Decision-making in surgical practice is a highly complex interaction between the client (patient) and his or her professional advisor (surgeon). The client approaches the professional advisor full of expectations and fears. The expectations may be realistic or unrealistic, and, in the same way, the fears may be justified or unjustified. Furthermore, the client may have a variety of priorities not immediately apparent to the surgeon and may wish to express different degrees of autonomy. For example, some patients may demand the final say in determining the balance between length and quality of survival, whereas others would be happy to allow complete abrogation of their responsibilities to the surgeon, who is invited to make the difficult decisions concerning the patient's utilities. From the viewpoint of the surgeon, the transaction is equally complex, and his or her decisions may be constrained by resource allocation and time available. However, in the final analysis, assuming a beneficent doctor, the most appropriate decision will be determined by the weight of evidence that can be adduced in favor of any intervention.

The very title - “Expensive Medical and Surgical Technology” - expresses in coded form a myth that is widely prevalent among those who consider the political and economic aspects of health care. Strangely, it is a myth that finds favor mainly at the extreme ends of the range of attitudes toward health care. Monetarists see it as an incentive toward increasing private provision of health care, while extreme egalitarians see it as another example of unjust privilege. The content of the myth is that there is a definable group of costly procedures, which can somehow be isolated from the general practice of medicine and surgery, after which such procedures can either be made the subject of special private provision or, alternatively, discarded from a rationed system of health care.

When it is well conducted, a randomized clinical provides the strongest evidence available for evaluating the comparative effectiveness of the interventions tested. Over the last two generations, we have learned much about various devices for strengthening them and about methods of avoiding between in their design, execution, analysis, and reporting. In a trial, we seek evidence for a causal link between treatment and observed outcomes. Becaues the controlled trial depends on an argument based on exculsion (i.e., no other causes or differences affected the experimental groups), we strengthen its inference by taking steps to exclude any such differences.

This article discusses a number of issues that deserve consideration: problems of multiplicity and generalizability, devices for strengthening trials, issues of power and sample size, the relationship between study design and reported gains, when to undertake a trial, the role of collaborative trials, and ways to make trials more feasible in clinical settings.

This article considers the distinction between “explanatory” and “pragmatic” aims in clinical trials—the distinction between testing a biological hypothesis and providing evidence to permit a choice between alternative treatment policies. The choice of treatments to compare, the selection of patients for the trial, the study size, and how the treatment comparison should be made are among the matters discussed. In general, where explanatory and pragmatic aims conflict, the pragmatic aim will often take priority.

This article reviews nonexperimental data bases, emphasizing the present uses and future opportunities of routinely collected information. Data bases are discussed in terms of appropriate research designs. Possibilities for expanding available information through new data collection and through record linkage are stressed. The relationship of nonexperimental data systems to randomized trials and to clinical decision-making is examined.

Many would agree on the need to inform patients about the risks of medical conditions or treatments and to consider those risks in making medical decisions. The question is how to describe the risks and how to balance them with other factors in arriving at a decision. In this article, we present the thesis that part of the answer lies in defining an appropriate scale for risks that are often quite small. We propose that a convenient unit in which to measure most medical risks is the microprobability, a probability of 1 in 1 million. When the risk consequence is death, we can define a micromort as one microprobability of death. Medical risks can be placed in perspective by noting that we live in a society where people face about 270 micromorts per year from interactions with motor vehicles.

Continuing risks or hazards, such as are posed by following unhealthful practices or by the side-effects of drugs, can be described in the same micromort framework. If the consequence is not death, but some other serious consequence like blindness or amputation, the microrisk structure can be used to characterize the probability of disability.

Once the risks are described in the microrisk form, they can be evaluated in terms of the patient's willingness-to-pay to avoid them. The suggested procedure is illustrated in the case of a woman facing a cranial arteriogram of a suspected arterio-venous malformation. Generic curves allow such analyses to be performed approximately in terms of the patient's sex, age, and economic situation. More detailed analyses can be performed if desired.

Microrisk analysis is based on the proposition that precision in language permits the soundness of thought that produces clarity of action and peace of mind.

At present there is no definitive test for appendicitis and no clear set of signs and symptoms, so that diagnosis is uncertain and probabilistic. A computer model for analyzing clinical data can improve physician accuracy in diagnosing appendicitis by assessing the outcome probabilities associated with the treatment options. Should a definitive test become available, the computer model will aid clinicians in determining when to test. Once our proposed model has been validated, it will not only improve physician decision-making, but also provide quality assurance feedback, generate reports for documentation, and compile a data base of many cases for study and reference.

Decision-making in medicine is a very complex process that demands input both from the doctor and from the client. From the doctor's point of view, the most important component for this decision concerns the quality of evidence available that the recommended intervention is the best available in terms of both cost and benefit. Good quality evidence demands good quality science. The randomized controlled trial is the expression of this scientific process at work within medical practice. This article reviews both the rationale and the ethics of randomized controlled trials in the epistemology of surgery. The ethical dilemma is accentuated because surgery by its very nature is invasive and often irreversible. As an illustration of the scientific and ethical dilemmas arising out of randomized controlled trials in surgery, a description of the CEA directed second-look laparotomy trial in the United Kingdom is provided. This trial may be judged essential because of the clash of attitudes between surgeons in the United States and the United Kingdom. It is unlikely that the truth lies entirely with one or other national groupings of surgeons, and this randomized trial will eventually resolve a conflict of ideas to the ultimate benefit of all patients with operable colorectal cancer.

