# Winning the numbers game

## Winning the numbers game McGill University

User Tools (skip):

Sister Sites: McGill website | myMcGill

November 9, 2006 - Volume 39 Number 06
| Help
Page Options (skip): Larger
Home > McGill Reporter > Volume 39: 2006-2007 > November 9, 2006 > Winning the numbers game

# Winning the numbers game

## From dementia to hedge funds, survival analysis has many uses

From studying our aging population to analyzing million-dollar investments, statisticians apply their expertise to many fields. Properly designed studies, managed jointly by researchers and statisticians, can greatly improve the quality of the research. Statisticians often help design studies as well as perform the final statistical analysis and interpret results.

iStock photo

Statisticians use several mathematical methods to analyze data. One method, survival analysis, is used to analyze data on the times tied to a specific event, for example, the times it took for a medication to have an effect on each member of a group of patients. Survival analysis has applications to many fields of research, including medicine, engineering, public health, anthropology and economics. Partnerships between investigators in these various fields and statisticians are common.

Masoud Asgharian and David Wolfson, professors in the Department of Mathematics and Statistics, collaborated with researchers involved in the Canadian Study of Health and Aging. In this large epidemiological study of dementia, which includes Alzheimer's disease, Asgharian, Wolfson and their co-authors addressed how long people with dementia survive following the onset of the disease. Using standard survival analysis, they found that the median lifespan of a patient with dementia was about seven years after the onset of the disease. This was consistent with previous studies. But the researchers realized that because of how this study, as well as all the other studies, had been designed, those who lived longer with dementia were more likely to be included than people who did not live as long with dementia. When Asgharian and Wolfson applied the appropriate mathematical modeling to remove the bias, they determined that the median survival after the onset of dementia was actually only about four years. Their non-standard statistical methods had revealed that the life expectancy, from disease onset, of a person with dementia was much lower than what was previously accepted in the medical community.

Asgharian is also working with financial experts interested in the lifetime of large global hedge funds. These funds are unregulated and require each investor to invest a minimum of \$1 million. Hedge funds are high-risk investments and financial experts are interested in discovering why, when they fail, they fail. Asgharian and co-authors used survival analysis and examined seven features of these funds to determine which features have an effect on their success. They found that only two features, the monthly returns and the amount of assets, have an effect on how long the funds last. Conversely, they discovered that hedge funds based on large amounts of borrowed money, known as high leverage, last as long as funds that use less borrowed money, known as low leverage. This was an interesting discovery since the failure of these high risk funds has often been attributed to high leverage or the fact that they were supported by large amounts of borrowed money.

From the survival analysis of investment funds to improving the screening of new drugs, there are several ongoing collaborations between the Department of Mathematics and Statistics and other research units at McGill. These partnerships are mutually beneficial for the investigators and the statisticians who continually learn about different research areas. Both Asgharian and Wolfson welcome collaborations with their colleagues at McGill because they enjoy "working with real world problems while at the same time developing mathematical methodology."

The Department of Mathematics and Statistics offers a free Statistical Consulting Service for faculty, staff and graduate students. They may be consulted for assistance with the design of studies and statistical analysis. The principal consultant, Jose Correa, can be contacted at Burnside Hall 1233, 514-398-3843, jcorrea@math.mcgill.ca.

WARM-SPARK (Writing About Research at McGill-Students Promoting Awareness of Research Knowledge) is a program supported by the VP Research Office, Associate Vice Principal (Communications), the faculties of Agricultural and Environmental Sciences, Arts, Engineering, Medicine and Science. See www.spark.mcgill.ca.

Search