What are the challenges in attributing causation

It is sometimes difficult to decide what constitutes a true adverse drug effect. As an extreme example, a patient in a diabetes trial of glycemic control was involved in a car accident that caused the death of the other driver. Is it possible that the accident could have been caused by the study subject suffering a hypoglycemic attack Should this fatality be classified as an adverse effect and the consequence of tight control of blood sugar If the universe of all possible adverse effects were...

What are the types of observational studies

Case reports are individual patient cases. They are the simplest form of an observational study. 2. Case series are compilations of several case reports. 3. Cross-sectional studies are based on measurements obtained at a single time point with no follow-up. 4. Case-control studies compare subjects who have a particular condition (cases) to those without that condition (controls). Investigators seek to determine after the fact (i.e., retrospectively) whether differences in past exposure to one...

What is the role of qualitative studies

Randomized clinical trials and observational studies rarely consider individual patients' perceptions or preferences. Qualitative studies fill this void. One must not forget that patients may regard risks and benefits very differently, and patients with the same medical condition may have very different symptoms, as well as completely different thresholds of tolerance for these symptoms. Some patients with asthma may suffer primarily from exertion dyspnea, while others have difficulty coping...

Which symptoms are most relevant

It is important to limit the number of questions patients are asked and to focus on those symptoms they perceive as the most troublesome or that are the most responsive to the therapy. Epigastric pain is the most frequently reported symptom for patients with an acute duodenal ulcer. Most patients want rapid pain relief and actual healing of the ulcer is a secondary concern. Therefore, it makes sense that patients' self-reported perception of epigastric pain should be the primary outcome in...

Do surrogate markers predict benefit in individuals

It has been generally assumed that only patients with hypercholesterolemia or hypertension benefit from lipid-lowering or antihypertensive treatment. Recent trial reports, however, have raised questions about these assumptions. The Heart Protection Study1 investigated simvastatin (Zocor) vs. placebo taken over 5 years in 20,500 subjects. The fairly unselected study population included those with normal and abnormal serum lipids, as well as those with and without a history of vascular disease....

Can net benefit be determined

In Chapter 2, we emphasized the importance of considering the benefit-to-harm balance when making treatment decisions. This weighing of favorable and unfavorable treatment effects often relies on different sources and types of information. The use of a composite outcome provides an opportunity to balance diverse treatment effects. It has not been conclusively documented that glycemic control in patients with Type 2 diabetes leads to a reduction in major cardiovascular events. In the...

Why clinical trials have a limited value for detection of unexpected adverse events

Many adverse effects of drugs are unexpected. For this reason, it may take several years before the occurrence of an adverse event is recognized and attributed to a specific medication. Terfenadine (Teldanex) had been on the market for more than a decade prior to the detection of a serious interaction with erythromycin or grapefruit juice that could cause a fatal arrhythmia.1 Phenylpropanolamine (PPA) had been marketed even longer before there was a signal that it might cause intracranial...

Should changes in a surrogate marker be extrapolated within a drug class

Favorable changes in a surrogate marker are sometimes insufficient to recommend full approval and widespread use of the first drug of a particular class. Clinicians and regulatory agencies prefer to see trial evidence of event reductions. When simvastatin (Zocor) was shown to reduce total and LDL-cholesterol, the FDA approved the drug, but asked for outcome trials. When the first large simvastatin trial, 4S,6 showed a convincing reduction in all-cause mortality in coronary patients, use of the...

Why are clinical trials unreliable for the detection of rare adverse events

When a new pharmaceutical agent is approved for marketing, approximately 1,000 to 5,000 patients have typically been exposed to it. This number is adequate to determine the frequency of common, predictable, and easily recognized adverse effects. However, the likelihood of detecting rare adverse effects is small in clinical trials. This becomes more of an issue if the rare event is serious. Approximately 3,000 patients are required to detect a single case with 95 probability if its true...

What went wrong

Ventricular extrasystoles on the electrocardiogram, serum cholesterol, and bone density serve as examples of surrogate endpoints. The presence of ventricular extrasystoles in coronary patients is associated with higher mortality and sudden death, elevated levels of serum cholesterol are related to the risk of acute myocardial infarction and premature death, and low bone density is linked to the risk of fractures. In principle, treatment that reduces the frequency of ventricular extrasystoles,...

Was the result for the primary outcome clearly stated in the conclusion and supported by the confidence interval

The major conclusion(s) of any clinical trial should be based on the findings of the pre-specified primary outcome. Because deviations from this fundamental principle are common, journal readers should be on their guard when reviewing trial results. Non-significant statistical findings for the primary outcome may be overinterpreted. Nominally statistically significant findings for one or more of several pre-specified secondary outcomes may be highlighted rather than drawing attention to a...

Was selection of the activecontrol fair

In evaluating an antihypertensive drug, the optimal comparator should be a low-dose diuretic, which has proven to be highly beneficial in reducing all vascular complications of hypertension, is fairly safe when properly used, and is very inexpensive.14 Since showing superiority over, or even equality with, a generic diuretic is difficult, a manufacturer of a novel class of antihypertensive agents or of a new member of an established class (so-called 'me-too' drugs) may tip the balance in favor...