Search PubMed⌕ Search

Biomedical subjects

Douglas G Altman

Publications and source records attributed to Douglas G Altman.

At least 37 records · Page 2Linked to original sources
PubMed2006

Decision tools in health care: focus on the problem, not the solution.

BACKGROUND: Systematic reviews or randomised-controlled trials usually help to establish the effectiveness of drugs and other health technologies, but are rarely sufficient by themselves to ensure actual clinical use of the technology. The process from innovation to routine clinical use is complex. Numerous computerised decision support systems (DSS) have been developed, but many fail to be taken up into actual use. Some developers construct technologically advanced systems with little relevance to the real world. Others did not determine whether a clinical need exists. With NHS investing 5 billion pounds sterling in computer systems, also occurring in other countries, there is an urgent need to shift from a technology-driven approach to one that identifies and employs the most cost-effective method to manage knowledge, regardless of the technology. The generic term, 'decision tool' (DT), is therefore suggested to demonstrate that these aids, which seem different technically, are conceptually the same from a clinical viewpoint. DISCUSSION: Many computerised DSSs failed for various reasons, for example, they were not based on best available knowledge; there was insufficient emphasis on their need for high quality clinical data; their development was technology-led; or evaluation methods were misapplied. We argue that DSSs and other computer-based, paper-based and even mechanical decision aids are members of a wider family of decision tools. A DT is an active knowledge resource that uses patient data to generate case specific advice, which supports decision making about individual patients by health professionals, the patients themselves or others concerned about them. The identification of DTs as a consistent and important category of health technology should encourage the sharing of lessons between DT developers and users and reduce the frequency of decision tool projects focusing only on technology. The focus of evaluation should become more clinical, with the impact of computer-based DTs being evaluated against other computer, paper- or mechanical tools, to identify the most cost effective tool for each clinical problem. SUMMARY: We suggested the generic term 'decision tool' to demonstrate that decision-making aids, such as computerised DSSs, paper algorithms, and reminders are conceptually the same, so the methods to evaluate them should be the same.

Attitude to Computers↗
PubMed2006

Dichotomizing continuous predictors in multiple regression: a bad idea.

In medical research, continuous variables are often converted into categorical variables by grouping values into two or more categories. We consider in detail issues pertaining to creating just two groups, a common approach in clinical research. We argue that the simplicity achieved is gained at a cost; dichotomization may create rather than avoid problems, notably a considerable loss of power and residual confounding. In addition, the use of a data-derived 'optimal' cutpoint leads to serious bias. We illustrate the impact of dichotomization of continuous predictor variables using as a detailed case study a randomized trial in primary biliary cirrhosis. Dichotomization of continuous data is unnecessary for statistical analysis and in particular should not be applied to explanatory variables in regression models.

Age Factors↗
PubMed2006

The Oxford Conception Study design and recruitment experience.

The Oxford Conception Study is a randomised controlled trial that aims to determine whether or not information about potential fertility from a device that monitors urinary hormones will increase the conception rate in women wishing to conceive. Three modified versions of a fertility monitor have been developed for the study. The monitor measures the levels of urinary oestrone-3-glucuronide (E3G) and luteinising hormone (LH), and the display indicates high or low fertility. The monitor requests all women to test their urine from day 6 to day 25 of the menstrual cycle inclusive. One-third of women are randomised to receive information from the fertility monitor about the early fertile time (from the first rise in E3G until the LH surge is detected), one-third receive information about the late fertile time (the onset of the LH surge and the following 2 days), and a third do not receive any information (control group). All the women are followed up for 6 months or until they are pregnant. A total of 1453 women have been recruited into the study, reaching the study recruitment goal for 80% power to detect a 10% difference in three-cycle pregnancy rate between the Late Fertile Time group (50%) and the Control group (40%), allowing for a 15% non-pregnancy drop-out rate. Follow-up of the women is currently ongoing. The primary analysis will compare the cumulative three-cycle pregnancy rate between each of the study arms. Time-specific conception probabilities will be estimated from coitus information recorded in 12-h intervals. The data from this study will also allow many additional questions to be addressed, including changes in intercourse patterns with feedback about the fertile days and other questions in relation to menstrual cycle function, sexual intercourse, stress, exposures to tobacco products, alcohol, caffeine and medications, fertility and pregnancy outcomes. In addition to presenting the study design, we review the recruitment experience for the Oxford Conception Study. We have achieved sustained and effective recruitment over time by primary use of recruiting via the Internet.

