Search PubMed⌕ Search

Biomedical subjects

S D Walter

Publications and source records attributed to S D Walter.

At least 55 records · Page 3Linked to original sources
PubMed1998

Incidence of and risk factors for ventilator-associated pneumonia in critically ill patients.

BACKGROUND: Understanding the risk factors for ventilator-associated pneumonia can help to assess prognosis and devise and test preventive strategies. OBJECTIVE: To examine the baseline and time-dependent risk factors for ventilator-associated pneumonia and to determine the conditional probability and cumulative risk over the duration of stay in the intensive care unit. DESIGN: Prospective cohort study. SETTING: 16 intensive care units in Canada. PATIENTS: 1014 mechanically ventilated patients. MEASUREMENTS: Demographic and time-dependent variables reflecting illness severity, ventilation, nutrition, and drug exposure. Pneumonia was classified by using five methods: adjudication committee, bedside clinician's diagnosis, Centers for Disease Control and Prevention definition, Clinical Pulmonary Infection score, and positive culture from bronchoalveolar lavage or protected specimen brush. RESULTS: 177 of 1014 patients (17.5%) developed ventilator-associated pneumonia 9.0 +/- 5.9 days (median, 7 days [interquartile range, 5 to 10 days]) after admission to the intensive care unit. Although the cumulative risk increased over time, the daily hazard rate decreased after day 5 (3.3% at day 5, 2.3% at day 10, and 1.3% at day 15). Independent predictors of ventilator-associated pneumonia in multivariable analysis were a primary admitting diagnosis of burns (risk ratio, 5.09 [95% CI, 1.52 to 17.03]), trauma (risk ratio, 5.00 [CI, 1.91 to 13.11]), central nervous system disease (risk ratio, 3.40 [CI, 1.31 to 8.81]), respiratory disease (risk ratio, 2.79 [CI, 1.04 to 7.51]), cardiac disease (risk ratio, 2.72 [CI, 1.05 to 7.01]), mechanical ventilation in the previous 24 hours (risk ratio, 2.28 [CI, 1.11 to 4.68]), witnessed aspiration (risk ratio, 3.25 [CI, 1.62 to 6.50]), and paralytic agents (risk ratio, 1.57 [CI, 1.03 to 2.39]). Exposure to antibiotics conferred protection (risk ratio, 0.37 [CI, 0.27 to 0.51]). Independent risk factors were the same regardless of the pneumonia definition used. CONCLUSIONS: The daily risk for pneumonia decreases with increasing duration of stay in the intensive care unit. Witnessed aspiration and exposure to paralytic agents are potentially modifiable independent risk factors. Exposure to antibiotics was associated with low rates of early ventilator-associated pneumonia, but this effect attenuates over time.

Adult↗
PubMed1998

Sample size and optimal designs for reliability studies.

A method is developed to calculate the required number of subjects k in a reliability study, where reliability is measured using the intraclass correlation rho. The method is based on a functional approximation to earlier exact results. The approximation is shown to have excellent agreement with the exact results and one can use it easily without intensive numerical computation. Optimal design configurations are also discussed; for reliability values of about 40 per cent or higher, use of two or three observations per subject will minimize the total number of observations required.

Analysis of Variance↗
PubMed1998

Confidence in life-support decisions in the intensive care unit: a survey of healthcare workers. Canadian Critical Care Trials Group.

