I am a Professor in the Department of Biostatistics and Medical Informatics and the Department of Population Health Sciences in the School of Medicine and Public Health at the University of Wisconsin-Madison. I have an affiliate appointment in the Department of Statistics.
I grew up in Duluth, Minnesota. I graduated from East High School in 1988. I received a BA in Mathematics and Economics in 1992 from the University of Minnesota-Duluth and an MS in Statistics in 1994 and a PhD in Statistics with emphasis in Biostatistics in 1998 from the University of Wisconsin-Madison. My PhD advisor was Murray Clayton.
I was a research scientist in the Statistical Data Analysis Center(SDAC) in the Department of Biostatistics and Medical Informatics at the University of Wisconsin-Madison, 1998-2005. I joined the faculty with a joint appointment in the Department of Biostatistics and Medical Informatics and the Department of Population Health Sciences in 2005.
I am an applied biostatistician focusing on problems in clinical and epidemiologic research. Current methodologic areas of interest include (1) multi-state models for incidence, progression and regression of ocular (and other) diseases, (2) small area estimation problems, particularly ranking, (3) spatial and spatio-temporal modeling, particularly cluster detection and high-dimensional variable selection and (4) age-period-cohort modeling.
Outside of the office, I’m an avid cyclist. You can check out my recent rides on Strava. I’m also a big movie fan. You’ll definitely see me at the Wisconsin Film Festival, one of my favorite events every year, and you can see lists of my favorite films and what I’ve been watching recently on Letterboxd.
PhD in Statistics (emphasis in Biostatistics), 1998
University of Wisconsin-Madison
MS in Statistics, 1994
University of Wisconsin-Madison
BA in Mathematics and Economics, 1992
University of Minnesota-Duluth
Prior research has shown that cancer risk varies by geography, but scan statistic methods for identifying cancer clusters in case-control studies have been limited in their ability to identify multiple clusters and adjust for participant-level risk factors. We develop a method to identify geographic patterns of breast cancer odds using the Wisconsin Women’s Health Study, a series of 5 population-based case-control studies of female Wisconsin residents aged 20-79 enrolled in 1988-2004 (cases=16,076, controls=16,795). We create sets of potential clusters by overlaying a 1 km grid over each county-neighborhood and enumerating a series of overlapping circles. Using a two-step approach, we fi=t a penalized binomial regression model to the number of cases and trials in each grid cell, penalizing all potential clusters by the least absolute shrinkage and selection operator (Lasso). We use BIC to select the number of clusters, which are included in a participant-level logistic regression model. We identify 15 geographic clusters, resulting in 23 areas of unique geographic odds ratios. After adjustment for known risk factors, confidence intervals narrowed but breast cancer odds ratios did not meaningfully change; one additional hotspot was identified. By considering multiple overlapping spatial clusters simultaneously, we discern gradients of spatial odds across Wisconsin.
Spatial and spatio-temporal cluster detection are important tools in public health and many other areas of application. Cluster detection can be approached as a multiple testing problem, typically using a space and time scan statistic. We recast the spatial and spatio-temporal cluster detection problem in a high-dimensional data analytical framework with Poisson or quasi-Poisson regression with the Lasso penalty. We adopt a fast and computationally-efficient method using a novel sparse matrix representation of the effects of potential clusters. The number of clusters and tuning parameters are selected based on (quasi-)information criteria. We evaluate the performance of our proposed method including the false positive detection rate and power using a simulation study. Application of the method is illustrated using breast cancer incidence data from three prefectures in Japan.
Background: Recent reports suggest that racial differences in breast cancer incidence rates have decreased. We examined whether these findings apply to breast cancer mortality while considering age, period, and cohort influences on both absolute and relative measures of breast cancer mortality. Methods: Using publicly available datasets (CDC Wonder, Human Mortality Database), we developed an age-period-cohort model of breast cancer mortality and breast cancer deaths as a proportion of all deaths during 1968-2019 among all women and by five race/ethnicity groups with sufficient numbers for estimation: Hispanic (all races), American Indian/Alaska Native and Asian/Pacific Islanders (regardless of ethnicity), non-Hispanic Black, and non-Hispanic White. Results: Initially increasing after 1968, age-adjusted breast cancer mortality rates have decreased among all racial/ethnic groups since 1988. The age-adjusted percent of all deaths due to breast cancer also has been declining for non-Hispanic White women since about 1990 while increasing or holding steady for other race/ethnic groups. In 2019, the age-adjusted percent of deaths due to breast cancer for women was highest for Asian/Pacific Islanders (5.6%) followed by non-Hispanic Black (4.5%), Hispanic (4.4%), non-Hispanic White (4.1%), and American Indian/Alaska Native women (2.6%). Conclusions: Breast cancer mortality disparities are now greater on both relative and absolute scales for non-Hispanic Black women, and using the relative scale for Asian/Pacific Islanders and Hispanic women, compared with non-Hispanic White women for the first time in 50 years.