Adolescent↗
PubMed2006

Polyp measurement using CT colonography: agreement with colonoscopy and effect of viewing conditions on interobserver and intraobserver agreement.

OBJECTIVE: This article presents inter- and intraobserver agreement for estimates of polyp diameter using CT colonography, including the effects of different visualization displays and prior experience. MATERIALS AND METHODS: Four observers, three of whom had prior experience with CT colonography, estimated the maximum diameter of 48 polyps using three different visualization displays: 2D colonography window, 2D abdominal window, and 3D surface rendering. Each re-measured a subset of 10 polyps. Polyps measured 2 to 12 mm according to a colonoscopic reference. Inter- and intraobserver agreement and agreement with the reference measurement were determined using the Bland-Altman method, paired Student's t testing, analysis of variance, and analysis of covariance (ANCOVA), and by calculating the components of variance. RESULTS: CT measurements overestimated polyp diameter, a phenomenon found least using the 2D abdominal display. Generally, 95% limits of agreement encompassed different size categories for individual polyps: the widest spanned 14.6 mm (-4.6 mm to 10.0 mm) for an experienced observer using the 3D display. When using the 2D abdominal display, no significant difference was found between estimates and the reference value for the other two experienced observers (p = 0.83 and 0.23). All the observers' measurements were significantly different from the reference when using the 3D display (p < 0.001). The novice was significantly different from the experienced observers in some analyses. Inter- and intraobserver agreement were poorest for the 3D display. CONCLUSION: Measurement of polyp diameter from CT colonography is subject to variation contingent on the observer's experience and the viewing display used. Although 3D visualization display is commonly used for polyp detection, it should not be used for measurement.

Colonic Polyps↗
PubMed2005

[Developing guidelines for reporting healthcare research: scientific rationale and procedures].

Healthcare research is extremely important, with the clear potential to influence treatment and prevention of disease, possibly on a global scale. Yet deficiencies in how research studies are reported are both well-documented and widespread across all medical specialties, types of interventions and study designs. Although randomised trials have received the most attention in this regard, similar concerns have been expressed about reporting of other types of research including diagnostic and epidemiological studies. If a journal article describes in enough detail what was done at each stage of a study, readers will have enough information to allow them to decide on the merits of the results for themselves. From this simple idea comes the scientific rationale of developing guidelines on how to report research. We consider desirable attributes of such guidelines, emphasising the importance of being evidence-based where possible. We describe recommended processes to produce guidelines that have evolved over several years during the preparation of a sequence of reporting guidelines including CONSORT, QUOROM, STARD and STROBE.

Authorship↗
PubMed2005

[The CONSORT statement for cluster randomised trials].

With the publication of the CONSORT statement there is now increased awareness of the need to adequately report the findings of randomised controlled trials. The CONSORT statement includes a checklist of items that should be addressed in the trial report. The original CONSORT statement was developed to ensure the appropriate reporting of parallel group randomised controlled trials in which individual participants are allocated to different intervention groups. In cluster randomised trials, however, groups of participants, rather than individuals, are randomly allocated to study groups. The process of allocating groups of participants raises additional reporting considerations and led to the publication of an extension to the CONSORT statement specifically for cluster randomised trials. In this paper we review the CONSORT extension to cluster randomised trials, outlining the special features of the cluster randomised trial which must be considered.

Authorship↗
PubMed2005

[Diagnostic (STARD) and prognostic (REMARK) studies].