OBJECTIVE: To examine the relationship between intensive care unit (ICU) healthcare workers' confidence and their decision to withdraw life support. DESIGN: Cross-sectional survey of Canadian intensivists, ICU housestaff, and bedside nurses. Respondents chose the level of care (from comfort measures only to full aggressive care) for 12 patients described in clinical scenarios, and rated their confidence in their decisions. SETTING: Thirty-seven Canadian university-affiliated hospitals. PATIENTS: None. INTERVENTIONS: We used discrete data analysis models to examine the association between the chosen level of care, confidence in the decisions, the clinical scenario, and healthcare worker group. MEASUREMENTS AND MAIN RESULTS: The response rate was 1,361 (76%)/1,795; for this analysis, we used data from 1,306 respondents with completed questionnaires. Responses for each scenario varied widely among respondents. The level of care chosen was dependent on the scenario, the healthcare worker group, and the confidence with which the decisions were made (p < .001 for each). Intensivists were less aggressive than the ICU nurses, who were less aggressive than the housestaff, but the magnitude of effect was small. Overall, respondents were very confident about their decisions 34% of the time. After adjustment for clinical scenario and chosen level of care, intensivists were more confident than nurses, who were more confident than housestaff (40% of intensivists, 29% of nurses, and 23% of housestaff were very confident). In general, healthcare workers tended to be more confident when they chose extreme levels of care than when they chose intermediate levels of care. Considerable variability in responses to scenarios remained even when we considered only those responses made with the highest level of confidence. CONCLUSIONS: While confidence in decisions about withdrawal of life support increases with seniority and authority, consistency of decisions may not. When given standard information, healthcare workers can make contradictory decisions yet still be very confident about the level of care they would administer.

Aged↗
PubMed1998

Reliability and accuracy of differentiating pervasive developmental disorder subtypes.

OBJECTIVE: To evaluate the ability of the DSM-IV criteria for the pervasive developmental disorders (PDD) to reliably and accurately differentiate PDD subtypes. METHOD: The sample consisted of 143 children with various types of developmental disabilities. A diagnosis of PDD and PDD subtype was made by one clinician using information obtained from the Autism Diagnostic Interview-Revised and the Autism Diagnostic Observation Schedule. The raw data from the Autism Diagnostic Interview-Revised, clinical notes (excluding diagnostic opinion), Autism Diagnostic Observation Schedule, IQ, and other available data were independently assessed by three experienced raters, each of whom then made a separate, blind diagnosis. If there was any disagreement, a consensus best-estimate (CBE) diagnosis was made after discussion. To assess reliability, the agreement between the three raters was calculated using k. Accuracy was assessed by calculating the agreement between the clinician's diagnosis and the CBE and by calculating the error rates associated with the three raters using latent class analysis. RESULTS: The current DSM-IV criteria show good to excellent reliability for the diagnosis of PDD, Asperger's disorder (AsD), and autism, but they show poor reliability for the diagnosis of atypical autism. The clinician (compared to the CBE) had little difficulty differentiating PDD from non-PDD children and autism from AsD but had more difficulty identifying children with atypical autism. The latent class analysis also showed that the average error rates of the three raters for a differentiation of atypical autism from autism were unacceptably high. CONCLUSIONS: Although the psychometric properties of the current DSM-IV criteria for autism and AsD appear quite acceptable, there is likely to be a high rate of misclassification of children given a diagnosis of atypical autism.

Adolescent↗
PubMed1997

Variation in baseline risk as an explanation of heterogeneity in meta-analysis.

A relationship of baseline risk to treatment effect size has been suggested as a possible explanation of between-study heterogeneity in meta-analyses. To address this question, we develop regression models to examine the relationship between the logits (or other response measure) in the intervention and control groups. A weighted least squares (WLS) approach is described that allows for the heterogeneous sampling variation in the two groups, together with a correction of the coefficients for sampling error. Two approximate maximum likelihood (ML) solutions are also obtained, with or without an assumption of equal variances between groups within studies. A closed form ML solution exists with the assumption of equal variances. Both methods appear preferable to a previously suggested regression model of the log odds ratio on the control event rate; the methods proposed here use the same scale of measurement for both study groups, and eliminate an artifactual correlation in the regression error structure. The ML approach may be preferable because of its symmetric treatment of study groups, but WLS is more easily implemented with standard software. The methods are illustrated with data from meta-analyses on pre-term delivery and on therapies to lower serum cholesterol.

Anticholesteremic Agents↗
PubMed1997

Effects of dependent errors in the assessment of diagnostic test performance.

Latent class models can be used to assess diagnostic test performance when there is no perfectly accurate gold standard test available for comparison. These models usually assume independent errors between the tests, conditional on the true disease state of the subject. Maximum likelihood estimates of the prevalence of the disease and the error rates of diagnostic tests are then obtained. This paper examines the impact of error dependencies between binary diagnostic tests on the parameter estimates obtained from the latent class models. The independence model often gives parameter estimates having relatively small bias, but in some situations (for example, when disease prevalence is low and the tests have low specificity, such as in population screening) the bias may be more serious.

