On the Pacific islands of Samoa, a fatal medical error and the spread of false stories about the measles, mumps, and rubella (MMR) vaccine has led to the deaths of 68 people due to measles, most of them children. Over the last year, the combination of the deaths of two infants due to negligent preparation of the vaccines and carefully-crafted ads questioning the safety of the MMR vaccine led to vaccination rates plunging from 60-70% to 31%. This deadly combo of tragic error and carefully-crafted spin is causing the worst health outcome from a societal perspective, the death of children, at a population level.
When we talked in the last blog post about understanding causes, we introduced a basic public health causal diagram to show how in medical research the goal is to be able to conclusively say that A causes B. In this case, at an individual level, the disease cause and effect are clear:
This simple diagram (it’s called a DAG, or directed acyclic graph) of disease cause and effect is what drives public health interventions and policy. Measles is a highly contagious and sometimes fatal disease with a clear pathology, the measles virus. A vaccine was developed in 1963, so now rather than it spreading through populations where 30% of infected individuals die, the measles death rate is now less than 0.5%. The MMR vaccine is a common preventive intervention and could be shown on the causal diagram this way:
In a confusing twist, the arrows are meant to reflect both causation and its opposite, prevention, e.g., the measles vaccine causes a lack of measles. When we scientists run our statistical programs on models like this, the first arrow would have a negative number attached to it while the second would have a positive one. Although the diagram may be a little befuddling, this is a robust and clear finding. The vaccine acts directly to boost the body’s immune response to the measles virus. It is almost universally effective in preventing measles.
Unfortunately, as we see from the tragedy in Samoa, understanding medical science is not that easy. Yes, the measles vaccine works, but for Samoa as a population right now, it’s not very effective because the spread of misinformation has affected parents’ ideas and behaviors about vaccination. Health behaviors affect health outcomes, so the causal situation is very nuanced. This is really important, because we have to understand this situation clearly to figure out how to deal with it.
On the causal pathway, it looks more like this:
And when you look at it from the perspective of multiple levels of causation, you can see how the individual actions (get a vaccine or not) are affected by environmental causes, including the cultural (in this case media/tech and scientific) context.
Why do I keep harping on this? Because from a public health perspective, most video game research is incredibly lacking in nuance because understanding media and its effects is not a specialty of health scientists. There’s often no nuance. It’s exasperating to see another study about the dangers of “screen time”, as if lumping the concepts of TV, phones, and video games on causal pathways made any sense at all. Can we expect Call of Duty: Modern Warfare to have the same effects as calling your mother or watching the news? Yet time spent with “screens” is often the thing we measure! It’s truly disturbing, and we should all be grateful to researchers like Przybylski and others at the Oxford Internet Institute who are careful to reveal this nuance by openly replicating studies using better designs.
Exposing the holes in public health research on media and technology use reflect an overall paradigm shift toward making sure science is done openly and is free from bias. Bad science does sometimes make it past the peer review process. The original article by Wakefield and colleagues linking vaccines with autism in a series of children (by the way, case series are considered one of the lowest forms of scientific evidence!) was later retracted, and it was discovered that Wakefield had committed financially-motivated fraud. Bad actors cause the worst and hopefully rarest kinds of bias, but unfortunately biased research and reporting are still out there and still affecting population health.
Thankfully, biomedical science does provide us with some basic tools to judge the quality of health research (or research outside of medicine/public health that’s still relevant). These last two blog posts talked about the underlying idea of causation and why we can’t ignore all the environmental factors that affect health at a population level. The MMR vaccine could have prevented most of the 68 deaths in Samoa, but many other factors interfered. Next time we’ll talk about the hashtag #itsaysinmice, PICOS, and why scientific reporting needs standards.
 In developed countries