GPs are not infallible, but we deal with most health problems reasonably effectively. Some of these are major – in our practice in the last few weeks we’ve diagnosed a case of HIV; a malignant melanoma; two new diabetics; an unusual hip pathology in an eight-year-old; two rare tumours; as well as a more commonplace uterine cancer.
We’ve also managed lots of other significant stuff – palliative care for the dying; uncontrolled asthma; falls by the frail elderly; crippling arthritis; adults and teenagers with mental health crises.
And amid all this, we’ve also seen many patients with issues that don’t require a fully trained doctor – sore throats, dental abscesses and contraception queries to name but a few.
With GP numbers falling year on year (the Tories’ pledge of 5,000 extra GPs by 2020 has been quietly forgotten) it is becoming ever more pressing to devise ways of filtering out the estimated 20 per cent of consultations that could be dealt with by a different professional. The challenge is how to do so without compromising patient care. Five of our recent major diagnoses – including the HIV case – presented in nebulous ways that appeared trivial. Were these patients to have been diverted to someone without medical training, their conditions could easily have gone unrecognised.
A neighbouring practice has adopted an increasingly popular solution. Every patient requesting an appointment speaks first to a GP on the phone. Those who need to be seen are booked in for appointments, others have problems that can be dealt with during the call, and a proportion are sent to allied professionals such as nurses, physios or pharmacists. This has reduced face-to-face GP consultation rates considerably but not workload, with the huge list of phone calls soaking up vast amounts of doctor time.
This is the background against which to view Health Secretary Matt Hancock’s enthusiasm for “GP at Hand”, the smartphone app currently being used by private company Babylon Health as the front door to its primary care service for more than 25,000 patients in London (among them one Mr M Hancock, MP). Hancock attracted a chorus of criticism last week for saying he wanted GP at Hand rolled out nationwide. He didn’t help himself by his bizarre decision to make the announcement from Babylon’s head office – spectacularly ill-judged partisanship that will have dismayed the 20-odd other technology companies currently developing similar “solutions” for primary care.
But it is Hancock’s naive faith in the ability of technology to resolve this crisis that provoked the most ire. Babylon touts GP at Hand as “artificial intelligence”, and caused controversy in June when it published data that purported to show its software performing as well as a human doctor, albeit under highly rarefied conditions. The ensuing social media video posts demonstrating GP at Hand missing barn-door diagnoses such as heart attacks and sepsis should have given Hancock pause for thought.
We have been here before with NHS 111. For all the talk of artificial intelligence, apps such as GP at Hand are essentially based on algorithms similar to 111’s Pathways, which attempt to identify cases that need a doctor’s involvement. However, the subtlety and variability of patients’ symptoms frequently confound algorithmic analysis, leading to the risk of significant conditions being missed.
NHS 111 mitigated this by setting low risk thresholds – anything that might be serious gets handled as though it is. The result has been a dramatic rise in A&E referrals and ambulance call-outs that has brought the NHS to its knees. No one knows where this app’s risk thresholds are set: that is crucial information, both in terms of workload and patient safety, to be established before it gets foisted on the NHS.
Until then, away from the emperor and his new artificial intelligence clothes, we desperately need more human GPs. Any solutions to that conundrum, Mr Hancock?