Preventing Malpractice in Internal Medicine: A Claims Review Panel Report

This report discusses several complex internal medicine claims that occurred despite generally attentive medical care. In each case, extraneous issues complicated the clinical factors.

Causes for the claims studied included critical comments made by subsequent treating physicians, poor physician-to-physician communication, patient and family anger, and inadequate response to life-threatening emergencies.

Iatrogenic Suits

Three cases appeared to have been triggered by remarks from subsequent treating physicians. In the first case, the condition of a 46-year-old man with a significant history of alcohol abuse was followed by his physician for three years. On several instances, a mildly elevated serum calcium was detected. On one occasion, a serum parathormone level was slightly elevated, but the physician did not recognize the significance of the laboratory study, and recorded the diagnosis in the chart as benign hypercalcemia. The patient’s calcium was often recorded in the normal range, and when elevated, the condition typically responded to the withdrawal of thiazides from the patient’s antihypertensive program.

Ultimately, the patient developed hemorrhagic pancreatitis and experienced a difficult hospitalization, requiring four months of inpatient care. The plaintiff’s expert was prepared to testify that the hyperparathyroidism, not the patient’s alcoholism, was the primary etiology of the pancreatitis.

Although the patient did have primary hyperparathyroidism that had been misdiagnosed, the absolute increase in the serum calcium was mild, and on many occasions, the value was actually normal. Even if the physician had recognized the serum calcium increase, surgery probably would not have been indicated. Although hyperparathyroidism is associated with pancreatitis, in this case, the patient’s alcoholism appears to have been a more significant factor. After a prolonged hospitalization, the patient made a full recovery and continued therapy with his physician for more than one year. Almost surely, this claim was generated by the remark of a consulting physician who noted the weak association of pancreatitis with hyperparathyroidism, rather than the stronger link with alcoholism.

In another case, a 67-year-old woman underwent nephrectomy for a staghorn calculus and renal abscess. In the four months prior to surgery, the patient’s platelet count varied between 566,000 and 686,000. The complete blood count (CBC) drawn immediately preoperatively, however, indicated that the platelet count had risen to 826,000 and was accompanied by a white count of 23,000. The patient suffered a thrombotic cerebrovascular accident on the evening after her operation. The hematologist who saw the patient in consultation postoperatively indicated that the thrombocytosis had played a part in the patient’s stroke. He placed the patient on aspirin and Persantine. Subsequent bone marrow examination showed a questionable early myeloproliferative syndrome, although platelet function was normal. The patient’s elevated platelet count was likely nonspecific and secondary to inflammation.

Both the platelet and white counts gradually returned to normal following surgery. Although there is some controversy on this point, reactive thrombocytosis is rarely thrombogenic and almost never requires treatment with counts less than 1,000,000. Even with counts above 1,000,000, it is hard to find convincing data of thrombosis risk. Indications for surgery were compelling, and preoperative antiplatelet treatment would have been questionable. Again, this claim probably would not have gone forward except for the consulting hematologist’s attitude. This sort of posture is especially frustrating when the expressed opinion is weakly founded.

A third case involved a 72-year-old woman who saw her internist regularly for a variety of medical problems. A thorough examination resulted in well-documented diagnoses of hypothyroidism and B₁₂ deficiency, which were treated. In May, the patient’s CBC was normal. The following November, the patient was found to have acute myelogenous leukemia and was transferred to a university hospital. The patient filed suit for delayed diagnosis of leukemia.

Review of the medical record showed no reason to suspect leukemia before the actual diagnosis. The CBC in May was normal, and the patient’s symptoms were not suggestive of severe underlying pathology. This patient had seen her physician regularly for two years, prompting the suspicion that this suit was precipitated by remarks made at the university hospital that suggested an earlier diagnosis would have been possible. Again, this case is particularly frustrating because the record fails to substantiate any breach in the standard of care.

While inappropriate communication between physician and patient may lead to litigation, the next group of cases shows that poor communication between doctors also constitutes a serious problem.

Doctor-to-Doctor Communication

In February, a 53-year-old patient was seen by his primary internist for indigestion. His condition had not improved with the usual antacid therapy. The patient was an obese heavy smoker who had a long and well-documented history of hiatus hernia and esophageal achalasia. Hospitalization was ordered, and an acute inferior myocardial infarction (MI) was documented. He was then transferred to a tertiary care center, where angiography demonstrated a totally obstructed right coronary artery and significant stenosis of the proximal left anterior descending. Suffering ongoing chest and abdominal pain, the patient was hospitalized for more than two weeks at the referral institution. Gastrointestinal (GI) and psychiatry consultations were obtained in addition to the extensive cardiac evaluation.

