The Imaginary Light

I'm writing to you (Mr. tumblr) as recent personal events have provoked some deep thoughts in me concerning the application and interpretation of statistical concepts in different fields, particularly in medicine and criminal justice.

Recently, my newborn son had a chaotic health incident in the hospital, and as a result, I've found myself involved in a complex hospital fiasco. Throughout this process, I've observed how statistical concepts, especially correlation and causation, have been handled in the medical field.

What struck me was the apparent dismissal of correlation as an invalid concept in the medical field, as per the hospital's investigation. This got me thinking about how these concepts are applied in different domains. In criminal justice, for instance, the correlation between two variables does not imply causation, a principle that is fundamental in avoiding misguided interventions.

Contrarily, in medicine, the stakes are too high to rely solely on correlation; causation needs to be established. Thus, a correlation may be disregarded if causation cannot be proved. It has made me consider how the nuances in each field's approach to these concepts can lead to differing interpretations and outcomes.

In criminal justice, proving causation often serves as the backbone of establishing a defendant's guilt. In cases of medical malpractice or negligence, establishing a causal link between the medical professional's action (or inaction) and the harm or damage that occurred to a patient is necessary to secure a conviction. The "correlation not causation" principle can play a vital role in these cases because there can be multiple factors at play that could contribute to a patient's harm. Let's delve into this scenario: We observed a correlation between a diagnosis of meningitis in a baby and a subsequent cessation of breathing. However, this correlation doesn't unequivocally indicate that the meningitis infection directly led to the baby stopping to breathe. There could be various other variables at play such as the baby's overall health status, their age, any pre-existing conditions, or potentially other factors.

One such additional factor could be the intervention that took place during the breastfeeding counselling. During this session, there was an attempt to position the baby's head on the mother's breast, an act which I (we) believe could have unintentionally led to suffocation. Moreover, the first confirmed and observed seizure didn't occur until after this incident, while the baby was in the Pediatric Intensive Care Unit (PICU). This fact further complicates the matter and suggests that the chain of events may not be as straightforward as initially perceived.

Both of these factors - the seizure possibly caused by the meningitis and the accidental suffocation during breastfeeding - could independently, or in combination, lead to conditions like HIE (Hypoxic Ischemic Encephalopathy). Hypoxia or HIE could indeed result from either of these factors alone or from their combined effects. However, a meticulous and comprehensive investigation is required to determine the precise cause and contribution of each factor.

The intricacies of these interactions highlight the importance of not prematurely and hastily attributing causality based on a single observed correlation. This situation reinforces the crucial principle that correlation does not automatically signify causation.

However, I believe that statistical learning, whether applied to medicine or criminal justice, remains scientific and doesn't contradict. Instead, it encourages us to think critically about data and conclusions drawn from it.

The hospital investigation has, in many ways, stirred up a great deal of critical thinking for me. While the incident itself was unfortunate, it has certainly opened up new avenues of thought and understanding in the realm of statistical analysis and its different applications. 

In a courtroom of facts, as a statistician, I always embark on a quest for truth. Like lawyers, we examine factors contributing to a conclusion based on employing statistical and scientific evidences and reasonings to prove claims with the least error, ensuring the truth shines through the complexities of information.

I should have been braver and held her hand. I should have told her how much I loved her and that she was the best thing I could ever of had. I should have told her that if she needed to go, I understood and that she would always be missed. I should have told her not to worry about me and that I would be fine. But I hid behind my fear, believing she must already know these things, pretending I’d still have time to say the things that needed to be said—how fucking stupid was that.