Deficiencies in how research studies are reported are both well-documented and widespread across all medical specialties and study designs. Although randomised trials have received the most attention in this regard, similar concerns have been expressed about reporting of other types of research including diagnostic and epidemiological studies. If a journal article describes in enough detail what was done at each stage of a study, readers will have enough information to allow them to decide on the merits of the results for themselves. From this simple idea comes the scientific rationale of developing guidelines on how to report research. Recommended processes to produce reporting guidelines have evolved over several years during the preparation of a sequence of reporting guidelines starting with CONSORT and QUOROM in the 1990s. We describe initiatives to develop reporting guidelines for diagnostic accuracy studies (STARD) and tumour marker prognostic studies (REMARK).

Biomarkers, Tumor↗
PubMed2005

Bayesian random effects meta-analysis of trials with binary outcomes: methods for the absolute risk difference and relative risk scales.

In a recent Statistics in Medicine paper, Warn, Thompson and Spiegelhalter (WTS) made a comparison between the Bayesian approach to the meta-analysis of binary outcomes and a popular Classical approach that uses summary (two-stage) techniques. They included approximate summary (two-stage) Bayesian techniques in their comparisons in an attempt undoubtedly to make the comparison less unfair. But, as this letter will argue, there are techniques from the Classical approach that are closer-those based directly on the likelihood-and they failed to make comparisons with these. Here the differences between Bayesian and Classical approaches in meta-analysis applications reside solely in how the likelihood functions are converted into either credibility intervals or confidence intervals. Both summarize, contrast and combine data using likelihood functions. Conflating what Bayes actually offers to meta-analysts-a means of converting likelihood functions to credibility intervals-with the use of likelihood functions themselves to summarize, contrast and combine studies is at best misleading.

Bayes Theorem↗
PubMed2005

[Selective reporting of positive outcomes in randomised trials--secondary publication.. A comparison of protocols with published reports].

INTRODUCTION: Selective reporting of positive outcomes is associated with bias. MATERIALS AND METHODS: We conducted a cohort study of 102 protocols of randomised trials approved by the Scientific Ethics Committees for Copenhagen and Frederiksberg in 1994-1995 and the corresponding 122 reports. RESULTS: The median proportion of incompletely reported outcomes per trial was 50%. Statistically significant outcomes were more often fully reported, odds ratio 2.4. In 62% of trials, at least one primary outcome was changed. DISCUSSION: Outcomes in randomised trials are often reported selectively.

Clinical Protocols↗
PubMed2005

Reporting recommendations for tumor marker prognostic studies (REMARK).

Despite years of research and hundreds of reports on tumor markers in oncology, the number of markers that have emerged as clinically useful is pitifully small. Often, initially reported studies of a marker show great promise, but subsequent studies on the same or related markers yield inconsistent conclusions or stand in direct contradiction to the promising results. It is imperative that we attempt to understand the reasons that multiple studies of the same marker lead to differing conclusions. A variety of methodologic problems have been cited to explain these discrepancies. Unfortunately, many tumor marker studies have not been reported in a rigorous fashion, and published articles often lack sufficient information to allow adequate assessment of the quality of the study or the generalizability of study results. The development of guidelines for the reporting of tumor marker studies was a major recommendation of the National Cancer Institute-European Organisation for Research and Treatment of Cancer (NCI-EORTC) First International Meeting on Cancer Diagnostics in 2000. As for the successful CONSORT initiative for randomized trials and for the STARD statement for diagnostic studies, we suggest guidelines to provide relevant information about the study design, preplanned hypotheses, patient and specimen characteristics, assay methods, and statistical analysis methods. In addition, the guidelines suggest helpful presentations of data and important elements to include in discussions. The goal of these guidelines is to encourage transparent and complete reporting so that the relevant information will be available to others to help them to judge the usefulness of the data and understand the context in which the conclusions apply.

Biomarkers, Tumor↗