So much information abounds the Internet these days, information that can be confusing. The confusion can result partly from the fact that recommendations from professionals are changing frequently based on new information, but there also is a lot of rubbish circulating about. Sometimes, the low quality of the information is obvious, but a good deal of the rubbish is presented in a way that can appear scientific. Thus, you can find articles and recommendations on a wide list of topics, ranging from vaccination during pregnancy to whether expectant mothers ought to limit themselves to foods lacking a particular component –gluten for example– or lacking ingredients produced with the most advanced genetic technologies. To make their points, despite a lack of scientific evidence, the authors and marketers of such articles typically succumb to a range of biases –prejudice slanting of information that a process known as peer-review (review of a manuscript by people in the authors’ field but who usually don’t know the identity of the author) is designed to screen out from papers that are published in scientific journals, but that tends to characterize writings that are not peer-reviewed. Here we will take a look at some of these categories of bias, and when appropriate give examples of that bias being applied to a topic of health and pregnancy, or in some cases generally to health.
Selection bias: This is something that often happens prior to the writing of a paper, during an experiment itself, or in the analysis of the results of a study that is conducted with information that is already available (for instance, women who have taken a certain medication during past pregnancies). Suppose, for example, that you are trying to find out whether a particular drug, or food ingredient causes spontaneous abortions (miscarriages). Suppose that you do this by giving some agent, coffee for instance, to pregnant volunteers, but it turns out that most of the women in your experiment volunteered because they have experienced spontaneous abortions in the past. Although you are creating an experimental condition by giving some women regular coffee and some women decaf, the experiment is bias toward the conclusion that caffeine causes spontaneous abortion, because you have selected women who are likely to be prone to spontaneous abortion in the first place, whether they drink highly caffeinated coffee, or not.
Publication bias: This is the phenomenon in which certain topics, or results are more likely to be written up for publication, and then accepted for publication, compared with other topics. Examples of topics and results more likely to be published include findings that a new medication can have a great benefit, or a worse toxicity compared with the standard treatment for a given condition. In contrast, a study that finds that a novel treatment produces effects comparable to the standard treatment, or doesn’t work quite as well as the standard treatment often is seen as boring, so it is less likely to be published. Furthermore, if the later type of finding does get published, it is not likely to receive attention in the media. You can get an idea of how this works, if you image a headline reading something like this: “New medication works almost as well as aspirin in alleviating tension headaches and produces similar side effects.”
Negativity bias: The idea here is that people we tend to notice when things go wrong much more than when things to right. We spend much more time trying to understand things that go wrong than things that go right. Diphtheria, whooping cough, polio –these and several other infectious diseases used to kill and disable so many children and young adults in the early 1900s that it was typical for families to have many children based on the expectation that some were going to die. Vaccination has completely changed this reality. It prevents these diseases, and as a result millions of people are alive and well who otherwise would not be. Autism, on the other hand, is common in our society. It may not be anymore common today than it was over a century ago, but with most children being healthy autism sticks out like a sore thumb, and so it is the subject of numerous articles and television documentaries. This is in sharp contrast to the lack of death and suffering from polio, measles, and diphtheria, because, after all, who is going to write about how people are not getting sick. Today, there is a negativity bias toward autism, and the tendency is to blame just about everything, from vaccines to the chem trails of aircraft to so-called “genetically modified foods” (GMOs –a nonsensical term as all things alive, including all crops that you eat, are genetically modified) —–even though there is not a shred of scientific evidence for any such connection; indeed, science shows that crops produced with modern genetic techniques are particularly safe. But when it comes to blaming vaccines for autism, other biases also come into play. Related to the vaccine issues is another type of bias of distorted risk perception. Today, few people have seen diseases such as diphtheria, measles, and polio and their consequences, and as a result have trouble contemplating the risk. But many people have seen autism, and consequently perceive it as a bigger threat than any disease that is preventable through vaccination.
Confirmation bias: Once a person’s mind is essentially made up, this is where we see confirmation bias, which means that we tend to notice information, articles, and other factors that support our preconceived idea, and we tend to ignore all information that supports the contrary. Anti-vaccination activists commonly blame vaccines for autism, particularly a vaccine called MMR, which is used to prevent measles, mumps, and rubella. It must be emphasized that there is not a shred of evidence for any connection between the MMR vaccine and autism, but back in the 1990s a very prestigious medical journal, The Lancet, published a study by researchers that seemed to suggest a connection. The journal later retracted the paper, because it turned out that the study results had been fudged –invented because the leader of the study has an arrangement with a pharmaceutical company to develop individual vaccines for measles, mumps, and rubella that could have replaced the MMR vaccine. Subsequent analysis has showed that the MMR vaccine absolutely does not cause autism, but more than two decades later, anti-vaccine activists are still citing that one de-published study, and ignoring all studies showing the opposite.
Omission bias: Once somebody has decided that he fears something –let’s continue with vaccination as the example– it becomes all the more difficult to get the person to accept the benefits of the feared thing. As a result, the person will tend not to accept a treatment, or preventive measure (such as a vaccine), unless he perceives the consequences of not getting the treatment to be much worse than the consequences of getting it. Given the other biases that we have mentioned here, including the fact that people today have little experience with the diseases that the vaccines are designed to prevent, people who fear vaccines often view the risk-benefit as kind of a 50/50 thing. Doing so, they reason that sickness resulting from doing nothing (omission of a vaccine or treatment) is not as bad as sickness resulting from taking action. In reality, the risk benefit of vaccines is nothing like 50/50. The benefit is actually enormous and the risk very minimal, but that’s not how it looks when one’s perception is distorted by multiple layers of bias.