Biomedical subjects
Publications and source records attributed to F Mosteller.
During 1984-85, laminectomy rates in Massachusetts demonstrated a 2.2-fold variation among districts. Thirty-five percent of laminectomies occurred in 7 of the 108 hospitals studied. Approximately 81% of laminectomies were performed by neurosurgeons. Rates of laminectomy decline with increasing age after 65, while rates of hospitalization for lower-back pain rise.
Age-specific (< 50, 50-69 years) 10-year survival proportions in four European trials for node-negative, Stage I and II breast cancer patients differ significantly. These patients, identified as having the same stage as localized-breast cancer patients in the End Results Registry of the U.S. National Cancer Institute, had pooled survival proportions for the two age groups (0.713, 0.579) close to those of the registry (0.713, 0.607).
A new effort in medical information has gathered and organized data from randomized clinical trials and other sources in obstetrics and made them available in several forms. This accomplishment implies the possibility of data-based medical practice that depends on systematic organization of medical findings. Possible limitations in data-based medical practice are treated and the benefits of such an approach are discussed. Will it be possible and useful to generalize this obstetric effort to all of medicine, and, if so, how? With the establishment of the Cochrane Centre, the United Kingdom has launched a program directed toward this end. New ideas for disseminating the gathered information are needed. Funding for core support will be needed both to create the medicine-wide information and to maintain it.
BACKGROUND: The large volume of published randomized, controlled trials has led to a need for meta-analyses to track therapeutic advances. Performing a new meta-analysis whenever the results of a new trial of a particular therapy are published permits the study of trends in efficacy and makes it possible to determine when a new treatment appears to be significantly effective or deleterious. We describe the use of such a procedure, cumulative meta-analysis, to assess therapeutic trials among patients with myocardial infarction. METHODS: We performed cumulative meta-analyses of clinical trials that evaluated 15 treatments and preventive measures for acute myocardial infarction. RESULTS: An example of this method is its application to the use of intravenous streptokinase as thrombolytic therapy for acute infarction. Thirty-three trials evaluating this therapy were performed between 1959 and 1988. We found that a consistent, statistically significant reduction in total mortality (odds ratios, 0.74; 95 percent confidence interval, 0.59 to 0.92) was achieved in 1973, after only eight trials involving 2432 patients had been completed. The results of the 25 subsequent trials, which enrolled an additional 34,542 patients through 1988, had little or no effect on the odds ratio establishing efficacy, but simply narrowed the 95 percent confidence interval. In particular, two very large trials, the Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico trial in 1986 (11,712 patients) and the Second International Study of Infarct Survival trial in 1988 (17,187 patients) did not modify the already established evidence of efficacy. We used a similar approach to study the accumulating evidence of efficacy (or lack of efficacy) of 14 other therapies and preventive measures for myocardial infarction. CONCLUSIONS: Cumulative meta-analysis of therapeutic trials facilitates the determination of clinical efficacy and harm and may be helpful in tracking trials, planning future trials, and making clinical recommendations for therapy.
OBJECTIVE: To examine the temporal relationship between accumulating data from randomized control trials of treatments for myocardial infarction and the recommendations of clinical experts writing review articles and textbook chapters. DATA SOURCES: (1) MEDLINE search from 1966 to present; search terms used were myocardial infarction, clinical trials, multicenter studies, double-blind method, meta-analysis, and the text word "random:"; (2) references from pertinent articles and books; and (3) all editions of English-language general medical texts and manuals and review articles on treatment of myocardial infarction. STUDY SELECTION: Randomized control trials of therapies for reducing the risk of total mortality in myocardial infarction (acute and secondary prevention). Review articles and textbook chapters dealing with the general clinical management of patients with myocardial infarction. DATA EXTRACTION: Two authors read the material and recorded the results; disagreements were resolved by conference. DATA SYNTHESIS: We used the technique of cumulative meta-analysis (performing a new meta-analysis when the results of a new clinical trial are published) and compared the results with the recommendations of the experts for various treatments for myocardial infarction. Discrepancies were detected between the meta-analytic patterns of effectiveness in the randomized trials and the recommendations of reviewers. Review articles often failed to mention important advances or exhibited delays in recommending effective preventive measures. In some cases, treatments that have no effect on mortality or are potentially harmful continued to be recommended by several clinical experts. CONCLUSIONS: Finding and analyzing all therapeutic trials in a given field has become such a difficult and specialized task that the clinical experts called on to summarize the evidence in a timely fashion need access to better databases and new statistical techniques to assist them in this important task.
The models and analyses used in this study represent an important step in the continued search for the optimum use of surgery for the treatment of lower back pain. The likelihood of patients who are hospitalized with lower back pain in Massachusetts receiving either laminectomies or spinal fusions or both was increased when any of the following demographic, socioeconomic, or medical characteristics were present: white, male, well insured, young, routine admission, admitted to a medium-sized hospital, admitted to a teaching hospital, admitted to a hospital with a high occupancy rate, and discharged home.
