PubMed Health⌕ Search

Biomedical subjects

S D Walter

Publications and source records attributed to S D Walter.

At least 19 recordsLinked to original sources
PubMed1992

The analysis of regional patterns in health data. I. Distributional considerations.

Regional patterns of health data such as cancer incidence rates are often examined for evidence of environmental effects. In this paper, three measures of spatial clustering are evaluated for use with epidemiologic data. In particular, the effects of variation in regional population structure on the distribution of these measures is considered. It is shown that substantial bias occurs if variation in regional population size is ignored (as has occurred in previous analyses). On the other hand, the methods are robust to small case frequencies and to variation in the regional age distribution. It is recommended that these regional differences be routinely taken into account, which can be done with relatively little additional computation. A companion paper (Walter SD. The analysis of regional patterns in health data. II. The power to detect environmental effects.

Bias↗
PubMed1992

The analysis of regional patterns in health data. II. The power to detect environmental effects.

Three measures of spatial clustering (Moran's I, Geary's c, and a rank adjacency statistic, D) were evaluated for their power to detect regional patterns in health data. The patterns represented various environmental effects: a latitude gradient; residence near a contaminated water supply; disease "hot spots"; relation to socioeconomic status and urbanization; and general spatial autocorrelation. While the methods had high power to detect certain patterns, they were also affected by factors such as the shape of the map, its regional structure, and the spatial distribution of explanatory variables. The power was sometimes low, even for strong geographic trends, particularly for D. Moran's I had the highest power most often. We conclude that use of these methods requires careful specification of the anticipated geographic pattern and awareness of idiosyncratic effects in the study of particular maps.

Bias↗
PubMed1992

Use of host factors to identify people at high risk for cutaneous malignant melanoma .

OBJECTIVE: To determine which host characteristics are risk factors for cutaneous malignant melanoma in order to aim prevention and early detection programs at people at high risk. DESIGN: Case-control study. SETTING: Southern Ontario. SUBJECTS: The 583 case subjects were aged 20 to 69 years and had had malignant melanoma newly diagnosed between Oct. 1, 1984, and Sept. 30, 1986. The 608 control subjects were randomly selected from a list of residents in the study area and were stratum matched for age, sex and municipality. INTERVENTION: Through in-person interviews the interviewer ascertained exposure to putative external risk factors and assessed skin colour and number of nevi on the arm, and the subject reported his or her natural hair colour at age 20 years, eye colour, skin reaction to repeated sun exposure, and freckle and whole-body nevus densities. RESULTS: Although all the host factors mentioned were significantly associated with melanoma risk when considered separately, only hair colour, skin reaction to repeated sun exposure, and self-reported freckle and nevus densities remained significant after backward logistic regression analysis. The odds ratio for melanoma was estimated to be 10.7 in people who had many nevi compared with those who had none (95% confidence interval [CI] 6.6 to 17.4), 4.0 in people who had red hair compared with those who had black hair (95% CI 1.9 to 8.2), 1.9 in people who had many freckles compared with those who had none or few (95% CI 1.3 to 2.8) and 1.4 respectively in people who burned and had a subsequent increase in tan and those who burned and had no increase in tan after repeated sun exposure compared with those who did not burn [corrected]. CONCLUSIONS: Four risk factors for malignant melanoma have been identified. Prospective evaluation of their predictive value should be done. In the meantime, however, these factors should be used to identify people apparently at high risk for malignant melanoma, who can then be targeted for early detection and prevention programs.

Adult↗
PubMed1992

The association of cutaneous malignant melanoma and fluorescent light exposure.

Data are presented from an interview case-control study (583 cases and 608 controls), performed in southern Ontario, Canada, from October 1984 to September 1986, on the association of cutaneous malignant melanoma with exposure to fluorescent light. Males showed a significant trend with cumulative years of occupational exposure and with various indices of exposure to domestic fluorescent light. The risk was more pronounced for lesions on the arms and for superficial spreading melanomas. There was no consistent association in females. These effects were similar when adjusted for other major risk factors for melanoma, including the amount of time spent outdoors occupationally. Comparisons of melanoma cases interviewed before or after diagnosis revealed no evidence of rumination bias. Comparisons of sample data from the same cases and controls by interview and mail questionnaire showed reasonable levels of reliability with no evidence of recall bias. A small sample of subjects was also selected for exposure validation with employers; this revealed very accurate recall of occupational exposure. On the basis of these results, previous epidemiologic studies, and clinical and animal evidence, the authors conclude that fluorescent light exposure remains a potential risk factor for melanoma.

