Evidence-based practice - are we there yet?

We visit doctors for two main reasons: for help with issues that are causing us pain or discomfort, and less commonly, to prevent such conditions.

When we take a medicine, have a blood test, x-ray or a procedure, we expect that these will benefit us and that the benefits will outweigh any potential risks or dangers.

We naturally expect our doctors to guide us according to what is best for us, safest and most effective. But, how would we feel if there was no clear evidence that treatments we were given effective or truly safe?

I suspect that few of us would agree to such a treatment knowingly, and would require that studies be undertaken before we would consider a treatment. 

This is the fundamental premise of evidence-based medicine - today widely accepted as optimal approach to health practice, and taught at medical schools, hospitals, and universities around the world.

Evidence-based medicine relies on high quality randomised trials that provide reliable evidence on the balance of risks and benefits to practitioners and patients. 

Despite wide acceptance of the importance of evidence-based medicine, many clinicians are happy to prescribe treatments where the benefits are uncertain, and also where there are still substantial questions about risk.

In some situations, this prevents us from ever getting the data we need improve treatments.

The EVOLVE trial: a case in point

The EVOLVE trial, reported on late 2012, exemplifies this very issue. It looked at the effects of a drug called cinacalcet in people with kidney failure.

Cincalcet lowers levels of the parathyroid hormone which is increased in most people with kidney failure and has been linked to an increased risk of fracture, cardiovascular events and death. It’s been hoped that reducing parathyroid hormone with cinacalcet will reduce these risks.

The drug was approved in the US in 2004 and has been widely used since without proof that it actually prevents fractures or the other things that patients are worried about.

Around 3880 people receiving dialysis for kidney failure volunteered to participate, in the EVOLVE trial. Half received the drug while the other half received a matching placebo) in a placebo-controlled design, crucial to maximising the reliability of the data need to run the trial effectively.    

The primary results failed to show a clear benefit for any patients, despite a substantial reduction in parathyroid hormone levels.

Ironically, analysis of the trial suggests that that it had failed to detect a real benefit because of two reaons: a large number of people randomised to take the drug had stopped taking it; and 20% of those people allocated to the placebo were prescribed the drug separately by their doctors.

When bias towards a treatment can be harmful

This leaves us with an ironic situation where the conviction of doctors and patients participating in the EVOLVE trial about the benefits of the treatment prevented the trial from demonstrating a benefit.

This also means that some of the potential risks, such as the risk of cancer, will also have been underestimated. Considering the effort of so many people and the millions of dollars involved in running the trial, we are not much further along in understanding the effects of this drug.
The implications for evidence-based medicine is that without more stringent and robust trial methodologies, doctors will continue making decisions based on poor quality data, potentially doing more damage than good.

When good intentions cause more harm than good

The US FDA approved the drug based on its effects on a blood test, and safety data derived from around 1000 people. Is this adequate for a drug likely to be used by hundreds of thousands of the sickest people in our society?

A rational look at this question suggests more data is needed before approving a drug like this.
Approval is likely to have stemmed from a range of factors, but one of key ones is pressure from doctors and their patients. This intention to ‘help’ has in this case caused more harm than good.

We need to raise the bar

The bar for approving new drugs needs to be higher and more consistent around the world if we are to practice what we preach when it comes to evidence-based treatments and not waste precious dollars and lives, especially in an era where the cost of healthcare is increasingly unsustainable.

Any drug being used to prevent fractures, cardiovascular events or other serious outcomes, needs to be backed by true randomised trial data to show that it is effective and safe – before being approved for use.
This has already been recognised by regulatory agents in the area of diabetes. Clear guidelines path for developing new medications were developed after a storm of criticism surrounding a publication on rosiglitazone, one of the most widely used diabetes drugs at the time shown to actually increase the risk of heart attack.

Using a similar intelligent, planned approach to the registration of drugs in other health areas will reduce the risk of repeating our mistakes, and may allow us to evolve to a true and consistent evidence-based approach to healthcare.
How do we do this at a reasonable cost is a question of intelligent design of clinical trials, transformation of the clinical trials industry, and a story for another day.