OBJECTIVES: Individual epidemiological investigations into the association between chlorination by-products in drinking water and cancer have been suggestive but inconclusive. Enough studies exist to provide the basis for a meaningful meta-analysis. METHODS: An extensive literature search was performed to identify pertinent case-control studies and cohort studies. Consumption of chlorinated water, surface water, or water with high levels of chloroform was used as a surrogate for exposure to chlorination by-products. Relative risk estimates were abstracted from the individual studies and pooled. RESULTS: A simple meta-analysis of all cancer sites yielded a relative risk estimate for exposure to chlorination by-products of 1.15 (95% CI: 1.09, 1.20). Pooled relative risk estimates for organ-specific neoplasms were 1.21 (95% CI: 1.09, 1.34) for bladder cancer and 1.38 (95% CI: 1.01, 1.87) for rectal cancer. When studies that adjusted for potential confounders were pooled separately, estimates of relative risks did not change substantially. CONCLUSIONS: The results of this meta-analysis suggest a positive association between consumption of chlorination by-products in drinking water and bladder and rectal cancer in humans.
A search of literature using the Connecticut Tumor Registry (CTR) for technology assessment produced eight articles. These articles represent a small but apparently increasing percentage (2%) of total CTR publications. By keeping pace with the development and dissemination of medical technologies, the population-based CTR has been useful for assessing them, especially their adverse effects.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Meta-analytic investigations sometimes use assessments of research quality according to a formal protocol as a tool for improving research synthesis. We asked whether a particular quality scoring system could have a direct use in adjusting the summary estimates of a treatment difference. In an empirical study of the relation of quality scores to treatment differences in published meta-analyses of 7 groups of controlled randomized clinical trials comprising 107 primary studies, we found no relation between treatment difference and overall quality score. We also found no relation between quality score and variation in treatment difference. The level of quality scores has increased at a rate of 9% per decade for three decades, averaging 0.51 on a scale of 0 to 1 for the 1980s, and leaving much room for improvement. Nevertheless, attention to quality of studies by editors, reviewers, and authors may be raising both the level of research done and quality of the reports.
Advances in science and technology are generally the product of multiple investigations. This article discusses statistical methods for combining empirical results from a series of different experiments or clinical investigations. We delineate the steps an assessor might take in combining data from different studies and provide references for topics not discussed in detail. The article reviews some of the most commonly used statistical techniques for combining results in the medical and social sciences.
We analysed 113 reports published in 1980 in a sample of medical journals to relate features of study design to the magnitude of gains attributed to new therapies over old. Overall we rated 87 per cent of new therapies as improvements over standard therapies. The mean gain (measured by the Mann-Whitney statistic) was relatively constant across study designs, except for non-randomized controlled trials with sequential assignment to therapy, which showed a significantly higher likelihood that a patient would do better on the innovation than on standard therapy (p = 0.004). Randomized controlled trials that did not use a double-blind design had a higher likelihood of showing a gain for the innovation than did double-blind trials (p = 0.02). Any evaluation of an innovation may include both bias and the true efficacy of the new therapy, therefore we may consider making adjustments for the average bias associated with a study design. When interpreting an evaluation of a new therapy, readers should consider the impact of the following average adjustments to the Mann-Whitney statistic: for trials with non-random sequential assignment a decrease of 0.15, for non-double-blind randomized controlled trials a decrease of 0.11.
We analysed the results of 221 comparisons of an innovation with a standard treatment in surgery published in six leading surgery journals in 1983 to relate features of study design to the magnitude of gain. For each comparison we measured the gain attributed to the innovation over the standard therapy by the Mann-Whitney statistic and the difference in proportion of treatment successes. For primary treatments (aimed at curing or ameliorating a patient's principal disease), an average gain of 0.56 was produced by 20 randomized controlled trials. This was less than the 0.62 average for four non-randomized controlled trials, 0.63 for 19 externally controlled trials, and 0.57 for 73 record reviews (0.50 represents a toss-up between innovation and standard). For secondary therapies (used to prevent or treat complications of therapy), the average gain was 0.53 for 61 randomized controlled trials, 0.58 for eleven non-randomized controlled trials, 0.54 for eight externally controlled trials, and 0.55 for 18 record reviews. Readers of studies evaluating new treatments, particularly for primary treatments, may consider adjustment of the gain according to the study type.
This article presents an overview of technology assessment in the United States. The authors argue that while there are numerous institutions carrying out assessments, the United States requires an overall plan that would provide a national system for technology assessment. If technology assessment were more organized and systematized, the authors argue, it would be more efficient and would reach the public and the medical world effectively.
When it is well conducted, a randomized clinical trial 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 pitfalls in their design, execution, analysis, and reporting. In a trial, we seek evidence for a causal link between treatments and observed outcomes. Because the controlled trial depends on an argument based on exclusion (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 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.
The meanings of 18 verbal probability expressions were studied in 3 ways: (a) frequency distributions of what single number best represented each expression; (b) word-to-number acceptability functions from what range of numbers from 0% to 100% best represented each expression; and (c) number-to-word acceptability functions from which expressions were appropriate for multiples of 5% from 5% to 95%. These results agreed highly with others and were highly consistent across methods. Expressions incorporating the stem probable were quantitatively synonymous with expressions incorporating the stem likely. Except for expressions using the word chance, positive expressions (e.g., likely) were closer to 50% in meaning than corresponding negative expressions (e.g., unlikely). This method proved very useful in deriving fuzzy-set membership functions for probability words, encouraging us in our ongoing codification effort.