Ultimately, the patient’s pain was believed to be primarily cardiac in origin. Despite this belief, medical therapy alone was recommended. The patient was referred back to his primary physician, who again hospitalized him after he presented with additional chest pain. The admitting note indicated that the physician had the impression that the patient did not have surgically correctable coronary artery disease. After 24 hours of observation, the patient was discharged. He died four days later of an acute MI.

The letter from the referring cardiologist indicated that the decision not to perform immediate coronary artery surgery would need reconsideration if there was recurrent chest pain. Despite this caveat in the written report, a different impression was apparently communicated during a telephone conference regarding this patient. The conversation led the primary physician to conclude that the consultants believed the patient could not be revascularized.

Review of the case showed ample blame to go around. First, the stenosis of the left anterior descending artery appeared to be more significant than was initially recognized. Second, the patient had ongoing pain during the two-week hospitalization—which should have been a signal for more aggressive treatment at that time. Third, the primary physician could have recognized the potential significance of the new chest pain episode and could at least have contacted the consulting cardiologist for further advice rather than simply discharging the patient after 24 hours of hospitalization. Moreover, the case is disturbing because the written records clearly indicate that the primary physician and consulting cardiologist did not appreciate the significance of the symptoms, nor did they appropriately evaluate the possibility of surgical intervention.

In another case, a 67-year-old woman was hospitalized with pulmonary edema and sepsis. The primary physician suspected that the gallbladder was the source of the sepsis. The subsequent extensive evaluation included endoscopic retrograde cholangiopancreatography and transhepatic cholangiogram, both of which were technically unsuccessful.

A radionucleotide biliary imaging study was normal. The infectious disease consultant recommended appropriate parenteral antibiotics, and the patient completed a six-week course of therapy. She then underwent cardiac catheterization and was found to have significant aortic insufficiency and mitral regurgitation. A prosthetic aortic valve was placed. Postoperatively, the patient became septic again and had a prolonged and difficult two-month hospitalization. One month after cardiac surgery, the patient had a right upper quadrant tenderness that required an emergency operation for a gangrenous gallbladder. The allegation was that cardiac surgery should have been postponed until the issue of biliary sepsis had been fully resolved.

Although there is room to debate the medical issues in this case, the initial consulting gastroenterologist left a well-documented record indicating his opinion that cardiac surgery should be postponed until the gallbladder could be thoroughly evaluated as a source of sepsis. It is not clear that the cardiac surgeons were cognizant of this recommendation. The second episode of sepsis was catastrophic.

Confounding Diagnoses

In working through the maze of differential diagnoses, Occam’s razor has always been an important clinical rule. Simply, it says not to make two diagnoses when one would suffice. That can be a serious error, however, when two conditions share similar symptoms but radically different treatments and outcomes. Clinical decision-making is particularly difficult when the patient has two well-established diagnoses.

The first case in this category involves a 53-year-old man with a well-documented three-year history of chronic back pain that left him medically disabled. The patient was seen by a gastroenterologist one day after referral by his family physician. During the visit, he complained of excruciating back pain and was too uncomfortable to cooperate with an abdominal examination. Instead, he was given parenteral analgesia and referred to a neurologist. He died at home that night. Autopsy showed a perforated gastric ulcer with peritonitis as well as an acute MI.

The records indicate that this patient was particularly difficult to treat. He had a long history of chronic back pain and drug abuse. It is difficult, however, to defend the failure of a gastroenterologist to examine the abdomen. The context should have provided valuable clues, and better doctor-to-doctor communication would also have been beneficial. The record does not indicate that the family physician’s examination was conveyed to the gastroenterologist, who, at the very least, should have excluded acute gastrointestinal pathology before referring the patient to yet another physician.

Establishing the etiology of chest pain is frequently difficult—as illustrated by the case of the man with indigestion described in the “Doctor-to-Doctor Communication” section above. The patient had recently experienced an MI and had severe esophagitis. Despite well-documented coronary artery disease, his symptoms were consistently interpreted as being referable to the GI tract. The patient died of an MI four days following hospital discharge.

These cases demonstrate the importance of paying careful attention to the nuances of a patient’s history, with particular attention to any change in established patterns. Each episode must be evaluated independently, and the most serious potential diagnosis must be considered. This becomes increasingly hard to accomplish in an era of cost containment, but that defense is extremely unlikely to convince a jury.

Difficult Diagnoses

A 60-year-old man was hospitalized with fever, nausea, vomiting, and low back pain radiating to the right leg. An extensive evaluation—including MRIs of the spine, a CAT scan of the abdomen, and a gallium scan—was performed. After seven days of hospitalization, the patient had an abrupt change in mental status. The admitting orthopedist requested neurology and infectious disease consultations. The possibility of a cerebral abscess was raised, but a CAT scan of the brain was negative. A lumbar puncture was requested but was not successfully accomplished for five days. When finally done, frank pus was found. A myelogram demonstrated an extensively infected epidural space extending from the foramen magnum to L5. A large decompressive laminectomy was performed and intensive antibiotic therapy was provided, but the patient became and remained quadriplegic.

