Doing the Groundwork before Designing a Trial

Recommendation 1: Before designing the trial, define the population at risk, the case and the impact of the potential target agent. 

Acute Respiratory Infections (ARIs) are usually benign self-limiting conditions, especially if caused by viruses. This characteristic should not make them attractive to the pharmaceutical industry or governments. 

Share Trust the Evidence

However, they are very common, and they all present as a syndrome. This is an aggregate of signs (e.g. fever, measured by a physician) and symptoms (e.g. cough) with a sudden onset. The cause of the syndrome is not easily definable, as on average one third are caused by rhinovirus, one third by a mixed bag of agents, and one third are of unknown origin. There are n number of agents involved, often unknown, and in about 20% of cases, two or more agents are at work in the syndrome.

Their ubiquitous nature plays well in the hands of those selling the latest remedy, as it means huge numbers and even larger, if you want to prevent things. In theory, the whole world is the market. Because viruses are unstable things (yes, things, not living beings), they mutate continuously. Creating a huge market and repeated interventions to catch up on the latest mutation. Yummy!

Donate

The syndrome also plays in the hands of governments with dishevelled health services, as they can blame any sudden crisis on the so-called “Flu”.

The point where the two interests converge is as follows: to avoid the impending inevitable “flu” crisis, you must swallow pills or be vaccinated. Big bucks for pharma and “job done” for governments pushing vaccination campaigns with extra cash.

So the absolute first requisite for an ethical trial is an analysis of the problem at hand by defining the risks of the target population for the critical outcomes of interest. The serious, adverse health outcomes that occur during or after critical illness, such as death, organ failure, or long-term physical dysfunction. But number one remains to define a case in a credible manner.

These variables need to be defined prospectively and backed up by health care footwork (a medical history), as well as robust testing done and reported in an appropriate and credible way. Current surveillance does not do this, as it is a nonlinear mish-mash of different sources and prone to ascertainment bias (search and thou shall find).

If you are considering trialling a vaccine for the complications of, say, influenza, you need figures for pneumonia cases where there is solid proof that influenza viruses were present in large quantities before or at admission to hospital, not several days later. The coding of such verified cases must not be conflated with “flu” cases to make it look much worse than what it is. 

This afternoon, we will summarise TTE posts showing how these simple rules have been ignored or distorted. 

This post was written by two old geezers who eye the chiasm between politics and scientific evidence.