Katherine Fradette

The New and Improved Two-Sample t Test

This article considers the problem of comparing two independent groups in terms of some measure of location. It is well known that with Students two-independent-sample t test, the actual level of significance can be well above or below the nominal level, confidence intervals can have inaccurate probability coverage, and power can be low relative to other methods. A solution to deal with heterogeneity is Welchs (1938) test. Welchs test deals with heteroscedasticity but can have poor power under arbitrarily small departures from normality. Yuen (1974) generalized Welchs test to trimmed means; her method provides improved control over the probability of a Type I error, but problems remain. Transformations for skewness improve matters, but the probability of a Type I error remains unsatisfactory in some situations. We find that a transformation for skewness combined with a bootstrap method improves Type I error control and probability coverage even if sample sizes are small.

Canadian breast cancer guidelines: Have they made a difference?

Background: A principal objective of the Canadian Clinical Practice Guidelines for the Care and Treatment of Breast Cancer was to reduce the variation in the way that breast cancer was being treated. To evaluate whether this goal has been reached, we examined variations among surgeons for 4 measures of surgical care and tested for differences in province-wide rates and in variations among surgeons before and after the guidelines were released. Methods: We studied a population-based cohort of 7022 women living in Manitoba in whom breast cancer was diagnosed from 1995 to 2003 inclusive. Demographic, tumour and treatment information was obtained from the Manitoba Cancer Registry. We examined 4 measures of care: breast-conserving surgery, axillary assessment in invasive disease, axillary node dissection in noninvasive disease and the adequacy of axillary node dissection. Generalized linear models were used to test for significant variations in care among surgeons and to test for differences in province-wide rates and variations in these rates among surgeons before and after introduction of the guidelines. Results: We found clinically significant variations in the province-wide rates of all 4 measures examined. These variations were statistically significant for all measures except axillary node dissection in noninvasive disease. No significant differences in either the province-wide rates or in variations in these rates among surgeons before and after introduction of the guidelines were found for any of the measures. Interpretation: Our results suggest that the Canadian breast cancer guidelines are not meeting their stated objective. New strategies for guideline dissemination and implementation may be required.

Comparing measures of the 'typical' score across treatment groups.

Researchers can adopt one of many different measures of central tendency to examine the effect of a treatment variable across groups. These include least squares means, trimmed means, M-estimators and medians. In addition, some methods begin with a preliminary test to determine the shapes of distributions before adopting a particular estimator of the typical score. We compared a number of recently developed adaptive robust methods with respect to their ability to control Type I error and their sensitivity to detect differences between the groups when data were non-normal and heterogeneous, and the design was unbalanced. In particular, two new approaches to comparing the typical score across treatment groups, due to Babu, Padmanabhan, and Puri, were compared to two new methods presented by Wilcox and by Keselman, Wilcox, Othman, and Fradette. The procedures examined generally resulted in good Type I error control and therefore, on the basis of this critetion, it would be difficult to recommend one method over the other. However, the power results clearly favour one of the methods presented by Wilcox and Keselman; indeed, in the vast majority of the cases investigated, this most favoured approach had substantially larger power values than the other procedures, particularly when there were more than two treatment groups.

Adaptive robust estimation and testing.

We examined nine adaptive methods of trimming, that is, methods that empirically determine when data should be trimmed and the amount to be trimmed from the tails of the empirical distribution. Over the 240 empirical values collected for each method investigated, in which we varied the total percentage of data trimmed, sample size, degree of variance heterogeneity, pairing of variances and group sizes, and population shape, one method resulted in exceptionally good control of Type I errors. However, under less extreme cases of non-normality and variance heterogeneity a number of methods exhibited reasonably good Type I error control. With regard to the power to detect non-null treatment effects, we found that the choice among the methods depended on the degree of non-normality and variance heterogeneity. Recommendations are offered.

