Variance components and their implications for statistical information in medical data

Discriminant analysis of Raman spectra for body fluid identification for forensic purposes

Detection and identification of blood, semen and saliva stains, the most common body fluids encountered at a crime scene, are very important aspects of forensic science today. This study targets the development of a nondestructive, confirmatory method for body fluid identification based on Raman spectroscopy coupled with advanced statistical analysis. Dry traces of blood, semen and saliva obtained from multiple donors were probed using a confocal Raman microscope with a 785-nm excitation wavelength under controlled laboratory conditions. Results demonstrated the capability of Raman spectroscopy to identify an unknown substance to be semen, blood or saliva with high confidence.

Bayesian statistics in medicine: a 25 year review

This review examines the state of Bayesian thinking as Statistics in Medicine was launched in 1982, reflecting particularly on its applicability and uses in medical research. It then looks at each subsequent five-year epoch, with a focus on papers appearing in Statistics in Medicine, putting these in the context of major developments in Bayesian thinking and computation with reference to important books, landmark meetings and seminal papers. It charts the growth of Bayesian statistics as it is applied to medicine and makes predictions for the future. From sparse beginnings, where Bayesian statistics was barely mentioned, Bayesian statistics has now permeated all the major areas of medical statistics, including clinical trials, epidemiology, meta-analyses and evidence synthesis, spatial modelling, longitudinal modelling, survival modelling, molecular genetics and decision-making in respect of new technologies.

Variations in medical care for HIV-related Pneumocystis carinii pneumonia: a comparison of process and outcome at two hospitals

BACKGROUND

Institutional variation in the quality of medical care may be evaluated by examining process measures, such as use of diagnostic procedures or treatment modalities, or outcome measures, such as mortality. We undertook this study to examine variations in both process and outcome of care for patients with HIV-related Pneumocystis carinii pneumonia (PCP) at two geographically diverse, HIV-experienced, public municipal hospitals.

DESIGN

Retrospective review of hospitalized patients diagnosed as having PCP cared for at two municipal hospitals from 1988 to 1990. At hospital A, charts of all patients diagnosed as having PCP were abstracted (n=209); at hospital B, a random sample of 15% were abstracted (=136).

RESULTS

Among all hospitalized patients diagnosed as having PCP, the frequency of making a definitive diagnosis of PCP (as opposed to treating empirically) differed markedly at the two hospitals (85% in hospital A vs 26% in hospital B; p<0.001), as did the use of intensive care (18% vs 3%; p<0.001) and "do-not-resuscitate" orders (39% vs 14%; p<0.001), although the timing of starting anti-Pneumocystis medications (89% vs 88% within the first 2 hospital days) and the use of corticosteroids (21% vs 23%) were similar. Despite differences in the process of care, survival rates were similar at the two institutions (75% vs 76%; p=0.8) and remained similar when logistic regression was used to control for demographic variables and severity of illness (odds ratio for survival, hospital B vs A, 1.2 [95% confidence interval, 0.7, 2.0]). The 95% confidence intervals (0.7, 2.0), however, were consistent with a considerable (and clinically significant) disparity in survival (from 30% lower to a twofold higher odds of survival). Sample size calculations showed that a sample of 10 cases in each hospital would be required to detect the observed difference in definitive diagnosis rates (85% vs 26%), but 722 cases in each hospital would be required to detect a relevant difference in mortality.

CONCLUSIONS

The process of care for hospitalized patients with PCP in these two institutions differed considerably, but the survival rates were not significantly different, even after adjusting for confounding factors. While sample sizes available at the individual institutions were sufficient for evaluation of the process of care, they did not provide the power necessary to evaluate outcomes. Comparisons of outcomes such as mortality between individual hospitals may not have the statistical power to exclude important differences.

Leadership for quality improvement in group practices

OBJECTIVES

The authors studied influence of physician leaders on their colleagues' performance using data from a randomized, controlled trial of quality assurance interventions in 16 primary care practices.

METHODS

The authors examined whether leaders performed better than their colleagues and looked for evidence of leaders' influence on their colleagues before intervention. As behavioral indicators of each leader's influence for improvement in response to quality assurance interventions, the authors (1) change in the leader's performance score and (2) an index of leader commitment derived from the leader's participation in quality assurance interventions; these indicators were used as covariates in a comparison in experimental practice sites of leaders' and colleagues' mean case performance scores before and after intervention.

RESULTS

Leaders did not outscore their colleagues or influence their colleagues' performance scores before intervention. In response to quality assurance interventions, a leader's change in performance score significantly improved colleagues' score if the leader improved. A positive leader commitment index predicted colleagues' improvement independently of the leader change score.

CONCLUSIONS

As hypothesized, physician leaders, by the example of their behavior, influenced colleagues' performance. However, leaders exerted their influence only after receiving external stimulation for quality improvement.

Performance measurement in prostate cancer care: beyond report cards

OBJECTIVES

Review and analyze various approaches to performance measurement in health care, demonstrating the value of provider-initiated performance measurement in which ongoing monitoring of both processes and outcomes of care coupled with the use of clinical guidelines enhances performance improvement efforts. Describe some of the issues and findings associated with the use of such a methodology in prostate cancer care.

METHODS

Literature review and case study.

RESULTS

There are a number of significant limitations in the use of a "report card" methodology to improve quality and efficiency in health care. The complementary approach of combining "instrument panels" and clinical guidelines within an overall continuous quality improvement framework appears to have resulted in improved clinical outcomes and reduced costs in a six-physician urology group located in a heavily managed-care penetrated market.

CONCLUSIONS

Performance measurement tools are integral to efforts to improve outcomes and efficiency in health care. Providers of care might consider adapting the kind of performance improvement methodology described in this article. Practice benefits including improved clinical and economic outcomes are likely to follow.

Application of an artificial neural network in epiluminescence microscopy pattern analysis of pigmented skin lesions: a pilot study

In vivo epiluminescence microscopy (ELM) is a non-invasive technique which improves the clinical diagnosis of naevi and malignant melanoma by providing diagnostic criteria that cannot be appreciated by the naked eye. The present study investigated whether ELM criteria pattern analysis can be employed in an objective, observer-trained, computer-aided diagnostic system, and whether artificial neural networks (ANN) can be applied to the diagnosis of pigmented skin lesions (PSL). The ELM criteria patterns of 200 PSL oil immersion images (60 common naevi, 60 dysplastic naevi, and 80 malignant melanomas) were analysed using a standardized questionnaire. One hundred randomly assigned PSL were used as a training set for an ANN, the remaining 100 PSL serving as the test set. The ANN was trained by backward propagation according to the histological diagnosis, and its performance was compared with that of human investigators. Out of the test set the human investigators correctly diagnosed 88% of PSL and the ANN 86%. In a dichotomized model comparing common, compound, and dysplastic naevi vs. malignant melanoma, i.e. benign vs. malignant PSL, the sensitivity and specificity of human diagnosis was 95 and 90%, respectively, whereas the sensitivity and specificity of the ANN diagnosis was 95 and 88%. Our data indicate that artificial neural networks can be trained to diagnose PSL at a human expert level, based on patterns provided by ELM criteria. We suggest that this technique offers a new approach to the diagnosis of PSL.