Read on Twitter
It is incredibly frustrating to read academic papers covering topics where you know what you are talking about. This recent paper on digital first GP consultation, for example: https://www.jmir.org/2020/6/e18203/pdf
A thread...
A thread...
The paper is by Salisbury, et. al. of Bristol University's Center of Academic Primary Care.
The worrying word there is "Academic", which, for reasons I will explain, suggests their goal is papers not practical advice...
The worrying word there is "Academic", which, for reasons I will explain, suggests their goal is papers not practical advice...
It is an important topic as GPs have moved very rapidly to total triage approaches as a way of avoiding face to face contact during Covid. So it is important to understand the implications of that...
But the paper isn't an observational study of real world evidence. It is a modelling study based on hypothetical results from the academic literature some of which were never robust in the first place...
The paper concludes that "Digital first approaches to primary care could increase general practice workload unless stringent conditions are met." this has already been quoted as a warning in the press...
It isn't a terribly bad conclusion, just a useless one. Those of us who work in the area know that getting benefits requires careful system design and management. Digital first is not a magic bullet...
But much of the data behind the model and its conclusions was based on an academic literature review not actual data on any digital consultation systems currently in use...
Take this input "digital consultations are predominantly used by patients in the 20-44 age group". That is certainly widely believed. It is also just plain wrong...
This is the demographic profile of about 3m @askmygp requests (patients have a choice of starting requests online or by phone, about 70% start online)...
That matches the demographics of normal GP demand and is not disproportionately dominated by the young. This data is available to researchers but this study chose to use old and unreliable academic papers instead...
And, as far as I can tell, the study ignored the benefit that widespread use of digital first tools makes access to practice by phone far, far easier for those who don't want to go online...
The study also claims that the biggest worry might be that easier access will increase demand. This factor seems to dominate others in their model and they warn that increased demand might nullify the benefits to GPs...
Again, though, this is a theoretical worry and ignores evidence from current systems that strongly suggests it isn't true and that the benefits to patients and GPs is large...
The paper argues "the twin goals to improve patient access and to manage workload pressures on GPs are likely to be in tension"...
But this argument ignores the actual evidence from operational online systems that the major source of inefficiency in the old "every request gets an f-to-f appointment" is that the majority don't need one...
And their problems can be solved quickly with shorter and simpler online or telephone responses. In @askmyGP practices (where ~90% of responses are done on the same day) only ~30% of patients even request f-to-f...
This observation alone invalidates much of the theoretical model which, by design, ignores the benefits to GP workload of total triage which allows GPs to spend their time based on need not "demand"...
An approach that also benefits patients who typically get much faster responses and are typically much more satisfied. We have > 20k pieces of patient feedback every month that demonstrate that...
Last month >93% of that feedback on the standard FFT question was "good" or "very good". And that data is also available to researchers who want to analyse it...
In short, theoretical models can help understand what factors are important to make online consultation useful for GPs and patients. But, if the inputs to the model ignore the real world evidence from systems...
several of which have processed millions of requests over several years, then the usefulness of the conclusions are low and not that relevant to actual practice in primary care.