Bias↗
PubMed1997

Attitudes toward the use of gender-inclusive language among residency trainees. The McMaster Residency Training Program Directors.

OBJECTIVE: To explore postgraduate medical trainees' attitudes toward the use of gender-inclusive language. DESIGN: Self-administered questionnaire. SETTING: Seven residency training programs at McMaster University, Hamilton, Ont., from July 1993 to June 1994. PARTICIPANTS: Of 225 residents in the programs, 186 responded to the survey, for a response rate of 82.7%. Men and women were equally represented among the respondents. OUTCOME MEASURES: Categorization of attitudes about the use of language as gender-inclusive or gender-exclusive; characteristics predicting a gender-inclusive attitude. RESULTS: Factor analysis and Cronbach's alpha (0.90) supported the existence of a construct related to attitudes about language use, the poles of which were categorized as gender-inclusive and gender-exclusive. The authors classified residents with respect to their attitudes to language use from their responses to the questionnaire. In univariate analyses, sex, residency program and country of graduation significantly predicted a gender-inclusive attitude (p < 0.01). Only the first 2 variables were significant in a multivariate model; residency program explained 18% of the variance and sex 3%. Residents in obstetrics and gynecology and psychiatry had the most gender-inclusive attitudes, whereas residents in surgery and anesthesia had the most gender-exclusive attitudes. CONCLUSIONS: Residents' values are reflected in the language they choose to use. Language use may provide an index of underlying attitudes that may create hostile environments for female trainees.

Adult↗
PubMed1997

Outcome assessment for clinical trials: how many adjudicators do we need? Canadian Lung Oncology Group.

Considerable effort is often expended to adjudicate outcomes in clinical trials, but little has been written on the administration of the adjudication process and its possible impact on study results. As a case study, we describe the function and performance of an adjudication committee in a large randomized trial of two diagnostic approaches to potentially operable lung cancer. Up to five independent adjudicators independently determined two primary outcomes: tumor status at death or at final follow-up and the cause of death. Patients for whom there was any disagreement were discussed in committee until a consensus was achieved. We describe the pattern of agreement among the adjudicators and with the final consensus result. Additionally, we model the adjudication process and predict the results if a smaller committee had been used. We found that reducing the number of adjudicators from five to two or three would probably have changed the consensus outcome in less than 10% of cases. Correspondingly, the effect on the final study results (comparing primary outcomes in both randomized arms) would have been altered very little. Even using a single adjudicator would not have affected the results substantially. About 90 minutes of person-time per patient was required for activities directly related to the adjudication process, or approximately 6 months of full time work for the entire study. This level of effort could be substantially reduced by using fewer adjudicators with little impact on the results. Thus, we suggest that when high observer agreement is demonstrated or anticipated, adjudication committees should consist of no more than three members. Further work is needed to evaluate if smaller committees are adequate to detect small but important treatment effects or if they compromise validity when the level of adjudicator agreement is lower.

Cause of Death↗
PubMed1997

The results of direct and indirect treatment comparisons in meta-analysis of randomized controlled trials.

When little or no data directly comparing two treatments are available, investigators often rely on indirect comparisons from studies testing the treatments against a control or placebo. One approach to indirect comparison is to pool findings from the active treatment arms of the original controlled trials. This approach offers no advantage over a comparison of observational study data and is prone to bias. We present an alternative model that evaluates the differences between treatment and placebo in two sets of clinical trials, and preserves the randomization of the originally assigned patient groups. We apply the method to data on sulphamethoxazole-trimethoprim or dapsone/pyrimethamine as prophylaxis against Pneumocystis carinii in HIV infected patients. The indirect comparison showed substantial increased benefit from the former (odds ratio 0.37, 95% CI 0.21 to 0.65), while direct comparisons from randomized trials suggests a much smaller difference (risk ratio 0.64, 95% CI 0.45 to 0.90; p-value for difference of effect = 0.11). Direct comparisons of treatments should be sought. When direct comparisons are unavailable, indirect comparison meta-analysis should evaluate the magnitude of treatment effects across studies, recognizing the limited strength of inference.

