Conventionally, medical malpractice is argued using inductive reasoning. There is nothing wrong with inductive reasoning. It is a component of critical thinking. However, it is just a component. Inductive reasoning is more subjective than objective and the influences of selection bias, avoidance bias, and confirmation bias are undetectable.
Deductive reasoning is also a component of critical thinking. It is not immune from these same biases; however, deductive reasoning is more robust and is more objective than subjective. It complements inductive reasoning to complete critical thinking and the influences of selection bias, avoidance bias, and confirmation bias are detectable.
The following generic medical malpractice case serves as an example. The plaintiff attorney and the plaintiff’s medical expert use inductive reasoning. There are 10 phases in the medical intervention and 10 counterpart phases in the standard of care. When making the comparison, only one phase in the medical intervention differs from its counterpart in the standard of care. There are no quantitative criteria when making this comparison. Inductive reasoning confirms that data in that phase depart from the standard of care and the departure is the proximate cause of the complication. Inductive reasoning only infers that the complication is a medical error and concludes that there is a preponderance of evidence that the medical intervention is medical malpractice.
Alternatively, the defense attorney and the defense’s medical expert use deductive reasoning. Deductive reasoning uses these 10 phases in the medical intervention and 10 counterpart phases in the standard of care and also finds one phase in the medical intervention that differs from its counterpart in the standard of care. Here is where inductive reasoning ends and deductive reasoning begins. To determine medical malpractice and to prove that the complication is a medical error, deductive reasoning uses hypothesis testing.
The background risk for this complication is researched to be 23 percent of the population at risk. To create a test sample, which represents the medical intervention in question, the defense’s medical expert uses the threshold risk ratio for a medical error, which is 100 percent / 23 percent = 4.35. This is the yardstick for the assignment of all relative risks. In the one phase of the medical intervention, which is different from its counterpart in the standard of care, it is determined that the difference in the performance of duty increases the risk of harm, but not enough to exceed the threshold risk ratio. The relative risk for that phase is determined to be 4 and the incident risk is 92 percent. The nine other phases are the same as their counterparts in the standard of care and relative risks are 1.
Hence, the 10 incident risks in the test sample are 23 percent, 23 percent, 23 percent, 23 percent, 23 percent, 23 percent, 23 percent, 23 percent, 23 percent, and 92 percent. Collectively, they represent the entire medical intervention. The population mean (µ) is 23 percent. Alpha is 0.05. The reason for this choice is that statistical significance is the sine qua non for science when doing hypothesis testing. The p-value = 0.171718.
The statistical test is the single sample T-test. The p-value is greater than alpha (0.05). The null hypothesis is retained. Type I error is 5 percent, which means that the probability that null hypothesis is a false positive is 5 percent. Type II error is 20 percent, which means that the probability the null hypothesis is a negative is 20 percent. Since the null hypothesis is retained, type I and type II errors support the notion that the null hypothesis is true and should be retained. Deductive reasoning concludes with 95 percent probability that the complication is an error-of-nature and that the medical intervention is not medical malpractice.
In deposition testimony, the plaintiff’s medical expert and the defense’s medical expert are both physicians and both are scientists. The plaintiff’s expert could have chosen statistical significance, the sine qua non in science, but, instead, elects preponderance of evidence, the sine qua non in law. The reason for doing so is relevant to a confirmation bias by the plaintiff’s medical expert.
Understanding that all reasoning can be influenced by bias, the defense attorney adapts inductive reasoning to deductive reasoning to examine for the effects of bias. This is completely acceptable. The medical intervention in the null hypothesis and the one in the standard of care are compatible and the p-value is 0.171718. Because the plaintiff’s expert uses preponderance of evidence, alpha is 0.5. The p-value is less than alpha. The null hypothesis is rejected.
However, deductive reasoning qualifies and quantifies type I and type II errors when inductive reasoning does not. An alpha of 0.5 is a type I error of 50 percent. This represents a selection bias which rejects the same null hypothesis. When type I error is 50 percent, type II error is 12.5 percent. This represents an avoidance bias of 87.5 percent, which rejects the same null hypothesis. A selection bias, and an avoidance bias are the effects of choosing an alpha of 0.5, the sine qua non in law, instead of an alpha of 0.05, the sine qua non in science. Choosing alpha of 0.5 is, also, a confirmation bias.
Complementing inductive reasoning with hypothesis testing exposes these biases and casts doubt on the conclusion of medical malpractice by the plaintiff’s medical expert and the plaintiff attorney, who rely exclusively on inductive reasoning.
“Whatever gets measured, gets managed.” Deductive reasoning measures type I error and type II errors and manages the merit, or lack thereof, in a medical malpractice lawsuit. It also measures and manages the biases of investigators. Jurors are not expected to be statistically literate; however, all medical experts and officers of the court are expected to be and should be. CCC+C may not change the game; however, it does level the playing field.
Howard Smith is an obstetrics-gynecology physician.
