The Salk Vaccine Field Trials

Lessons in the Design of Experiments

The U.S. government undertook the polio vaccine field trials in 1954 in an effort to curb the spread of the highly infectious poliomyelitis or infantile paralysis disease.  

The trials were described as among the most elaborate and most publicized of its kind ever done. I am drawn to it because of the insights we learned in the way we devise plans to prove whether or not a proposed solution is effective. 

The methods we employ may have a significant bearing on the outcome of the study and thus deserve careful consideration. 

Oftentimes, the goal of a medical trial is to ascertain the efficacy of a particular treatment. An adequate way of doing this is by applying the treatment on two groups and comparing each groups’ outcomes. It is crucial that the groups be similar to each other as much as possible in order to reasonably infer that any difference in the outcome between the groups can be attributed solely from the effect of the medical treatment and not from any other cause. 

Be cautious of other causes

Aside from the attention given to the study of the main treatment as a possible cause of the outcome under study, it is also important to put some effort in finding other possible causes that might influence the results of the trial.

If a trial results in an outcome that favors the possibility of the treatment being effective, we want to be very sure that the success can only be attributed to the treatment being the only cause and not from any other inconspicuous factors.

Conversely, if the results show that the treatment failed, we want to be sure that there exists no other factor that may have contributed to the undesirable outcome and that the treatment’s ineffectiveness was the primary cause of the failure.

A SEASONAL DISEASE

It is said that the incidence of polio varied from year to year. There are years when the number of people who contracted the disease would soar and there are years when that number would plummet.  It would be unwise to infer in favor of the treatment if none of the treatment patients contracted the disease.

Remember that the treatment itself should be viewed with skepticism at the start of the trial. And it should be placed in equal footing along with other possible factors of a polio infection. This is the reason why the study is being done to begin with.

If the polio vaccine was ineffective in actual reality and the observed abrupt decline in polio incidence among the treatment patients just so happens to coincide with a seasonal decline in the prevalence of the disease, then any inference alluding in favor of the treatment should cause us to demur.

A DISEASE OF THE HYGIENE

The Salk vaccine trials divided the trial participants into two groups. One group will receive the treatment while the other group did not.  This arrangement would have presented an ethical dilemma as it would be controversial to withhold a possible cure to children in the control group who would not receive the vaccine.  But the ethical problem was settled when parental permission became a requirement for vaccination. The government assigned children whose parents gave consent to the treatment group while those who did not went to the control group.

Assigning children based on parental consent is biased against the vaccine. Parents of well-to-do families were more likely to consent to having their child be vaccinated.  And so children assigned to treatment groups were more likely to come from more affluent families and thus were reared under comparatively more hygienic surroundings.  Their immunity is not as built up to fight against the disease.

The contention is that children who were assigned to the treatment group are more prone to get ill as compared to their control group counterparts who might have developed active immunity from the polio virus due to previous contraction of the disease.

Since it is not known whether or not the vaccine is effective, if the outcome of the trial show that the proportion of children who contracted polio was greater in the treatment group as compared to the control group we might make a mistake by imputing this to the inefficacy of the vaccine where in fact it just so happens that the treatment children were particularly more vulnerable and that the control children were just more resistant to the disease. This is what is meant by the set up being biased against the vaccine.

We have thus far described some scenarios that explain how the results of a study may be compromised by the approaches that are applied in measuring any possible relationship between the independent factor and the observed outcome.

What is common among these scenarios is another factor whose presence has the potential of  demonstrably distorting the outcome of the experiment.

These factors are unforeseen and can sometimes come from sources that are external to the variables of interest.

The confounding variable’s influence gets “mixed in” with that of the independent variable’s influence. So it defeats the aim of completely isolating the effect of a treatment upon the outcome of interest. Therefore, any statement regarding the perceived association or relationship between the treatment and the outcome would be inaccurate.

What do we get from this discussion? It reminds us that a carefully designed experiment will benefit greatly by conducting a thorough determination and examination of possible extraneous actors that might influence the observed effect of an examined treatment.

Thinking out of the box and maybe consulting with peers to elicit their views about what other variables outside of the independent factors one has already listed might garner some ideas about what to watch out for that might affect the results of the study.