In regression analysis for spatio-temporal data, identifying clusters of spatial units over time in a regression coefficient could provide insight into the unique relationship between a response and covariates in certain subdomains of space and time windows relative to the background in other parts of the spatial domain and the time period of interest. In this article, we propose a varying coefficient regression method for spatial data repeatedly sampled over time, with heterogeneity in regression coefficients across both space and over time. In particular, we extend a varying coefficient regression model for spatial-only data to spatio-temporal data with flexible temporal patterns. We consider the detection of a potential cylindrical cluster of regression coefficients based on testing whether the regression coefficient is the same or not over the entire spatial domain for each time point. For multiple clusters, we develop a sequential identification approach. We assess the power and identification of known clusters via a simulation study. Our proposed methodology is illustrated by the analysis of a cancer mortality dataset in the Southeast of the U.S.
It is fairly common to rank different geographic units, e.g. counties in the USA, based on health indices. In a typical application, point estimates of the health indices are obtained for each county, and the indices are then simply ranked as if they were known constants. Several authors have considered optimal rank estimators under squared error loss on the rank scale as a default method for general purpose ranking, e.g. situations where ranking units across the full spectrum of performance (low, medium, high) is important. While computationally convenient, squared error loss on the rank scale may not represent the true inferential goals of rank consumers. We construct alternative loss functions based on three components: (1) the inferential goal (rank position or pairwise comparisons), (2) the scale (original, log-transformed or rank) and (3) the (positional or pairwise) loss function (0/1, squared error or absolute error). We can obtain optimal ranks for loss functions based on rank positions and nearly optimal ranks for loss functions based on pairwise comparisons paired with highest posterior density (HPD) credible intervals. We compare inferences produced by the various ranking methods, both optimal and heuristic, using low birth weight data for counties in the Midwestern United States, from 2006 to 2012.
Background: The atopic march refers to the co-expression and progression of atopic diseases in childhood, often beginning with atopic dermatitis, although children may not progress through each atopic disease. Objective: We hypothesized that future atopic disease expression is modified by atopic dermatitis phenotype, and that these differences result from underlying dysregulation of cytokine signaling. Methods: Children (n=285) were enrolled into the Childhood Origins of ASThma birth cohort and followed prospectively. Rates of atopic dermatitis, food allergy, allergic rhinitis, and asthma were assessed longitudinally from birth to 18 years of age. Associations between atopic dermatitis phenotype and food allergy, allergic rhinitis, asthma, allergic sensitization, exhaled nitric oxide, and lung function were determined. Peripheral blood mononuclear cell responses (IL-5, IL-10, IL-13, IFN-γ) to dust mite, phytohemagglutinin, Staphylococcus aureus Cowan I, and tetanus toxoid were compared among atopic dermatitis phenotypes. Results: Atopic dermatitis at year 1 was associated with an increased risk of food allergy (p=0.004). Both persistent and late-onset atopic dermatitis were associated with an increased risk of asthma (p=<0.001), rhinitis (p<0.001), elevated total IgE (p=<0.001), percentage of aeroallergens with detectable IgE (p<0.001), and elevated exhaled nitric oxide (p=0.002). Longitudinal analyses did not reveal consistent differences in PBMC responses among dermatitis phenotypes. Conclusion: Atopic dermatitis phenotype is associated with differential expression of other atopic diseases. Our findings suggest peripheral blood cytokine dysregulation is not a mechanism underlying this process, and immune dysregulation may be mediated at mucosal surfaces or in secondary lymphoid organs. Keywords: Atopic dermatitis; allergic sensitization; atopic dermatitis phenotypes; atopic march; progression of atopic disease.
Background: Farm exposures in early life reduce the risks for childhood allergic diseases and asthma. There is less information about how farm exposures relate to respiratory illnesses and mucosal immune development. Objective: We hypothesized that children raised in farm environments have a lower incidence of respiratory illnesses over the first two years of life than non-farm children. We also analyzed whether farm exposures or respiratory illnesses were related to patterns of nasal cell gene expression. Methods: The Wisconsin Infant Study Cohort included farm (n=156) and non-farm (n=155) families with children followed to age 2 years. Parents reported prenatal farm and other environmental exposures. Illness frequency and severity were assessed using illness diaries and periodic surveys. Nasopharyngeal cell gene expression in a subset of 64 children at age two years was compared to farm exposure and respiratory illness history. Results: Farm vs. non-farm children had nominally lower rates of respiratory illnesses (rate ratio 0.82 [0.69,0.97]) with a stepwise reduction in illness rates in children exposed to 0, 1, or ≥2 animal species, but these trends were non-significant in a multivariable model. Farm exposures and preceding respiratory illnesses were positively related to nasal cell gene signatures for mononuclear cells and innate and antimicrobial responses. Conclusions: Maternal and infant exposure to farms and farm animals was associated with nonsignificant trends for reduced respiratory illnesses. Nasal cell gene expression in a subset of children suggests that farm exposures and respiratory illnesses in early life are associated with distinct patterns of mucosal immune expression. Keywords: Farm; children; gene expression; nasal epithelial cells; respiratory illness; virus.
