Narrative and Numbers

Ben Hoban is a GP in Exeter.

Patients adopt a variety of approaches when talking to their doctor: they may tell a story, with all the contextual details that by turns help, hinder and misdirect; they may describe their symptoms, inviting us to provide an interpretation; they may name the problem, trusting that terms like vertigo and wheeze mean the same to both of us; or they may just quote numbers, that ultimate form of abstraction, whether referring to their blood pressure, their waistline, or how much pain they’re in. Communication takes place across the whole of this spectrum, at one end rich in meaning but strictly personal, and at the other, precise but generic. As we move from narrative to numbers, we limit the opportunities for misunderstanding, but also for genuine understanding; we communicate more clearly, but we also communicate less.

This tension crops up in other areas too. The Quality and Outcomes Framework (QOF) considers only quantitative measures of patient care, for example. All things being equal, it is preferable to have lower rather than higher blood pressure, but not meeting a target could indicate either a poorly performing practice that doesn’t know the rules, or an excellent one that looks beyond them and empowers patients to make their own decisions: there is complexity here which QOF points cannot adequately capture. Similarly, it is good to learn from our patients and colleagues, but the questionnaires approved for collecting feedback invite bland, standardised responses which generate another number: what does a percentage score really mean, other than that a test has been passed or failed? Does this really inform our learning? Patients attending an NHS health check are in the same position, having reached a point in life where age or changing circumstances prompt them to consider their health more carefully, and wanting to know where they stand. The outcome of their appointment, an estimate of 10-year cardiovascular risk, is so abstract that it can take considerable time to unpack, and the officially endorsed advice they receive reflects the wholesale price of statins more than any meaningful idea of whether they are in reasonable health for their age. Given the limited utility of numbers in healthcare, how is it that they have come to dominate our working life so much?

The most obvious answer is that we privilege a scientific, biomedical outlook, which views measurable realities as more valid than unmeasurable ones. Evidence Based Medicine relies on quantifying the effectiveness of potential treatments in order to know which we should use. This is of course good in principle, but still depends on measuring meaningful outcomes in a representative group of patients: it doesn’t necessarily tell us whether the person in front of us will benefit at all, or on whose terms. It is easy to find fault with what feel like more subjective approaches, but counting and measuring are prone to their own biases; the idea that basing decisions on numbers alone is somehow objective is in fact known as the McNamara fallacy.¹

A natural consequence of this belief in the superiority of numbers is our tendency to measure things, even when it is unlikely to help. The streetlight effect calls to mind a home-owner who has lost their keys somewhere dark, but looks for them where the light is better.² When we look for a diagnosis by measuring someone’s haemoglobin, liver function or whatever, and then reassure them that the results are normal, we are effectively saying that all is well, the keys can’t be lost because we looked under the streetlight and couldn’t find them! Faced with the needs of someone who is unaccountably fatigued, or in pain, or otherwise struggling with life, it is hard to gaze with them into the abyss and acknowledge our own helplessness; far easier to do some blood tests instead. The abstraction of numbers confers a certain distance from these everyday horrors, and a sense of agency in a job that often robs us of agency. Atul Gawande advised doctors who want to stay healthy and effective: count something.³

We walk a tightrope in medicine, balancing every day the unique and complex needs of individual patients with the standardised requirements of the rule-book that governs their care. There is danger in tipping too far in either direction, in concentrating too much on either the narrative or the numbers, but I believe that we are currently tipping towards the latter, and that this represents one element of the current crisis in the health service. The increasing use of laboratory investigations in primary care,⁴ for example, generates significant additional work for everyone involved, but reflects a wider cultural shift from a more personal and nuanced model of care to one that is more scientific and intolerant of ambiguity. If we want to regain some measure of control over our working lives, we can either embrace this shift or push back against it. Knowledge, certainty and structure can all contribute to a sense of agency, but it is a contingent, brittle kind, which does not long survive contact with everyday life. A more robust variety grows instead from understanding, engagement and responsiveness to the needs of the people we are trying to help, and sometimes the numbers can be left to look after themselves.

References:

1. Michael H Basler, Utility of the McNamara fallacy, BMJ2009;339:b3141 doi: https://doi.org/10.1136/bmj.b3141
2. Pat Croskerry, The Importance of Cognitive Errors in Diagnosis and Strategies to Minimize them, Med. 2003; 78: 775-780
3. Better: A Surgeon’s Notes on Performance, Atul Gawande, Metropolitan books 2007
4. SullivanJ W, Stevens S, Hobbs F D R, Salisbury C, Little P, Goldacre B et al. Temporal trends in use of tests in UK primary care, 2000-15: retrospective analysis of 250 million tests BMJ 2018; 363 :k4666 doi:10.1136/bmj.k4666

Featured photo by Raphael Schaller on Unsplash