Physicians through the ages have practiced their trade with more or less regard to the effects of their treatment on their patients' sense of well-being, if not as much as on the disease itself.

Until recently, however, little attempt has been made to measure the effect of disease upon quality of life and how this quality is or is not improved by the treatment. Several means have now evolved and in areas of medicine where the treatment may profoundly affect the patient's sense of well-being, effort is being made to gauge these consequences.

The benefit for patients with operable breast cancer treated with adjuvant systemic therapy is small, if reduction of early mortality within the context of randomized control trials is used for treatment comparison. One might consider that the 75%–85% of patients who die despite treatment are overtreated, as are patients who remain alive even without therapy within a given time frame. Larger treatment benefits in terms of avoided or delayed breast cancer relapse have been demonstrated even at early phases of follow-up in the vast majority of adjuvant trials. Exposure of all patients to adjuvant therapy at a time at which no symptoms of disease are present is detrimental in terms of quality of life.

Based on our assumption that the quality of life of the patient is typically altered both by subjective toxic effects of adjuvant treatment and by the appearance of relapse, we developed a method of comparing treatment effects in terms of time without symptoms of disease and toxicity of treatment (TWiST). Because the impact of treatment on relapse rates appears earlier than survival effects in all adjuvant therapy trials, and because the value of time without relapse in terms of the quality of life of the patients is as yet poorly defined, we have generalized our method of comparing treatment attitudes to include individual qualitative judgment values. The experience gained from integrating quality-of-life issues into clinical trials for breast cancer might also be applied to other diseases characterized by a chronic course, toxic treatments, and gains in periods of relative or absolute freedom from toxic effects and progressive disease.

The Cancer Research Campaign (CRC) Trial demonstrates a decreased incidence of local recurrence for patients randomized to prophylactic radiotherapy following mastectomy. Irradiated patients alive after 5 years, however, have a small increased risk of dying from causes other than breast cancer. How to estimate the cost-benefit of such therapy is discussed.

Of the four randomized controlled trials of mass mammography completed to date, the New York HIP trial used obsolete technology and was not designed to assess mammography in the absence of physical examination; the Swedish Two-County trial had serious methodological flaws; the U.K. trial was incompletely documented but showed no significant benefit; and the Malmö trial, which was perfect in design and execution, showed no benefit.

Paradoxicially, with each successive trial and an increasing sophistication of mammography, the benefit of screening was becoming not greater but less. This may be summarized by comparing the number of women who would have to be offered mammography for one breast cancer death prevented or postponed in the respective trials: 5,000 in the HIP trial, 13,000 in the Two-County trial, 18,000 in the U.K. trial, and 68,000 in the Malmä trial.

The enthusiasts in favor of mass mammography have created an atmosphere of false optimism. They have not informed the public about the very small potential benefit nor about the risks associated with screening, especially the risks of false positive diagnoses and of unnecessary surgical operations. This is an unacceptable situation, and the ethical vacuum in which population screening takes place needs to be filled with guidelines to protect women who are “targeted” for screening.

Coronary bypass surgery developed as another in a line of surgical procedures dating back more than 60 years. The medical profession at first assessed this procedure with time-honored anecdotal techniques. Gradually, for a variety of reasons, improved methods of comparisons worked their way into assessments of bypass surgery. Randomized controlled trials met resistance but have been very influential. Assessment of percutaneous transluminal coronary angioplasty has benefited from the knowledge generated during the last 25 years, but clinicians have been slower to apply the most advanced techniques.

Carotid endarterectomy and EC/IC bypass grafting have been widely adopted for patients considered at risk from stroke, without good evidence of efficacy. Unjustified claims for surgery usually derive from overestimating the dangers of the disease without surgery, while perioperative risks are underestimated. Inadequate follow-up and choosing irrelevant outcome measures often add to the confusion. All these factors apply to surgery for stroke. A trial of EC/lC bypass in 1,377 patients from three continents took 8 years to complete and showed no benefit in patients randomized to surgery. Reluctance to accept this result led to detailed critiques of this trial in several journals, largely based on the discovery that many patients had been operated on in some centers without having been randomized. In reply, the investigators showed that these cases did not affect the “resounding negativity of the results.” Lack of good data about the prognosis of patients with TlAs or minor strokes was the fundamental reason for so much misplaced surgical effort. This applies equally to carotid endarterectomy, for which large trials are currently being completed. Had there been a reliable data base of patients at risk of stroke, prospectively collected and followed, the efficacy of these two operations could have been determined much sooner, and inappropriate diffusion might have been prevented.

This article develops arguments for the use of decision theory, rather than intuition, to determine the size of trials. It is wrong to expect doctors to ignore personal preferences in favor of clinical experiments unless the trial is capable of showing differences in treatment effect that would influence clinical practice substantially. It follows from our analysis that if delta (the treatment effect that the trial is designed to detect) is sufficient to alter clinical practice, then the alpha and beta errors of a trial should be equal. This applies even if a new treatment is to be compared with conventional therapy or if a treatment with high “costs” is compared with a less invasive or more inexpensive method.