Hutch News

Finding the meaning of the missing

Biostatistician brings passion for collaboration and expertise in analyzing missing longitudinal data

Jan. 1, 2007
Dr. Brenda Kurland

Dr. Brenda Kurland develops and uses statistical methods to analyze longitudinal data, especially data missing due to patient dropout or death. Kurland has collaborated with Drs. Karen Syrjala, Stephanie Lee and others on quality-of-life and long-term follow-up studies of transplant patients.

Photo by Stephanie Cartier

When Dr. Brenda Kurland recalls her childhood in West Virginia, she remembers her parents, both chemists, working at home late at night. Her father would lean over her mother's shoulder as she edited a chemistry textbook, offering opinions and perspectives. It was a great model, Kurland said, and a picture of equality and collaboration she internalized and sought in her professional world. At the Hutchinson Center, she found an environment that feels like home.

Kurland, a biostatistician, joined the Clinical Research Division last year to share her expertise in the analysis of longitudinal data. Though her academic and career path has been rather circuitous, she is unequivocally satisfied with where she's landed.

"I've had positions where researchers come to you with data, and they know the result they want," Kurland said. "They just want a value, and they give you about 15 minutes to do it. They don't believe a statistician needs to understand the data, and they are very surprised and annoyed if you find a problem."

"But here, statisticians are given tools, time and respect. I've heard a lot of lip service in other places, but at the Center, researchers really mean it. It's a wonderful, collaborative environment."

'Reformed math geek'

Kurland, who calls herself a "reformed math geek," sought probability and statistics classes for fun during her teenage summer breaks when she ran out of high-school math courses to take. While pursuing undergraduate and master's degrees at Harvard, Kurland studied cognitive psychology and language and literacy development. But she found her number sense made her valuable in the school's language and psychology labs, so she kept herself busy doing statistical work.

As her education evolved, Kurland realized the cutting-edge application of statistics involved medical research. She headed to the University of Washington and completed a second master's degree and doctorate in biostatistics. At the same time, she worked as a research assistant with the Public Health Sciences Division's Southwest Oncology Group Statistical Center.

"I didn't know this is where I'd end up, but I loved the Center," Kurland said. "It's a great place and an important place to work." She admired the strength of the research and took advantage of the high standards for statistical collaboration. Drawing on both her psychology and statistical skills, Kurland pursued Alzheimer's research as a postdoctoral fellow. She worked at the National Alzheimer's Coordinating Center, the cog of 30 Alzheimer's centers and part of the UW's Epidemiology Department. But Kurland found aging research underdeveloped and with few opportunities for collaboration, so she returned to the Center last year, primarily to provide statistical support for the Seattle Cancer Care Alliance's solid-tumor program.

Kurland develops and uses statistical methods to analyze longitudinal data, especially when data are missing due to patient dropout or death. Her research is especially applicable in studies where deaths occur and are related to longitudinal response, but death is not the topic of interest. Her focus lends itself to quality-of-life and long-term follow-up studies of transplant patients. She has collaborated on studies with Drs. Karen Syrjala, Stephanie Lee and others looking at post-transplant physical, cognitive and sexual functioning and self-rated health over time.

Kurland's work involves distinguishing between missing data — such as a patient not showing up for a physical or not being reached by phone — and data cut short by a patient's death. When data are missing, the goal is to obtain the same results as if all data had been collected. When data are cut short by death, the goal is to avoid implying that the data could continue beyond death, since many bodily functions deteriorate rapidly as death approaches.

Understanding complicated variables

She also models study results that look at multiple variables at once, such as the probability of surviving a certain number of years post-transplant and being able to care for one's self. "We're usually looking at multiple observations at time," Kurland said. "It's always very complicated."

Colleagues appreciate Kurland's statistical specialization. "Brenda brings a needed talent and training to the Clinical Research Division with her knowledge of analyzing longitudinal data and managing the issues that arise when our research cohort changes in size," said Syrjala, Biobehavioral Sciences director and co-director of the Center's Survivorship Program. "In my work, I almost always want to follow people over time to see how their health and lives unfold. Surprisingly perhaps, most medical-oncology science focuses on a single endpoint such as recurrence or survival. So I've waited for someone in our division with this specialized training for over a decade. Beside the quality of her training, Brenda is fun to work with and brings a wonderful new energy to the division."

Dr. Dave Mankoff, a radiology, medicine and bioengineering professor at the UW who sees patients at the SCCA, considers Kurland a valuable colleague as he tests the ability of positron emission tomography (PET) to characterize breast cancer and monitor breast-cancer treatment. This imaging concentrates on molecular and biochemical features of tumors.

"Our studies make heavy use of biostatistics, and Brenda has been an outstanding and welcome addition to our efforts," he said. "Her expertise, combined with her understanding of clinical trials and her openness to new ideas, makes for a truly unique combination. And her patience as a teacher to the 'biostatistically naïve' has been very much appreciated. Her contributions to our work have been essential, and in several cases, we would not have been able to do the analysis and publish our work without her help."

Kurland's desire to feed her scientific hunger is palpable, as is her quest to be an essential part of the bigger picture in research. She recalls bounding into the office of Dr. Barry Storer, director of Clinical Statistics, asking, "Am I doing a good enough job? Am I contributing?" Storer said, "You've been here six weeks. Relax!"

Her goal for each day? "I really want to contribute to sound science and to develop my own statistical methods wherever they lead me," Kurland said. "I want to be able to translate those methods into actual practice. It's an evolving science. We methodologists keep plugging away and over time, things that were too fancy become commonplace."

"I have to keep asking, 'What question am I trying to answer? What's the best way to get it?' I find the question only reveals itself through collaboration; then I lock myself in my office to figure out how to do it."


For the Media

News releases >
Media coverage >
Contact us >


Story Archive


Publications

Quest
Our quarterly magazine

Annual Report
Fiscal year highlights

Science Spotlight
Monthly review of Center-authored papers


Fred Hutch News

Get updates via email.