Trends in Time to Diagnosis of Colon Cancer and Impact on Clinical Outcomes

BACKGROUND There has been a rapid increase in screening for colorectal cancer (CRC) over the past several years in North America. This could paradoxically lead to worsening outcomes if the system is not adapted to deal with the increased demand. For example, this could create increased wait times for endoscopy and delayed time to CRC diagnosis, which could worsen clinical outcomes such as stage at diagnosis and⁄or survival. No previous Canadian study has evaluated the association between time to CRC diagnosis and clinical outcomes. METHODS The present historical cohort study used Manitobas population-based cancer registry and Manitoba Health administrative databases. The effect of time to diagnosis on patient survival was evaluated using Cox regression analysis with adjustment for stage at diagnosis, grade of CRC, age, sex, socioeconomic status, comorbidity index score and year of CRC diagnosis. The association between time to diagnosis and CRC stage at diagnosis was evaluated using multivariate logistic regression analysis. RESULTS The median time to CRC diagnosis increased significantly from 72 days (95% CI 61 days to 83 days) in 2004 to 105 days (95% CI 64 days to 129 days) in the first three months of 2009 (P=0.04). There was no significant association between time to diagnosis and survival. Individuals with the longest time to diagnosis were less likely to have stage III⁄IV CRC at diagnosis (quartile 4 versus quartile 1: OR 0.50 [95% CI 0.33 to 0.75). CONCLUSION Time to CRC diagnosis is continuing to increase in Manitoba. Although the present study did not detect a significant negative clinical effect of increasing time to diagnosis, additional studies are required.

Abstract P3-07-13: Outcomes of women with small, early-stage breast cancer in Manitoba from 1995-2011

OBJECTIVES: The objectives of this study were to describe the distribution, management, and outcomes of women with early-stage breast cancer in Manitoba during the period from 1995-2011. Specifically, we looked at tumours 1cm or smaller in size (T1mic/T1a/T1b) across the spectrum of molecular phenotypes: estrogen-receptor (ER) status, progesterone-receptor (PR) status, and human epidermal growth factor receptor 2 (HER2) status. This is the first study to evaluate these objectives in Manitoba, and will contribute significantly to the limited data of long-term outcomes for smaller-sized breast tumours. METHODS: Using the Manitoba Cancer Registry, we created a retrospective cohort of all patients with primary breast cancer of 1.0 cm or less, between 1995 and 2011. Women with previous or synchronous contralateral breast cancers were excluded. Data included patient demographics, diagnostic procedure, tumour size, nodal disease, treatment modalities (surgery, radiotherapy, hormone therapy, and chemotherapy), ER status, PR status, and HER2 status. Linkage to the Manitoba Health Drug Program Information Network (DPIN) database allowed capture of all hormonal therapies. Patient outcomes were evaluated, including risk of recurrence, overall survival and relative survival. Risk of recurrence was illustrated using cumulative incidence of recurrence, accounting for the competing risk of death. Kaplan-Meier curves were used to illustrate overall survival. The relative survival estimate was used to estimate the probability of surviving from breast cancer while controlling for differences in mortality due to other causes. RESULTS: Our study captured 2,341 women. Mean age at diagnosis was 62. ER+, PR+, and HER2+ status were 71% (11% unknown), 61.6% (11% unknown), and 8.6% (31% unknown), respectively. Clinically, 78% were Stage I disease, and by tumour size overall: T1mic 11.5%, T1a 18.9%, and T1b 69.7%. Ninety-eight percent had primary surgery, 58% had primary radiation, 48% received hormone therapy, and 16% received chemotherapy. At last follow-up, 21% of all patients were deceased. Relative survival estimates revealed that the survival of the overall cohort was not different than the general population, when comparing based on age groupings for Manitoban females. Recurrence occurred in 6.6% (156) of all cases. ER+, PR+, and HER2+ status were 63% (4% unknown), 58% (4% unknown), and 48% (5% unknown), respectively. By staging, 65% were Stage I disease and by tumour size 60% were T1b. Overall, 64% of recurrences were node-negative. For the HER2+, node-negative subpopulation at diagnosis: 21% of the T1mic group recurred, 41% of T1a recurred, and 19% of T1b recurred. In all recurrences, 39% of patients were deceased at last follow-up. CONCLUSIONS: Small, early-stage breast cancers are common and a significant proportion of patients recur. HER2-positivity appears to be an important risk factor for recurrence and may independently warrant treatment with trastuzumab, regardless of primary tumour size. Citation Format: Aly-Khan Lalani, Katherine Fradette, Rashid Ahmed, Debjani Grenier, Marshall Pitz. Outcomes of women with small, early-stage breast cancer in Manitoba from 1995-2011 [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P3-07-13.