Adult↗
PubMed1992

The natural history of lung cancer estimated from the results of a randomized trial of screening.

The results from a randomized controlled trial of screening for lung cancer in Czechoslovakia have been used to estimate parameters of the natural history, using a model to simulate the disease process and the effects of screening. The results suggest that the period before clinical presentation during which lesions can be detected by screening is very short (seven to eight months). This implies that to detect three-quarters of all lung cancers by screening, two examinations per year are necessary, and that such a program would advance diagnosis by six months if there were complete participation. The results of the trial itself suggest that the benefit, in terms of a reduction in mortality from lung cancer, is likely to be very small.

Adult↗
PubMed1991

Reliability of interviewer and subject assessments of nevus counts in a study of melanoma.

Several types of data are presented concerning the reliability of counting or estimating the density of nevi (moles), a major risk factor for melanoma, using methods typically employed in epidemiologic studies. First, interviewer-derived counts of nevi on the arm produced estimates of inter-observer, inter-subject, temporal and random variability, and their interactions. Second, interviewer-derived arm counts and respondent self-reports of whole body nevus density were compared. Finally, we compared male and female cases and controls with respect to their reported rates of having a relative with a malignant mole. Overall, the intra-observer reliability ranged from 55 to 81%, and was better for observers with more experience. The correlation between the interviewer counts and respondents' self-reported estimates was 0.41. The data on malignant moles in relatives suggest higher reporting rates in male cases and lower reporting in male controls relative to their female counterparts, but there is little difference by sex in the reporting of one's own nevus density.

Adult↗
PubMed1991

Statistical significance and fragility criteria for assessing a difference of two proportions.

This paper compares the traditional methods of statistical inference on the data from biomedical studies with a proposed index of fragility in the results. In general, for any given study there are 8 possible combinations of conclusions regarding statistical significance, quantitative significance and fragility. The 8 possibilities are considered in turn with respect to how studies in each group might be interpreted. Numerical examples show that not all 8 possibilities need be attainable with a given study design, and that the relative likelihood of them occurring can vary widely. It is concluded that the fragility index may be a useful adjunct to conventional statistical inference, with certain intuitive appeal, but that more empirical experience is needed with the fragility method.

Analysis of Variance↗
PubMed1991

Mapping mortality and morbidity patterns: an international comparison.

A set of 49 national, intranational and international health atlases was surveyed to characterize their mapping methodology with respect to the populations covered, the diseases represented, the mapping techniques, and statistical methods. Little consistency was found concerning the choice of data function to be mapped, minimum event frequency requirements, method of age standardization, or map colour systems. Many atlases did not include basic epidemiological information; for instance, approximately half the atlases did not quote population denominators. There was a tendency to emphasize statistical significance over rate values, and to focus on high rather than low risk. Very few atlases included supplemental information on environmental factors. Most adopted a descriptive stance, and attempted no aetiological interpretation. We conclude that inter-atlas comparisons are made very difficult by methodological differences, and that even regional comparisons within atlases should be made cautiously. We propose a set of methodological guidelines for consideration in future atlases.

Epidemiology↗
PubMed1991

The ecologic method in the study of environmental health. I. Overview of the method.

This paper summarizes the salient features of the ecologic method, with emphasis on its application in the study of environmental health. Various types of ecologic design are described, with examples. Finally, the main advantages and disadvantages are indicated. A companion paper discusses the methodology of ecologic designs in more detail and describes a census of data sets with potential suitability for the ecologic study of water quality and human health.

Data Interpretation, Statistical↗
PubMed1991

The ecologic method in the study of environmental health. II. Methodologic issues and feasibility.

This paper reviews some methodological aspects of ecologic studies of human health, with emphasis on investigations of environmental quality. A recent census of Canadian and U.S. data sets potentially suitable for this type of study is summarized. It is concluded that despite the considerable utility of the ecologic design for this purpose, substantial practical difficulties are common in their implementation. Particular problems are the relative scarcity of relevant environmental data and complications associated with rendering them compatible with health data.

Canada↗
PubMed1991

A comparison of several point estimators of the odds ratio in a single 2 x 2 contingency table.

