Search PubMed⌕ Search

Biomedical subjects

S W Duffy

Publications and source records attributed to S W Duffy.

At least 109 records · Page 6Linked to original sources
PubMed1995

Comparison of single reading with double reading of mammograms, and change in effectiveness with experience.

In the prevalence round of screening for breast cancer at a single centre, the benefit of the second radiological report has been assessed using data from 33,734 women, their screening results and knowledge of the interval cancers. The service was set up under the UK National Breast Screening Programme, and the data show evidence of learning by both individuals and the team as a whole. Of the 269 cancers detected by screening, 33 would not have been diagnosed if the only report available had been the first. The recall recommendation rate associated with the first report was 6.9% and 236 cancers were detected. The recall recommendation rate associated with all queries of all the reports was 10% (3354 queries). Had these 3354 queries all been investigated (instead of the 1423 actually recalled) only a further three cancers would have been detected. Sensitivity of the programme as a whole was substantially better than that of individual radiologists, while the specificity was maintained. The decision pathway by which recalls were agreed between the two radiologists resulted in a low recall rate (4.2%) for the programme as a whole, and is a critical factor in gaining the benefit of the improved sensitivity without a concomitant deterioration in the specificity. With the passage through the prevalent round, recall rates steadily fell, the malignant to benign biopsy ratio improved and sensitivity increased. The second radiological report yielded 14% additional cancers diagnosed and contributed very significantly to good sensitivity and so to the effectiveness of screening. Economic analysis of the results will be reported in a further paper.

Breast Neoplasms↗
PubMed1995

Correction of risk estimates for measurement error in epidemiology.

Epidemiologists are under considerable pressure to acknowledge the presence of measurement error in the determination of risk factors. Repeatability and validation studies are often prescribed in conjunction with epidemiological studies. We describe some practical uses for repeatability and validation study data, in terms of correcting risk estimates for measurement error. Commonly available methods are described, with their advantages and shortcomings. A user-friendly computer program to carry out the analyses described accompanies the paper.

Case-Control Studies↗
PubMed1994

Preserved foods and nasopharyngeal carcinoma: a case-control study among Singapore Chinese.

The association between the consumption of preserved foods in infancy, childhood and adulthood and the risk of nasopharyngeal carcinoma was investigated in a case-control study among Singapore Chinese: 200 cases (73% male) were compared with 406 hospital controls (71% male). Significantly increased risks were observed in association with frequent adult consumption of salted soy beans (OR 7.3 for consumption 4 or more times per week, compared with never), canned pickled vegetables (p = 0.01; OR 4.5 for 4 or more times per week), "sze chuan chye", a salted Chinese tuber, (OR 2.4 for 4 or more times per month) and "kiam chye", salted mustard greens (OR 2.7 for 4 or more times per week). A protective effect of high vitamin-E intake in adulthood was observed (OR 0.5 for the highest third of the control range, as compared with the lowest).

Adult↗
PubMed1994

Childhood cancer incidence in relation to distance from the former nuclear testing site in Semipalatinsk, Kazakhstan.

Rates of childhood cancer between 1981 and 1990 in the 4 administrative zones of Kazakhstan were studied to assess the relationship, if any, with distance from nuclear testing sites. Risk of various cancers among children aged 14 years or younger were estimated in relation to distance from (1) a site where testing in air was performed before 1963, (2) a site where underground testing took place thereafter, and (3) a reservoir, known as "Atom Lake," created by 4 nuclear explosions in 1965. Risk of acute leukaemia rose significantly with increasing proximity of residence to the testing areas, although the absolute value of the risk gradient was relatively small. The relative risk for those living less than 200 km from the air-testing site was 1.76 compared with those living 400 km or more away from the site. Similar relative risks were observed for the underground site and "Atom Lake." There was also some evidence of increased risk of brain tumours in association with proximity to the test sites. In 2 of the 4 zones studied, there was substantial regional variation in acute leukaemia rates which was not attributable to distance from the test site. The findings may be affected by potential confounders, notably urban/rural status and ethnic factors.

Adolescent↗
PubMed1994

A case-control study of diet and lung cancer in Kerala, south India.

