As shown in my earlier post, when prosecuting Byrom vs. Johns Hopkins Bayview Hospital with inductive reasoning, as is traditional, the medical intervention is compared to the standard of care in a very general and subjective way. The medical intervention “more likely than not” departs from the standard of care. “More likely than not” corresponds to a level of confidence of around 51 percent and a type-1 error of around 49 percent. The defense attorney, also, uses “more likely than not.” One does not cast doubt on “more likely than not” by using “more likely than not.” The outcome is a coin toss resulting in a $229-million plaintiff verdict.
If, however, Byrom vs. Johns Hopkins Bayview Hospital is defended with deductive reasoning, decision-making is completely objective. There are ten duties from day one until discharge that collectively represent the duty to do no harm. The standard of care and the medical interventions are separated into ten phases, each corresponding to a specific duty. Phases in the standard of care are compared to corresponding phases in the medical intervention. The comparison is of specific duties, and the differences are quantitative.
The standard of care is empirical. It could mean different things to different practitioners depending on the circumstances. The medical intervention is not empirical; it is verifiable in medical records and cannot be disputed. The medical intervention is or should be a facsimile of the standard of care.
From this comparison, a test sample, which represents the medical intervention, emerges. In the Byrom case, only one phase differs from the standard of care: the technical phase. There is only a maternal indication for Cesarean section. In the technical phase of the standard of care, there are maternal and fetal indications for Cesarean section.
In this test sample of ten phases, nine have the background risk of 10 percent. The “background risk” is the risk of cerebral palsy in a population of preterm newborns. Only the technical phase has the incident risk of 12 percent. The “incident risk” is the risk of cerebral palsy by departing from the standard of care.
The null hypothesis is “the medical intervention comports with the standard of care.” When using a level of significance of 0.05 in the single sample t-test, the p-value is 0.171718. The null hypothesis is retained. Deductive reasoning has a level of confidence of 95 percent and a type-1 error of 5 percent.
Plaintiff attorneys and the medical experts they hire are clearly critical of the standard of care used by the defendants. They also assume that this 26-week fetus is normal before delivery. They never once state their reason for this assumption or their version of the standard of care. The only thing finders of fact know about their version is that it includes maternal and fetal indications for Cesarean section. So does the standard of care using deductive reasoning, but in only one phase—there are ten phases.
What if the plaintiff attorney uses the same deductive reasoning, i.e., hypothesis testing, to reject the null hypothesis? Rejecting the null hypothesis proves a departure from standards of care with 95 percent confidence.
The medical intervention is the same because it is what defendants document in medical records. Documented are a history of no prenatal care during the first trimester in Liberia and sonographic findings of oligohydramnios, absent end-diastolic umbilical artery blood flow, and intrauterine growth restriction at 26 weeks. These are reasons to suspect that this fetus is not normal.
The background risk, the incident risk, and the level of significance are all the same because they are part of the heuristics of the t-test. The only thing in question is the standard of care. After all, it is empirical.
The plaintiff attorney’s version of the standard of care must explain which phase or phases justify the assumption that this 26-week fetus is normal before delivery despite what is already documented in medical records. Furthermore, the standard of care must explain why a classical Cesarean section is the safest and most effective option and why the mother must agree. At long last, the plaintiff attorney and the medical expert, who is retained, will articulate their version of the standard of care for all to see.
When their standard of care is compared to the medical intervention in dispute, the test sample has four phases with the background risk of 10 percent and six phases with the incident risk of 12 percent. The p-value is 0.007478, and the null hypothesis is rejected.
However, the distortion of the standard of care necessary to reject the null hypothesis is completely transparent for finders of fact to see. In fact, this degree of distortion crosses the line of professional misconduct. Plaintiff attorneys are not so bold as to use deductive reasoning for this very reason; so, they use inductive reasoning, which obscures this misconduct.
In the final analysis, hypothesis testing proves that a medical intervention comports with the standard of care with 95 percent confidence. Hypothesis testing also makes it virtually impossible to make a medical intervention, which comports with the standard care, appear as if it departs from the standard of care.
On February 2, 2021, the Maryland Court of Special Appeals overturned Byrom vs. Bayview Hospital. “Because we conclude that the evidence presented at trial was not sufficient to support findings of either negligent treatment or breach of informed consent, we hold that the trial court erred … We reverse the judgments.”
This conclusion presumes that the appellate judges may have used deductive reasoning themselves. The best way to expose the shortcomings of plaintiff attorneys and the medical experts, who they hire, is with deductive reasoning. If deductive reasoning was traditional decision-making instead of inductive reasoning, Byrom vs. Johns Hopkins Bayview Hospital would likely never have been filed.
Howard Smith is an obstetrics-gynecology physician.
