their synthesis was entirely narrative – not to be unexpected given the very disparate nature of the measures and designs and outcome groups – but seems sensible, and the outcomes can be transferred with caution: children and families are not little adults, and healthcare providers from different disciplines are not all the same.

The main outcomes were pretty well summarised by the authors: “clinician’s empathy is associated with higher patient satisfaction, better psychosocial adjustment, lesser psychological distress and need for information, particularly in studies with patient-reported measures and retrospective designs. On the contrary, results indicate that empathy is not related to patient empowerment (e.g. medical knowledge, coping)”.

So – empathy does seem to be linked to satisfaction. But what about harder outcomes? And what about making it happen more (if it’s good) – does this just reflect a natural talent or can it be taught?

There is a Cochrane intervention review which addresses these issue: again with a very wide range of patients, health care providers and outcomes under consideration. Of the 43 included trials 11 addressed satisfaction, and showed only  a small and heterogenous improvement in these measures. Mortality outcomes were not measured in any trial, with the majority of outcomes focusing on improvements in mental health and wellbeing – which did improve – by one quarter of a standard deviation. (This is considered a small-moderate improvement; but a good size for an educational intervention.)

So do we conclude that the Prince of Wales was right and that we just need to learn to care more?

It would be daft to say we shouldn’t shift the quality of caring upwards, but the evidence for how to make it happen and for that to really make a difference is not as string as I’d like.

– Archi