Defensive medicine will be difficult to eliminate because it is the result of nearly 40 years of continuous legal assault on the medical profession. A physician practicing in a high-risk specialty today can expect to spend 25 percent of his or her entire career defending malpractice claims, despite the fact that four out of five claims are found to be without merit. Doctors find the medical legal process so destructive, it is understandable they feel justified in going to virtually any length to avoid it—including ordering tests that their medical judgment tells them are unnecessary but which our legal system says are required. Richard E. Anderson, MD, FACP, chairman and chief executive officer of The Doctors Company, provides his perspective on this issue in the article below—originally published in the Archives of Internal Medicine, Volume 159, November 8, 1999, Pages 2399–2402—which remains relevant today.
Physicians practice defensive medicine all the time. Some of us know it, acknowledge it, and accept it. Others believe this to be quite untrue and find the concept “morally unconscionable”1 and “reprehensible.”2 Precisely because the concept is elusive and the practice is pervasive, it is incumbent upon us to define the process and understand its effects. Not until then will we be able to do something about it.
First, I will explore the meaning of defensive medicine and why it has become so prevalent. Then I will assess the costs of defensive medicine and examine how the practice can be widespread, yet often go unrecognized.
Defensive medicine can be simply defined as medical responses undertaken to avoid liability rather than to benefit the patient. This violates the Hippocratic Oath, does violence to the doctor-patient relationship, and is manifestly self-serving for the physician. However, this definition is not particularly close to the way doctors approach most clinical problems. Virtually all medical tests and procedures offer at least some possible benefit to the patient. When do they become defensive?
The Office of Technology Assessment (OTA) approaches defensive medicine in a manner that is somewhat closer to clinical reality: “when doctors order tests, procedures, or visits, or avoid high-risk patients or procedures primarily (but not necessarily solely) to reduce their exposure to malpractice liability.” The basis of the choice need not be conscious and the practice is not necessarily without benefit to the patient.
Though this definition makes clear that defensive medicine is undesirable, the costs remain difficult to measure. There is no “gold standard” against which defensive medicine can be compared. Who defines “ideal medical practice”—physicians, patients, the courts, or society at large?
The OTA acknowledged “there is simply no acceptable method for measuring the full extent and national use of defensive medicine.”3 Faced with the problem, they surveyed physicians about hypothetical clinical scenarios designed to raise issues of potential malpractice liability. The OTA concluded that no more than 8% of diagnostic testing is “consciously defensive.”3 This misses the larger point, however. It is impossible to estimate how many physicians perceived the defensive pathway to be medically correct, but failed to acknowledge the degree to which these “medical” standards had been influenced by fear of liability. This 8% figure, moreover, represents tens of billions of dollars of health care spending.
Kessler and McClellan4 used a different method for measuring defensive medicine. After controlling for outcomes, they compared the cost to Medicare for patients hospitalized with either acute myocardial infarction or ischemic heart disease in states with and without effective tort reform. Effective reforms were defined as those that directly reduce malpractice awards, especially limits on either noneconomic (pain and suffering) or total awards, and the collateral source rule, which prevents duplicate recovery of costs already compensated.
In states with effective tort reform, Medicare costs were 5.3% less for a new diagnosis of acute myocardial infarction and 9% less for ischemic heart disease.4 If applied nationally, effective liability reform would save $600 million per year from the $8 billion Medicare spends annually for cardiac disease. If these savings were extrapolated across the health care system, they would amount to $50 billion per year, without any compromise in outcome. Thus, the authors have identified an annual cost of 5% to 9% of the health care budget for defensive medicine in the United States today. This figure is strikingly close to the 8% calculation of the OTA described above.
This must, however, be considered a minimum estimate. Even the best current tort reforms do not eliminate defensive medicine. California, whose Medical Injury Compensation Reform Act of 1975 (MICRA) is considered a model for national malpractice tort reform, still has a 44% higher claim frequency than the other 49 states, although costs per action are considerably reduced (The Doctors Company data on file). There are more than four claims for every ten California neurosurgeons, obstetricians, and orthopedists every year. It is difficult to imagine that this environment is free of defensive medicine.
The threat of litigation is ubiquitous in the practice of medicine today. Nationally, there are more than 17 claims for every 100 full-time practicing physicians every year; in high-risk specialties, the number is considerably higher (The Doctors Company data on file).
