In medical malpractice, inductive reasoning regards the standard of care as the duty to do no harm. If there is a complication from a medical intervention and the medical intervention differs from the standard of care in any conceivable way, the difference alone is sufficient to conclude that the medical intervention departs from that duty.
However, unanticipated threats require nimbleness. This is a calculated risk. It causes a difference, but it does not cause harm. Nonetheless, a calculated risk can be misrepresented as the cause.
No medical malpractice lawsuit illustrates this phenomenon better than Byrom v. Johns Hopkins’ Bayview Hospital.
Sixteen-year-old Erica Byrom comes to the U.S. from Liberia in August 2014, when 18 weeks pregnant. At 23 weeks, she begins prenatal care. Two weeks later, she is admitted to Bayview Hospital for pre-eclampsia.
Sonograms are consistent with chronic antenatal conditions associated with neurologic injury. There is at least a 65% chance that the fetus has already sustained neurologic injury. She steadfastly refuses a Cesarean section unless her own life is in immediate jeopardy.
On October 24, after 22 hours of induced labor, there is a normal spontaneous vaginal delivery of a 670-gram, 26-week, female infant with a 1-minute Apgar score of 0. After resuscitation, the newborn is admitted to the NICU and is later found to have cerebral palsy.
A trial begins in June 2019. Using inductive reasoning, the plaintiff’s attorney argues that “more likely than not, at 26 weeks the fetus is normal prior to birth. If not for the failure to perform a Cesarean section, Zubida Byrom would still be normal.” It is as though the antenatal history about Liberia and subsequent sonographic findings are not factors. Also using inductive reasoning, the defense attorney argues that “more likely than not, at 26 weeks the fetus has neurologic injury prior to birth. Even if there is a Cesarean section, Zubida Byrom would still have neurologic injury.” The antenatal history and subsequent sonographic findings are factors.
Two plaintiff medical experts testify that, because of the duty to do no harm, a Cesarean section should have been performed. A vaginal delivery is a medical error.
By convention, preponderance of evidence is 50% confidence plus a vague value, and that “vague value” only needs to be a scintilla to tip the scale. In inductive reasoning, 51% confidence suffices to prove a departure from the duty to do no harm. The defense attorney uses the same metrics to cast doubt on this proof. On July 1, 2019, the jury returns a $229-million plaintiff verdict.
This would have been different with deductive reasoning. First of all, medical interventions are harmonious processes involving 10 duties. These duties interact as a collective unit, and each duty represents a particular phase. The duty to do no harm is the collective result of these 10 phases. The ten phases are: 1. Presentation, 2. Investigation, 3. Interpretation, 4. Diagnosis, 5. Discrimination, 6. Informed Consent, 7. Selection, 8. Technical, 9. Resolution, 10. Discharge.
The accepted background risk for cerebral palsy in a preterm low birth-weight newborn is 15.2%. Cerebral palsy can still occur even with the standard of care because the 15.2% is inescapable. Nevertheless, because the standard of care is error-free, the risk of cerebral palsy from each phase in the standard is the background risk. The medical intervention is designed to be a facsimile of the standard of care.
When the 10 phases of the standard of care are compared to corresponding phases in the medical intervention, if a phase is the same, that phase in the medical intervention is represented by the background risk, or 15.2%. If a phase is different, that phase in the medical intervention is represented by the “incident risk” for the complication. In Byrom v. Johns Hopkins, the incident risk is 28.3%.
The aforementioned 65% chance of neurologic injury serves as the basis for this “incident risk.” If there is a 65% chance that a Cesarean section would not alter the outcome, it follows that there is a 35% chance that it will. Hence, the risk/benefit ratio for performing a Cesarean section is 1.86. In this case, not performing a Cesarean section increases the background risk of cerebral palsy in a preterm low birth-weight newborn from 15.2% to 28.3%.
The null hypothesis is “if there is no departure from the standard of care, there is no statistically significant difference between the incident risk resulting from the medical intervention and the background risk resulting from the standard of care.”
The null hypothesis is tested using the single-sample t-test. The null hypothesis is either retained or rejected. The level of significance is 0.05, corresponding to 95% confidence. The population mean is the background risk, which is 15.2%.
When the phases in the standard of care are compared to their counterparts in the medical intervention, there is a “test sample,” which collectively represents the entire medical intervention for Erica Byrom. It includes the 10 results from the comparison. The test sample is 15.2%, 15.2%, 15.2%, 15.2%, 15.2%, 15.2%, 15.2%, 28.3%, 15.2%, and 15.2%. Except for the technical phase, in which a Cesarean section is performed in the standard of care and is not performed in the medical intervention, there are no other differences between counterparts.
The result is the p-value. The p-value is 0.171718. Because the p-value is greater than the level of significance, which is 0.05, the null hypothesis is retained. The medical intervention comports with the standard of care with 95% confidence.
Hypothesis testing casts sufficient doubt on the plaintiff attorney’s inductive reasoning, which is no less than 51% confidence.
When using hypothesis testing, no medical expert with any integrity would prepare a certificate of merit favoring the plaintiff attorney. Instead of a verdict, Byrom v. Johns Hopkins would never go to trial.
On February 2, 2021, the Maryland Court of Special Appeals overturns this verdict, but it takes 16 months.
Howard Smith is an obstetrics-gynecology physician.
