Saturday, 21 January 2017

Dangerous charlatanry? Why not automate it…


Another interesting recent article on the BBC about thedangers of NHS trusts using referral services to review GP referrals with the intention of delaying or blocking treatment to save money. Although not this specific process, I do have painful first-hand experience of a GP referral being delayed without proper clinical evaluation.
About six years ago, during the events that led to my CKD diagnosis, I suffered a bout of severe back pain. I went to my GPs who prescribed medication, wrote me a referral to the local NHS back clinic and recommended I try their resident (private) osteopath. I took my medication, I saw the osteopath who provided a little relief but couldn’t work miracles, and I waited for the back clinic appointment.
A few weeks later I got a letter back from someone, their job title was something like ‘Senior Physical Therapist’, saying they’d reviewed the referral and decided I wasn’t worth an appointment at this time, but I’d be put on some kind of pre-waiting list for a place on the official waiting list (which I think was some sort of wheeze to game the genuine waiting list target). This ‘senior’ healthcare professional had basically read the referral from the GP and overrode his decision (which was based on clinical evidence) without examining me or even making any attempt to discuss my clinical symptoms.
During this time my back got much worse, Tramadol, Gabapentin and Diclofenac could only take the edge off. I was hunched over like an old man, I couldn’t walk properly, I couldn’t sit properly, I had to sleep on the floor. If it wasn’t for the fact that I was able to work from home I’d have been signed off sick for a prolonged period (I worked sprawled out on the lounge floor). Due to the lack of activity my weight ballooned by 15 kilos, the consequences of which are still with me today.
I went back to the GPs, I saw the other partner, who read the referral by his colleague, read the half-baked response and assessed the consequences to me which were now well past the point of medication or basic physical therapy. He asked me if I had medial insurance, I did, he wrote a referral to a private consultant, a couple of days later I had an appointment (handily he did a satellite clinic nearby), a couple of days after that I had an MRI which showed I had two bulging discs pressing on my sciatic nerve. About a week after the private referral I had a short procedure to inject cortico-steroid into my spine and the improvement was rapid.
A month or so later I did actually make it onto the waiting list for the NHS back clinic, but by this time the worst was over, my private health insurance had dealt with the acute problem. So instead of being an assessment to treat the acute problem I eventually ended up with a course of physical therapy designed to help me recover from it. Don’t get me wrong, I am grateful for that physical therapy, it taught me long-term techniques for strengthening and protecting my back which are very useful whenever I have any twinges. But the issue remains that without private healthcare I would have been left to rot in crippling agony due to a non-evidence based pseudo-scientific approach from a so-claimed health professional. So that’s how I know it’s dangerous to delay or veto referrals from without clinical assessment. It may save some money, if the patient is lucky enough to go private, or if more likely, they are unlucky and they simply die (being dead is a quick way of cutting healthcare costs).
Now I quite clearly believe the NHS shouldn’t be using dodgy processes like this, but if it’s going to use them in the name of efficiency, there are cheaper and more efficient ways. Working in Business Intelligence I know tools and techniques which could automate the assessment of referrals. These processes are not based on the first hand clinical evidence (that’s the GP consultation), they are based on textual analysis of what the GP wrote, which is something quite different.
With just a little up front involvement from the sort of healthcare ‘professionals’ currently reviewing referrals algorithms could be developed to recognise the patterns that delay treatment and the patterns that progress it. Plug those algorithms into a text mining data science tool like RapidMiner and you have a cheap way of assessing thousands of referrals an hour. Sure, from a clinical perspective it’s just as dangerous and pseudo-scientific as the current referral review process but it’s much cheaper, more consistent and also more objective as a computer doesn’t care for such things as ethnicity, religion or personal persuasions.
So my question is if the NHS is going to use dangerous processes to be ‘more efficient’ and to ‘save money’ why not race straight to the bottom and go with outright automation? With my kind of solution the NHS could do away with most of the costs of the healthcare 'professionals' currently rejecting referrals and instead spend that money on productive treatment. It won’t help the people who need treatment who get their referrals rejected on dodgy grounds, but at least the money currently siphoned off to these charlatans could be used so less referrals get rejected overall?

No comments:

Post a Comment