AIDS-Related Opportunistic Infections↗
PubMed1997

Crossover trials with a cumulative response.

We consider the use of a crossover design when the response is long-lasting and cumulative. Examples include the treatment of chronic diseases such as cerebral palsy, and educational and psychological intervention studies. Cumulative response models are developed for the two-treatment, two-period design. Various estimators for the treatment and carryover effects are compared, and ANOVA tables are derived. The crossover design including pretreatment observations is found to be more cost-efficient than the one-period randomised design. Without pretreatment data, it may be more or less efficient, depending on the reliability of the data, and the costs of subject recruitment, treatment, and observation.

Analysis of Variance↗
PubMed1997

Doctors' statistical literacy: a survey at Srinagarind Hospital, Khon Kaen University.

Medical doctors need to keep abreast of new developments in medicine. This is often done by reading medical journals and carrying out research activities that require an understanding of statistical methods. This study was designed to assess the knowledge of statistics among doctors in Thailand. A pretested, self-administered questionnaire with nine multiple-choice questions on basic statistical issues was used. In a survey of university hospital staff, there were 365 doctors, including 156 specialists, 152 residents and 57 final year medical students (externs). The overall response rate was 40.0 per cent. The overall median number of correct answers was 4.0 (95% CI 3.0, 4.0). Specialists had a significantly higher median score, 4.0 than residents, and externs, 3.0's, (p = 0.02). Respondents who had previously attended statistical workshops had a significantly higher median score (5.0) than those who had not (3.0) (p < 0.01). These results indicate that doctors in our hospital have insufficient knowledge of the basic statistical concepts that are commonly used in medical journals. Continuing education in statistics for doctors during residency and post doctoral training must be given serious consideration.

Adult↗
PubMed1997

A logistic model for trend in 2 x 2 x kappa tables with applications to meta-analyses.

There recently has been an increased interest in examining the relationship between the baseline (control) risk of an adverse outcome and the magnitude of the treatment effect (Brand and Kragt, 1992, Statistics in Medicine 11, 2077-2082; Davey Smith, Song, and Sheldon, 1993, The British Medical Journal 306, 1367-1373; Senn, 1994, Statistics in Medicine 13, 293-294). To facilitate such an examination, we propose a logistic model in which the relationship between the treatment effect, as measured by the log odds ratio, and the baseline risk is specified parametrically. This procedure is founded on a product-binomial likelihood and generates maximum likelihood estimates of the baseline event rates and two parameters characterizing the trend in the treatment effect. We fit this model to data from a meta-analysis involving the treatment of women at risk of preterm labor and contrast our findings with those of an earlier analysis.

Biometry↗
PubMed1996

Residents' experiences of abuse, discrimination and sexual harassment during residency training. McMaster University Residency Training Programs.

OBJECTIVE: To assess the prevalence of psychological abuse, physical assault, and discrimination on the basis of gender and sexual orientation, and to examine the prevalence and impact of sexual harassment in residency training programs. DESIGN: Self-administered questionnaire. SETTING: McMaster University, Hamilton, Ont. PARTICIPANTS: Residents in seven residency training programs during the academic year from July 1993 to June 1994. Of 225 residents 186 (82.7%) returned a completed questionnaire, and 50% of the respondents were women. OUTCOME MEASURES: Prevalence of psychological abuse, physical assault and discrimination on the basis of gender and sexual orientation experienced by residents during medical training, prevalence and residents' perceived frequency of sexual harassment. RESULTS: Psychological abuse was reported by 50% of the residents. Some of the respondents reported physical assault, mostly by patients and their family members (14.7% reported assaults by male patients and family members, 9.8% reported assaults by female patients and family members), 5.4% of the female respondents reported assault by male supervising physicians. Discrimination on the basis of gender was reported to be common and was experienced significantly more often by female residents than by male residents (p < 0.01). Ten respondents, all female, reported having experienced discrimination on the basis of their sexual orientation. Most of the respondents experienced sexual harassment, especially in the form of sexist jokes, flirtation and unwanted compliments on their dress or figure. On average, 40% of the respondents, especially women (p < 0.01), reported experiencing offensive body language and receiving sexist teaching material and unwanted compliments on their dress. Significantly more female respondents than male respondents stated that they had reported events of sexual harassment to someone (p < 0.001). The most frequent emotional reactions to sexual harassment were embarassment (reported by 24.0%), anger (by 23.4%) and frustration (20.8%). CONCLUSION: Psychological abuse, discrimination on the basis of gender and sexual harassment are commonly experienced by residents in training programs. A direct, progressive, multidisciplinary approach is needed to label and address these problems.