A total of 281 male lung-cancer patients were identified from the hospital cancer registry in the Regional Cancer Centre in Trivandrum. The controls were selected from the visitors and patients' bystanders in the hospital. The recruitment of cases and controls started in 1990, and the present study used the cases registered in the first year. The questionnaire administered to cases and controls collected information on tobacco smoking and alcohol habits. Dietary data were collected using a food frequency questionnaire and were analyzed by multiple logistic regression producing odds ratio estimates of the relative risk and deviance chi-squared tests of significance. Analysis was done on the computer package, EGRET. All models included age, education, religion and smoking to adjust for the effect of confounding. Green vegetables and bananas were found to have a protective association with lung cancer. The odds ratio associated with the highest quartile of vegetable consumption compared with the lowest was 0.32 (95% confidence interval 0.13, 0.78). Forward stepwise regression analysis indicated pumpkins and onions as the most consistently significant protective factors. Animal protein foods and dairy products were found to have a predisposing effect on lung cancer in this study. The expected influence of smoking on lung cancer (a considerable increase in risk among smokers) provided evidence of the reliability of the data. In conclusion the results from this study show that diet has a role in lung cancer aetiology, although the association is weak compared to the effects of smoking.

Adult↗
PubMed1994

Incidence of female breast cancer in relation to prevalence of risk factors in Denmark.

The extent to which changes in prevalence of risk determinants for female breast cancer could explain the temporal variation in incidence was examined using incidence figures from Denmark for the years 1943-1989. Significant increases in incidence were observed for more recent time periods and birth cohorts. Using deviance statistics from Poisson regression measures of variability explained, only a small proportion of the increase in incidence could be accounted for by fertility rates, average age at menarche and exposure to exogenous hormones. Dietary factors, on the other hand, accounted for the greater proportion of the variation observed by time period or birth cohort. In particular, there was a strong positive association of incidence with alcohol consumption and a negative association with carbohydrate intake.

Adult↗
PubMed1994

Design and analysis considerations in a cohort study involving repeated measurement of both exposure and outcome: the association between genital papillomavirus infection and risk of cervical intraepithelial neoplasia.

Cohort studies commonly involve a single determination of exposure, and one or more assessments of whether or not the outcome of interest has occurred. This approach is appropriate when the exposure does not change with time, and when one can readily determine the time of outcome (for example, mortality). If either of these conditions is not met, however, we may introduce bias into the estimate of effect, since we may misclassify individual members of the cohort with respect to exposure, outcome or both. We can reduce this bias by measuring exposure and outcome on more than one occasion. In this paper, we illustrate the design and analysis issues that arise in such circumstances, by reference to an ongoing prospective study of the relationship between genital papillomavirus infection and risk of cervical intraepithelial neoplasia. This study entails annual assessments of the status of the study subjects with respect to both conditions. In particular, we examine the implications that use of this design has on the statistical power of the study.

Adult↗
PubMed1994

Trends in incidence of various cancers in Bulgaria, 1981-1990.

BACKGROUND: Bulgaria has undergone considerable social changes in the last 40 years, including a transition from a 75% rural to a 75% urban population. These changes might be expected to be reflected in disease rates. The Bulgarian cancer registry has computerized data on cancer incidence throughout the country from 1981 onwards. METHODS: Incidence rates in Bulgaria from 1981 to 1990 of cancers of the lung, stomach, large bowel, prostate and bladder in males and of cancers of the breast, lung, stomach, large bowel, cervix and corpus uteri in females were analysed with particular attention to time trends, age-specific changes in rates and urban/rural differences. Poisson regression was used for statistical analysis. RESULTS: Male rates of cancers of the lung, large bowel, prostate and bladder increased significantly over the period, with average annual increases of 0.4%, 2.9%, 1.3% and 2.3% respectively. In females, rates of cancers of the breast, cervix and corpus uteri increased significantly, with average annual increases of 1.3%, 1.9%, and 2.9%. In both sexes, stomach cancer incidence declined significantly, by 3.0% per year for males and 3.6% per year for females. Rural rates were lower than urban rates for most cancers, particularly in the higher age groups. The decline in male stomach cancer rates was confined to rural areas. The increases in rates of lung cancer and prostate cancer in males and of colorectal cancer and breast cancer in females were most rapid in urban areas. CONCLUSIONS: Substantial changes in rates of various cancers have taken place from 1981 to 1990, in particular increases in rates of breast cancer and large bowel cancer in women resident in urban areas. It is suggested that dietary changes may be responsible for some of these changes.

Adolescent↗
PubMed1993

New Swedish breast cancer detection results for women aged 40-49.

