Patient Risk

Unnecessary testing may lead physicians to pursue and treat borderline and false-positive laboratory abnormalities. This observation does not imply that all standard screening tests should be discontinued. Some are beneficial, such as the mammogram for all women over 40 or 50 years of age,^{[108] [109] [135]} the test for occult blood or colonoscopy in stool for all people over 40 years of age,^{[126] [127] [129]} and the Papanicolaou (Pap) smear for sexually active females.^{[126] [127] [129]} However, few studies have examined whether increased testing and the follow-up on false-positive test results adversely affect patients. In one study addressing this issue, Roizen and associates^[82] retrospectively examined the adverse effects of chest radiographs on patients. For 606 patients, 386 additional chest radiographs were ordered without indication of need. Among those 386 patients, the discovery of only one abnormality (an elevated hemidiaphragm probably caused by phrenic nerve palsy) may have resulted in improved care for that patient. On the other hand, the existence of lung shadows on chest radiographs of three patients led to three sets of invasive tests, including one thoracotomy, but no discovery of disease. These procedures caused considerable morbidity, including one pneumothorax and 4 months of disability, for these three patients.

Tape and Muslin^[149] found a similar result when examining the benefits and risks of chest radiographs obtained preoperatively in Rochester, New York. Of 341 patients admitted for vascular surgery, nine had radiographic findings that led to clinical action. Specifically, three patients (two with congestive heart failure and one with pulmonary fibrosis) may have benefited from the findings. However, all three patients were known by history to have the disease shown on chest radiographs. In addition, six patients were subjected to a potentially detrimental clinical response. Two had a false diagnosis of tuberculosis, with subsequent therapy for one patient; two others had false diagnosis of nodules; and the last two had falsely normal chest radiographic readings. All the beneficial effects attributed to preoperative chest radiographs accrued to patients who had an obvious clinical history of pulmonary or cardiac disease. Orkin^[150] has further explained the basis of the risk of testing asymptomatic patients.

Similarly, in another study, even though few patients benefited by preoperative testing that was warranted by history or risk factors (91 of 1,746), even fewer were harmed by such testing (one patient) (Apfelbaum JL et al, unpublished data). In contrast, testing of asymptomatic patients was more risky than beneficial to patient health. Specifically, 1 in every 2,000 preoperative tests (1 in 300 patients) led to patient harm because of the pursuit of abnormalities indicated by those tests; 1 in 10,000 tests (1 in 1,746 patients) led to benefit (Apfelbaum JL et al, unpublished data).

In another study, Turnbull and Buck^[113] examined the charts of 2,570 patients undergoing cholecystectomy to determine the value of preoperative tests. With four possible exceptions, history and physical examinations successfully indicated the need for all tests that ultimately benefited the patients. For those four patients, it is doubtful that any benefit actually occurred as a result of preoperative tests. Among them was one patient who had emphysema detected only by chest radiograph; this patient had preoperative physiotherapy without subsequent postoperative complications. Two patients had unsuspected hypokalemia (potassium levels of 3.2 and 3.4 mEq/L in blood) and received treatment prior to operation. Current data in the literature indicate that no harm occurs to patients undergoing surgery with this degree of hypokalemia and that severe harm may be caused by treating such patients with oral or intravenous administration of potassium ( Table 25-7 ).^{[151] [152] [153] [154] [155]} The fourth patient possibly benefiting from preoperative testing had an asymptomatic hemoglobin concentration of 9.9 g/dL and was given a blood transfusion prior to cholecystectomy. Because cholecystectomy is not normally associated with major blood loss, one might conclude that this patient also received no benefit from preoperative laboratory testing and its pursuit but was exposed to the risk of transfusion. Thus, it is not clear that any patient in this study benefited from preoperative screening tests given without indication for need by history or physical examination.

In another study, only two patients (who had eradication of asymptomatic bacteriuria) benefited from the 9,720 screening tests that were given.^[156] At least one patient was seriously harmed from pursuit and treatment of abnormalities on screening tests. In this patient, atrial fibrillation and congestive heart failure developed after institution of thyroid therapy (for borderline low thyroxine levels) and free thyroxine index tests. It is unclear whether these investigators examined other patients for potential harm arising from the pursuit and treatment of abnormalities on screening tests.

To determine benefits and risks, screening mammography has been evaluated in a real-life practice setting.^[135] Although yearly screening was beneficial, over 20% of the women who did not have disease were subjected to a breast biopsy. Furthermore, more than 49% of these "normal" women would have been subjected to a breast biopsy if each had had a yearly mammogram and clinical breast exam. Thus, even when benefits exceed risk, there is substantial risk in routine testing.