The National Practitioner Data Bank immortalizes every paid claim against a physician and this history is repeated endlessly in each application for hospital privileges and managed care contract. In some states, malpractice records are now available on the Internet.
It is critical that we acknowledge the strength of this aversive signal. Throughout our training, we are taught the noble traditions of medicine, selflessness in the care of patients, and discipline in the pursuit of beneficent medical knowledge. Yet malpractice suits are cast in terms of negligence, violation of duty, and incompetence. Although patient compensation is usually the expressed goal of litigation, physician punishment is also prominent in many cases (“I don’t care about the money, I just don’t want him to do that to anyone else”). Physicians take malpractice suits very personally. Rather than being seen as a “fact of life” or a “cost of doing business,” malpractice suits often threaten the core of a physician’s self-identity.
Despite this, empirical studies that have compared changes in clinical practice between sued and nonsued physicians have seemed to produce inconsistent results.5,6,7,8 For example, Charles et al. found that similar percentages of sued and nonsued physicians (68% v. 60%) ordered more tests than clinical judgment would indicate. However, sued physicians were more likely to report symptoms of personal stress, to limit exposure to high-risk patients, to consider early retirement, and to discourage their children from becoming doctors.6 Baldwin et al. found that personal malpractice history did not influence clinical decisions in low-risk obstetrics,8 but Localio et al. found a significant increase in the use of cesarean section by physicians who had been sued previously.7
These differing conclusions can be explained by the fact that physicians are so averse to malpractice suits nearly all clinical judgments are influenced. The question is not whether, but by how much they are affected. In lower risk medicine, the gap between best medical practice and defensive medicine is naturally diminished.
Although claims frequencies are already high by almost any measure, there are those who feel these suits are appropriate9 and that more, rather than less, litigation is needed.10 Best known are the authors of the Harvard Medical Practice Study. After reviewing the records of 31,429 hospitalizations in New York State in 1984, they concluded “the chances that a claim would be filed by a patient with an identifiable negligent injury is only one in fifty.” They perceived a “tort gap” and asserted that “the underlying assumption that too many groundless malpractice suits are initiated is unfounded.”10
If this were so, it would be difficult to argue there is too much defensive medicine, but is it true? First, their definition of injury required only twenty-four hours of disability. It is hard to understand the authors’ implication that such transient events should be the subject of litigation. Minor and serious events were counted equally, whether or not a physician was directly involved. A slip and fall in a hospital corridor was weighted the same as a surgical error resulting in paraplegia.
Equally important, the methodology used in the Harvard study was unreliable. Once identified, adverse events were independently rated for negligence by two medical reviewers chosen without regard to their specialties. Although only a “more likely than not” (51%) likelihood of negligence was required, the two reviewers agreed with each other only 10% of the time.11 Furthermore, when a subset of 318 records was reevaluated, a second set of reviewers “did not identify the same group of adverse events as the first team, but they did find about the same incidence of adverse events and adverse events due to negligence.”12 The authors took this as evidence confirming the reliability of their methodology. This is akin to saying that as long as the incarceration rate matches the crime rate, it does not matter whether we convict the innocent or the guilty. In fact, this idea is remarkably similar to the physician’s view of medical liability proceedings as a lottery. If lawyers file enough suits, they will ultimately win some percentage of them.13
Perhaps most remarkable of all, the authors found no correlation between the outcome of the actual medical malpractice suits in the study and the presence or absence of medical negligence.14 Only the degree of injury was associated with outcome.
The Harvard study, therefore, tells us little about the reality of malpractice litigation except that jury verdicts are independent of medical negligence. Add to this the fact that 70%–80% of all claims close without payment to the plaintiff,15 and it is understandable that physicians feel most malpractice suits are unfounded and, therefore, actions taken to avoid entanglement in our system of medical-legal jurisprudence can be justified. These facts also clearly belie the shibboleth that the “primary malpractice problem continues to be malpractice.”9
“Defensive medicine” implies that a doctor deviates from the standard of care primarily to protect against liability, but what is the standard of care? We think of it as a reasonable medical response appropriate to the clinical problem and context. This definition is similar to the formal legal notion of the standard of care being what other similarly qualified physicians would do under similar circumstances. In truth, the standard of care consists of two components: “medically indicated” and “legally required,” reflecting the standards of the community.16 In other words, the standard of care becomes the corpus of medically reasonable tests and procedures which produce outcomes acceptable to the individual patient and to society as a whole.
