11 October 2016
Let’s breathe life into real-life research
It’s long been drummed into us that to definitively answer most clinical questions we need large, prospective, double-blind, randomised trials – ideally placebo-controlled.
But such trials are painfully slow, with recruitment alone often literally taking years. Neither do RCTs come cheap, and increasingly it’s pharmaceutical companies that commission and pay for medical research.
While it’s trendy to be critical of big pharma, many people I know, including me, would be six foot under without them. Further, the body of literature about most illnesses would be far worse off without pharma-funded research.
But nothing’s perfect, and downsides of pharma funding include publication bias, the paucity of head-to-head trials, and the increased tendency for pharma-sponsored research to produce positive findings for their drugs.
A BMJ investigation recently revealed that Janssen withheld data from the FDA about a faulty INR device used in one of the crucial trials into rivaroxaban (Xarelto) versus warfarin to prevent stroke.
The trial found rivaroxaban was non-inferior, but the faulty device may have made warfarin look worse than it was. It was later recalled. So after a huge amount of effort and cost, patients and clinicians are still left with more than a modicum of uncertainty.
But if UK physician and researcher Dr Ben Goldacre had his way, this type of research question would be approached quite differently. Rather than spending gazillions and years of work on a clinical trial, we’d investigate questions about drugs already in use in real life using real GP patients.
In the above case for example, patients with non-valvular atrial fibrillation could be randomised to warfarin or rivaroxaban in general practice, and followed using electronic data – from their electronic health record, MBS and hospitals – and voila, a question answered by low-cost unobtrusive research.
Speaking in Sydney last month, Dr Goldacre said this approach side-stepped the cost and red tape of cumbersome trials, and was the thinking behind a UK project called Care.Data, which he was involved in.
The idea was that the wealth of patient data stored in GP surgeries would be linked to other data and used for all manner of purposes – to guide the allocation of resources, monitor disease spread, uncover substandard facilities, streamline inefficiencies and investigate the effectiveness of drugs and procedures.
A victim of appalling communication to the public, ultimately Care.Data failed, but let’s hope the concept of real life research didn’t die with it.