In studying this case, a panelist for The Doctors Company cited a good clinical rule: “A lumbar puncture should be performed at any time the question arises as to whether a lumbar puncture should be performed.” Despite the technical difficulties presented by this patient’s spinal stenosis, an earlier lumbar puncture would probably have improved this patient’s prognosis.

The second case in this category involved a 14-year-old girl who presented to the emergency room just after midnight with chills, fever, and emesis times two. Her temperature was 102 degrees, and her initial blood pressure was 74/40. The blood pressure did not improve after more than 4,000 cc of IV fluid. Only then was a CBC drawn and a decision made to admit the patient. Zinacef and erythromycin were begun, but by that time, the patient had developed generalized ecchymoses. She died approximately two hours later.

The diagnosis of meningococcemia can be extremely difficult. Meningococcal infections can vary from mild upper respiratory syndromes to the Waterhouse-Friderichsen syndrome. This occurs in 10 percent of patients who have meningococcemia and presents as rapidly progressive vasomotor collapse. The classical rash is macular, erythematous, nonpurpuric, and blanching. It ranges from 2 mm to 15 mm and typically presents on the trunk and extremities. Petechiae and purpura follow. High doses of penicillin G or ampicillin remain the treatment of choice and must be instituted as soon as symptoms emerge.

Cases Involving Telephone Communication

The panel’s first case in this category involved a 27-year-old woman in her eighth month of pregnancy who spoke with her obstetrical group one Saturday, and again the next day, complaining of a typical flu-like syndrome. The patient was treated conservatively and told that if her symptoms worsened, she should go to the emergency room or be seen at her doctor’s office on Monday. The patient was not heard from again until she presented to the emergency room at 3:00 a.m. on Tuesday. At that time, her white count was 700 and her blood pressure was 88/40. She was found to have an overwhelming Streptococcus pneumoniae, from which she expired.

The care in this case appears to have been appropriate. Evidence suggested nothing in the initial phone conversations that would have mandated direct physician evaluation. Pneumococcal sepsis, however, remains a potentially lethal disease, even in the antibiotic era.

It is impossible to avoid clinical decision-making by telephone, but extra caution is always necessary. Explicit instructions for follow-up should be given and documented.

The second such case in this category involved a 42-year-old dentist who called his physician friend complaining of a severe stomachache. A remarkable series of phone calls ensued in which the patient was advised to obtain Prilosec tablets from the hospital pharmacy and then to remain in the emergency room waiting area to see if the tablets effectively removed the pain. If they did not, he was to be evaluated. Unfortunately, the patient left the hospital after obtaining the medication and expired behind the wheel of his car as he drove off the hospital grounds. The coroner found the cause of death to be an acute MI.

This appears to be a case in which the doctor-patient friendship, and perhaps the patient’s limited medical knowledge, led to an unusual course of action. Although the physician recognized the severity of the potential diagnoses and gave the patient a specific set of instructions, the patient died without having been seen by a physician.

Common Problems in Internal Medicine

As with the case described above, patients sometimes make requests that are best not granted. In another example, a 67-year-old man presented to his physician for evaluation of what appeared to be a sinus headache. He had a long history of hypertension but requested Sudafed as well as antibiotics. The physician instructed the patient to take one Sudafed and to recheck his blood pressure using his cuff at home. The physician told him that if his blood pressure had not changed, he was to take the second tablet. The patient died of a subarachnoid hemorrhage soon after taking the second pill.

Despite a clear warning on the package that Sudafed should be used cautiously by patients with hypertension, the literature actually suggested that small doses do not typically elevate blood pressure. Because he cautiously yet unnecessarily defied conventional wisdom, the prescribing physician would probably have to defend his actions in front of a jury. This case also emphasizes the difficulty of evaluating patients who have a headache. Not every patient can or should be screened with a CAT scan or MRI, but a careful history and physical examination are necessary for each patient presenting with a headache. The physician must have an accurate understanding of the nature and severity of the headache, particularly compared with prior episodes.

The final case involved a 76-year-old patient who died of complications of mediastinitis following an esophageal perforation during endoscopy. In this case, the physician met all of the caveats:

• The procedure was indicated.
• The patient signed a detailed informed consent.
• The complication was promptly recognized and treated.

Under these circumstances, suits such as this one, though unfortunate, are almost always won.

Summary

Even complex cases involving internists frequently revolve around straightforward issues of communication.

J3244 2/05