Malpractice suits play a major role in influencing the standard of care. Cesarean section rates in the United States rose from 4.5 per 100 births in 1965 to 24.1 per 100 in 1986.7 Electronic fetal monitoring is widely used even though its routine clinical utility has been severely questioned.17 Essentially all explanations of this phenomenon relate to the perceived threat of litigation, not the value of the intervention. The medical standards have changed to incorporate the perceived legal standards. This is not inherently good or bad, but physicians and society must acknowledge it explicitly. Doctors are trained to make medical decisions using clinical skills and judgment to solve medical problems. We think in terms of a medically correct or best answer, though we recognize limits on diagnostic certainty.18 Increasingly, however, the physician practices the standard of care defined by the courts and the court of public opinion.
Two recent publications—the first concerning Lyme disease, the second regarding breast cancer—illustrate why explicit acknowledgment of defensive medicine is important.
Fix et al. found serologic testing for Lyme disease was frequently ordered without clinical benefit and that empirical antibiotic therapy was overutilized.19 They concluded that “physicians managing patients with tick bites in communities with endemic disease are either unaware of recommendations against routine prophylactic antibiotic therapy, disagree with them, or find them infeasible in practice.”19 An accompanying editorial suggested the findings were due to patient expectations and to the limited time physicians allocate for patient education.20
Although these explanations are undoubtedly true in part, there was no acknowledgment of the effect of the threat of malpractice litigation on physician behavior. Untreated Lyme disease can be serious and is the patient’s prime concern in initiating the office visit. Should the disease actually develop, the patient’s diagnosis, however uninformed, would prove more accurate than the doctor’s. The immediate adverse consequences of a prescription are usually small. Physicians can be assured by the concept that they are preventing the worst outcome at relatively little risk and cost, and avoiding a potential lawsuit at the same time. This is almost certainly an unacknowledged example of defensive medicine changing clinical practice. In this case, the authors feel the standard of care as defined by the scientific evidence is not being met. However, the medical community has chosen a different standard.
Similarly, a widely publicized study recently reported the high rate of abnormal findings in women undergoing annual mammography.21 The cumulative estimate of a false positive result was nearly 50% in women receiving 10 mammograms. Neither the study authors nor the editorialist22 attempted to explain the findings, but both indicated they were of concern. The press was less restrained, with front page articles in major newspapers variously decrying “needless fear,”23 “unnecessary biopsy,”24 and “stressful, time-consuming and expensive further testing.”25
Although the vast majority of breast biopsies performed for mammographic abnormalities are benign,26,27,28 allegations involving the failure to diagnose breast cancer are the most common basis for malpractice suits today.15 Misreading mammograms is the most frequent malpractice allegation made against radiologists.29
Moreover, as with Lyme disease, the lesion is usually discovered by the patient and the positive diagnosis her greatest fear. We must not be surprised, therefore, when radiologists are cautious in interpreting mammograms and surgeons have a low threshold for biopsy.
The authors note that abnormal mammogram readings are nearly twice as common in the United States as elsewhere, yet sensitivity is similar.21 This suggests the fear of missing a diagnosis here is high, although the ability to detect cancer is similar. This situation has costly consequences. Mammography false positives increased the expense of screening by 33% and added greatly to the burden of patient anxiety incurred in the process.
We are left with the hypothesis that the standard of care in the community has incorporated a high degree of defensive medicine. If we are dissatisfied with this result, we will first need to acknowledge it, and then work to change the medical-legal climate which produced it.
Physicians may practice defensive medicine without guilt because the action is either unacknowledged or not considered unethical. We are rewarded for the unexpected diagnosis by impressed colleagues and grateful patients. Additional medical evidence, even if it merely confirms clinical judgment, cannot be said to be contrary to the patient’s welfare. It is easy for physicians who overutilize tests and procedures to regard themselves as thorough and complete in their approach to patient care. Rightly or wrongly, physicians may believe that medical practices that appear to protect against involvement with our erratic and unpredictable system of medical-legal jurisprudence are warranted.
Medical standards of care are increasingly determined by legal and community standards not based on scientific evidence. This approach adds significantly to the enormous expense of contemporary health care. Where this is unrecognized, medical outcomes may be unaccountably changed for the worse. Current examples include overprescribing for Lyme disease and the high percentage of false positives in breast cancer screening.
Nearly all medicine today is practiced in a liability sensitive context. The extent to which this phenomenon is appropriate requires serious evaluation. Before this can be accomplished, it must first be recognized.