Intellectual disability (ID) commonly co-occurs in children with autism. While diagnostic criteria for ID require impairments in both cognitive and adaptive functioning, most population-based estimates of the frequency of co-occurring ID in children with autism– including studies of racial and ethnic disparities in co-occurring autism and ID – base the definition of ID solely on cognitive scores. The goal of this analysis was to examine the effect of including both cognitive and adaptive behavior criteria on estimates of co-occurring ID in a well-characterized sample of preschool children with autism. Participants included 3,264 children with research or community diagnoses of autism enrolled in the population-based Study to Explore Early Development (SEED) Phases 1-3. Based only on Mullen Scales of Early Learning (MSEL) Composite cognitive scores, 62.9% (95% confidence interval [CI]: 61.0, 64.7%) of children with autism were estimated to have co-occurring ID. After incorporating Vineland Adaptive Behavior Scales, Second Edition (VABS-II) Composite or domains criteria, co-occurring ID estimates were reduced to 38.0% (95% CI: 36.1, 39.8%) and 44.9% (95% CI: 43.0, 46.8%), respectively. After incorporating VABS-II Composite or domains criteria and adjustment for selected socioeconomic variables, the increased odds of meeting ID criteria for non-Hispanic Black children and Hispanic children relative to non-Hispanic White children observed when only MSEL criteria were used were substantially reduced, though not eliminated. This study provides evidence for the importance of considering adaptive behavior as well as socioeconomic disadvantage when describing racial and ethnic disparities in co-occurring ID in epidemiologic studies of autism.
Purpose: Alcohol consumption increases health risks for patients with cancer. The Covid-19 pandemic may have affected drinking habits for these individuals. We surveyed patients with cancer to examine whether changes in drinking habits were related to mental health or financial effects of the pandemic. Methods: From October 2020 to April 2021, adult patients (age 18-80 years at diagnosis) treated for cancer in southcentral Wisconsin were invited to complete a survey. Age-adjusted percentages for history of anxiety or depression, emotional distress, and financial impacts of Covid-19 overall and by change in alcohol consumption (non-drinker, stable, decreased, or increased) were obtained via logistic regression. Results: In total, 1,875 patients were included in the analysis (median age 64, range 19-87 years), including 9% who increased and 23% who decreased drinking. Compared to stable drinkers (32% of sample), a higher proportion of participants who increased drinking alcohol also reported anxiety or depression (45% vs. 26%), moderate to severe emotional distress (61% vs. 37%) and viewing Covid-19 as a threat to their community (67% vs. 55%). Decreased (vs stable) drinking was associated with higher prevalence of depression or anxiety diagnosis, emotional distress, and negative financial impacts of the pandemic. Compared to non-drinkers (36% of sample), participants who increased drinking were more likely to report emotional distress (61% vs. 48%). Conclusions: Patients with cancer from Wisconsin who changed their alcohol consumption during the Covid-19 pandemic were more likely to report poor mental health including anxiety, depression, and emotional distress than persons whose alcohol consumption was stable. Implications for cancer survivors: Clinicians working with cancer survivors should be aware of the link between poor mental health and increased alcohol consumption and be prepared to offer guidance or referrals to counseling, as needed.
Background: Self-identified African American or Black women (“Black women”) have persistently higher breast cancer mortality than women from other self-reported racial/ethnic groups. These mortality differences are partially explained by the effects of systemic racism on cancer risk and carecontrol processes. We quantify the relative contributions of tumor factors, screening and treatment these cancer control processes on cancer mortality disparities. Methods: We used three Cancer Intervention and Surveillance Modeling Network (CISNET) simulation models to estimate the separate contribution of demographics, incidence, subtype, screening and treatment inequity on modeled mortality among multiple birth cohorts of Black women. Model input parameters were based on national data from registries and clinical trials. Results are summarized as the mean and range across the three models. Results: Results were very similar across the three models. Racial differences in tumor subtype and stage distributions in the absence of screening accounted for a median of 23% (range across models 13-24%) and screening accounted for a median of 3% (range 3-4%) of the modeled mortality in Black women. Treatment parameters accounted for the majority of modeled mortality for Black women: median 17% (range 16-19%) for treatment initiation and median 61% (range 57-63%) for real-world effectiveness. Conclusion: Our model results suggest that policies that decrease the effects of systemic racism on treatment access could increase breast cancer equity. The findings also highlight that improvements must extend beyond policies targeting equity in treatment initiation to include high-quality treatment completion. Future modeling research will be useful to test the effects of different specific policy changes on mortality disparities.