The relative performance of the unconditioned maximum likelihood estimators (UMLEs), conditional MLEs (CMLEs), and Jewell-type estimators of the odds ratio (OR) and its logarithm were investigated in sets of single 2 x 2 contingency tables. The tables were generated by complete enumeration of all possible cell frequencies consistent with a single fixed margin. The bias, mean squared error (MSE), and average absolute error (AAE) were computed for all estimators using the individual table probabilities as weights. The results showed that, for the OR, Jewell's estimator usually had smaller bias, MSE, and AAE than either of the MLEs. While the differences were often slight for MSE and AAE, for bias it was sometimes substantial. For the log(OR), the UMLE usually had the lowest bias, and its MSE and AAE were only slightly greater than those for the other estimators. Overall, we recommend estimation on the log scale using the UMLE. If OR is to be estimated, Jewell's method had strong merit, although it is nonsymmetric with respect to the table orientation. In view of this, the UMLE may again be favoured in some situations.

Bias↗
PubMed1991

The estimation of sensitivity and specificity in colorectal cancer screening methods.

The sensitivity and specificity of three screening tests for colorectal cancer were evaluated using latent class analysis. This type of analysis is useful in situations where screening tests are performed on each person and follow-up diagnostic test results are not available for individuals with negative test results. Traditional methods of evaluation, which assume these individuals to be disease free, may be biased under these circumstances. In addition to providing the parameter estimates, the latent class technique gives standard errors and permits significance tests for differences in sensitivity and specificity. It was found that the radial immunodiffusion technique was significantly more sensitive (p less than 0.001) than either the rehydrated or nonrehydrated Hemoccult II tests for detecting occult blood in patients with cancer or adenoma. While comparable to the rehydrated Hemoccult II test in terms of specificity, the radial immunodiffusion technique was significantly less specific (p less than 0.01) than the nonrehydrated Hemoccult test. Similar results were observed when restricting the analysis to cancer only.

Canada↗
PubMed1990

A comparison of multivariable mathematical methods for predicting survival--I. Introduction, rationale, and general strategy.

This paper and the two following papers (Parts I-III) report an investigation of performance variability for four multivariable methods: discriminant function analysis, and linear, logistic, and Cox regression. Each method was examined for its performance in using the same independent variables to develop predictive models for survival of a large cohort of patients with lung cancer. The cogent biologic attributes of the patients had previously been divided into five ordinal stages having a strong prognostic gradient. With stratified random sampling, we prepared seven "generating" sets of data in which the five biologic stages were arranged in proportional, uniform, symmetrical unimodal, decreasing exponential, increasing exponential, U-shaped, or bi-modal distributions. Each of the multivariable methods was applied to each of the seven generating distributions, and the results were tested in a separate "challenge" set, which had not been included in any of the generating sets. The research was intended not merely to compare the performance of the multivariable methods, but also to see how their performance would be affected by different statistical distributions of the same cogent biologic attributes. The results, which are presented in the second and third papers, were compared for selection of independent variables and coefficients, and for accuracy in fitting the generating sets and the challenge set.

Cohort Studies↗
PubMed1990

A comparison of multivariable mathematical methods for predicting survival--III. Accuracy of predictions in generating and challenge sets.

This paper concludes a study of "performance variability" when four methods of multivariable analysis--multiple linear regression, discriminant function analysis, multiple logistic regression, and two arrangements of Cox's proportional hazards regression--were applied to the same stratified random samples of "generating sets" containing seven different statistical distributions of cogent biologic attributes in a composite staging system for a large cohort of patients with lung cancer. Each model developed from the generating sets was also applied for predictions in a previously sequestered "challenge set". Across the different generating sets, the multivariable methods showed good agreement with one another in the stepwise choice of first two powerful predictor variables, but not in the sequence of subsequent choices or in the standardized coefficients assigned to the same collection of "forced" variables. In concordance of predictions for individual patients in the generating sets, the overall proportions of disagreement for pairs of methods ranged from 0 to 28%, and kappa values ranged from 0.49 to 1.00. The accuracy of individual predictions showed relatively similar results when the different methods were applied to the same generating set. Across the generating sets, the different methods showed similar total results but substantial variations in predictions for alive and dead patients. When the models from the generating sets were applied for predictions in the challenge set, the results showed an analogous pattern: similar accuracy within models for overall and live/dead predictions, but substantial variations in live/dead predictions across models derived from different generating sources. The results showed that the multivariable methods often had good agreement with one another in predictions for groups but not for individual persons; and that no single method was superior to the others or to the composite staging system. We conclude that multivariable analytic methods may be most effective and consistent if used to find the few most powerful predictor variables, omitting the many other variables that may be "statistically significant" but less cogent. The powerful predictors may sometimes be best constructed, before the analysis begins, as composite variables containing appropriate unions or ordinal arrangements of elemental candidate variables.

Cohort Studies↗