I don’t know how others celebrated International Statistics Day, but I decided to dedicate an entire morning clinic to the cause.
If I’m honest, I usually drift into the old habit of seeing each patient as an individual, but for once I made a concerted effort to treat everyone as a statistic.
I didn’t want any atypical patients ruining my outcomes, so I worded up my receptionist Joan about who to book in.
“Please exclude anyone who has heart failure, doesn’t speak English or is on two or more medications.” She raised a dubious eyebrow. “Oh, and I want them allocated randomly.”
Joan’s eyebrow went decidedly bell-curve. “Randomly how?”, she protested. I gave her a coin to toss and a friendly wink. She means well.
The first patient on my list was “Baby 4136 (gender blinded)”. Joan had received my memo about de-identification.
Mum had some concerns about her infant’s undescended testis – oops, lost the blinding already – but I was more interested in its growth charts. They were beautiful. A head circumference straight down main street.
At mum’s eventual insistence, I popped little 4136 onto the bed and palpated his unusually-shaped testes – or to use statistical parlance, I adjusted his odds.
He had a hydrocoele on the right, twice the usual size, while his left scrotum was entirely empty.
I started with the good news. “On average, his testis size is normal.” I then reluctantly broke the data down into two discrete variables and sent off a referral.
The next patient was of the opposite gender (Joan’s coin had done its job), aged 50-59. Her opening line was “I’m feeling pretty ordinary.”
“Excellent!” I replied. An awkward pause ensued, until I realised her mistake.
“Oh, do you mean worse than ordinary?”
It comforted me knowing that thousands of others were probably also insisting on similar accuracy on International Statistics Day.
She continued “As you know, my ex-husband has been gravely ill, doctor. Do you think there’s any hope?”
“That depends on what you’re hoping for,” I replied with calculated precision. I was doing the day proud.
She went on to ask for a carer’s certificate, but just then I recalled a recent study linking hypertension with having an ill family member. You might call it a relative risk.
The quoted expert had said he was 95 per cent confident GPs weren’t measuring enough BPs in these people.
“Before you continue…” I interrupted, cuff poised. Sure enough, her systolic BP was 141. Twice.
The new US hypertension guidelines recommend starting medication at ≥140, so I did, using a series of bar graphs to indicate each side effect probability.
After 30 minutes’ exhaustive discussion (I had to explain multiple regression twice), she left, only to phone back later.
“Any chance you’ve sent off that carer’s certificate?”
“Yes,” I replied. I hadn’t sent it, but the precise question was about chance.
There’s a difference between a lie and a damned statistic.
This article was first published in Medical Observer, April 2019