The issue of whether to screen women aged 40-49 for breast cancer is debated usually in terms of the potential mortality reduction achievable by the application of screening in this age group. Theories regarding why a significant reduction in mortality has not been observed in trials relate to the biologic behavior of tumors in this age group and the screening process itself. Survival curves with respect to node status, size, and grade of the tumor were compared among age groups in the Swedish two-county trial. In the Kopparberg part of this trial, for the 40-49 age group, predicted survival was calculated from the size, node status, and grade of cancers detected during the trial in comparison with those found in two later series of tumors, one from the 1989-1992, Kopparberg screening program, the other from the British Columbia screening program, the other from the British Columbia screening program that began in 1988. The Kopparberg arm of the Swedish two-county study used single-view mammography with extended processing but without grid; the two more recent programs used two-view mammography with extended processing and the grid. Both the Kopparberg programs used a 2-year interval. The effects of grade, node status, and size on survival in the 40-49 age group were very similar to their effects in older age groups. Predicted survival from the latter Kopparberg series was essentially the same as that for the earlier. The mortality reduction in this age group in the Kopparberg part of the Swedish two-county trial was 26%. The survival results indicate no biologic reason why screening should not be able, theoretically, to reduce mortality. Nonsignificant reductions in mortality have been observed in the Kopparberg part of the two-county trial and in the overview of Swedish trials. The similar predictive results for the two-view and one-view trials suggest that the most likely way to achieve further reductions in mortality is to reduce the interval between screens, possibly to 1 year.

Adult↗
PubMed1993

Cancer incidence in the native peoples of far eastern Siberia.

Cancer incidence rates were examined in the native peoples of the far north-east of Siberia for the years 1977-1988. Particularly high rates of cancers of the stomach, lung, oesophagus and cervix were observed. For stomach cancer, the male and female age-standardized (to the world population) rates were 103.9 per 100,000 and 50.0 per 100,000 respectively. The corresponding lung cancer rates were 109.4 and 45.7, and for oesophageal cancer 83.9 and 35.0. The age-standardized cervical cancer rate was 38.5 per 100,000. Rates of these cancers were considerably higher than in native Alaskan peoples, although the latter had higher rates of breast and colorectal cancers. The rates were also much higher than those of the migrant peoples from Russia and elsewhere who have settled in the far north-east over the past 3 centuries, particularly at younger ages. Male rates of stomach and lung cancer were highest in the paleo-Asiatic peoples of the north, whereas male oesophageal rates were highest in the Taiga people. In females, rates of stomach and oesophageal cancers were highest in the paleo-Asiatic peoples, and rates of lung cancer were highest in the Taiga nationalities. Cervical cancer rates were highest in the Amuro-Sakhalin nationalities of the south. Further research is needed at individual levels to explain the very high risks and the differences among the ethnic groups.

Adolescent↗
PubMed1993

Risk of leukaemia after chemotherapy in a case-control study in Moscow.

In a case-control study of second primary cancers in Moscow, there were 165 cases and 294 controls, matched for site of first primary, duration of follow-up since first primary and relapse history. Of the cases, 18 were of acute, non-lymphocytic leukaemia (ANLL), with 39 matched controls. Risk of ANLL was assessed with respect to chemotherapy for the first primary tumour. The chemotherapeutic agents investigated were nitrogen mustard, cyclophosphamide, procarbazine, doxorubicin, bleomycin, vinblastine, vincristine, prednisone and combinations. Increased risks were associated with use of nitrogen mustard (odds ratio = 9.94, not significant), doxorubicin (odds ratio = 11.25, 0.1 > P > 0.05) and vincristine (odds ratio = 26.57, P < 0.05). Despite the small number of cases and potential confounding by other agents, these findings, together with those of previous studies, suggest that some non-alkylating agents may predispose to second malignancies.

Adolescent↗
PubMed1992

Pickled vegetables in the aetiology of oesophageal cancer in Hong Kong Chinese.

Oesophageal cancer is common in Chinese populations, but individual-based epidemiological studies have provided little explanation. A case control study with 400 cases and 1598 controls (800 hospital and 798 general practice) was conducted among Hong Kong Chinese. In multivariate analyses, statistically significant effects on risk were detected for several potentially preventable exposures with high attributable risks (ARs). These included preference for consuming drinks or soups at high temperature (AR = 14%), infrequent consumption of green leafy vegetables (AR = 15%) and citrus fruits (AR = 26%), ingestion of pickled vegetables (AR = 29%), tobacco smoking (AR = 44%), and alcohol drinking (AR = 48%). On the assumption of multiplicative risk, the combined AR due to these exposures was 89%. This is the first case control study to show an association between pickled vegetable consumption and oesophageal cancer risk. The finding considerably strengthens the evidence for carcinogenicity of N-nitroso compounds in man and helps to explain the very high risk among Chinese.

Aged↗
PubMed1992

Cluster randomization in large public health trials: the importance of antecedent data.

Large-scale public health trials are often randomized by geographic or administrative clusters, for reasons of financial or organizational exigency. In this paper, we deal with the situation where the dependent variable is a count of events, such as mortality from, or incidence of a given disease. Simulation results show that this design may decrease power by more than 50 per cent. The lost power can largely be replaced by incorporating information on the dependent variable, within clusters, before the start of the trial. The pretrial and trial data can be analysed by negative trinomial models.

Breast Neoplasms↗