Adult↗
PubMed1996

Empirical evidence of bias in infertility research: overestimation of treatment effect in crossover trials using pregnancy as the outcome measure.

OBJECTIVE: To determine whether crossover trials with simple pooling of data over different study periods leads to a different estimate of treatment effect compared with parallel group trials in infertility research using pregnancy as the outcome measure. DESIGN: An observational study using nine overviews that included trials with both crossover and parallel group designs. These overviews comprised 17 crossover and 17 parallel group trials. In total, there were 5,291 outcomes including 775 pregnancies. The association between study design and treatment effect estimate was analyzed using multiple logistic regression, controlling for differences in the therapeutic interventions and variations in the methodological quality of the trials. SETTING: Infertile patients in an academic research environment. PATIENTS: Infertile patients undergoing treatment efficacy evaluation in controlled trials. INTERVENTIONS: Random allocation to a variety of treatments including clomiphene citrate, hCG, IUI, tamoxifen, and bromocriptine. MAIN OUTCOME MEASURE: Estimate of bias between study designs, based on the interaction of study design and treatment in the logistic regression model. RESULTS: Crossover trials produced a larger average estimate of treatment effect compared with trials with a parallel group design, overestimating the odds ratio by 74% (95% confidence interval, 2% to 197%). CONCLUSION: The use of a crossover design for evaluating infertility treatments with outcomes that prevent patients from completing later phases of the trial should be avoided because it leads to exaggerated estimates of treatment effect and may result in erroneous inferences and clinical decisions. Furthermore, the type of study design should be taken into account when assessing the methodological quality of therapy trials in infertility.

Bias↗
PubMed1995

Methods of reporting statistical results from medical research studies.

There has recently been vigorous debate about the most appropriate way to report statistical results from medical studies, particularly concerning the relative merits of significance tests, confidence intervals, or other methods such as likelihood. Arguing that p values have often been abused in the past, some biomedical journals now require authors to emphasize confidence intervals instead. A review of the debate reveals points in favor of each of the proposed methods, and more than one type of analysis may be required to obtain a full perspective on the data. Furthermore, a change in editorial policy alone is unlikely to eliminate future statistical errors. It therefore appears unsatisfactory and coercive of editors to mandate the exclusive use of any one approach, based on speculative estimates of the likely rate of its correct usage by future authors. Instead, further work is needed to better understand how consumers of medical research interpret studies, particularly how they are influenced by the authors' methods of summarizing the data. Until then, editorial flexibility is required, so that authors can provide readers with enough information to form their own conclusions from a given set of results. Evidence from papers recently published in the American Journal of Epidemiology suggests that authors are now tending to report results using both significance testing and estimation techniques. This seems entirely appropriate, to avoid the strictures implied by rigid adherence to only one method.

Confidence Intervals↗
PubMed1995

Use of certainty of opinion data to enhance clinical decision making.

Clinical decision making and diagnostic testing procedures are frequently evaluated using indices such as sensitivity and specificity. These indices require an assumption that the underlying disease state or the diagnostic test result is binary, which may be an unreasonable assumption in many clinical situations. They also fail to recognise the potential for different degrees of diagnostic confidence. In this paper, we report some data from a clinical study of postpartum mothers in which, in addition to the diagnostic result, an expression of certainty or strength of belief was ascertained. We illustrate how the certainty of opinion data can be combined with the diagnostic opinion in order to enhance the interpretation of the study, and the implications for improving the accuracy of clinical decision making in general.

Decision Making↗