HARVARD GAZETTE ARCHIVES

Nancy Krieger, one of the event's organizers, said statistical data on health disparities can be a double-edged sword because statistics can be used to discriminate against minority groups. (Staff photos Kris Snibbe/Harvard News Office)

School of Public Health Dean Barry Bloom said the information conveyed by health statistics is critical in identifying both disparities in health and in designing programs to address them.

Godfrey Woelk of the University of Zimbabwe said the main challenge of gathering health statistics is to make disparities visible to the general public so that action can be taken.

In today's globalizing world, Bloom said, an increasing gap is found between those who "have" and those who "have not," and that gap extends to health care. While universities don't have the power to address these inequalities directly, they can stimulate discussions and use their convening power to bring together people who may be able to prompt action.

"It seems pretty clear that disparities within and between countries are increasing," Bloom said.

Health information in the form of statistics is important, Bloom said, because such data describe problems that otherwise are just anecdotal.

"I truly believe that without credible, scientific evidence, everything else is just opinion," Bloom said.

HSPH Dean Bloom said statistical information is critical in identifying disparities in health.

Nancy Krieger, associate professor of society, human development, and health, and one of the event's organizers, said gathering statistical data on health disparities can be a double-edged sword because statistics can be used to discriminate against minority groups as well as to help them.

At the other extreme, though, is the mind-set that without data, a problem doesn't exist.

"For health disparities to matter, they have to be counted," Krieger said.

The most delicate issue, she said, is how to design health data collection systems that gather the needed information while guarding against misuse.

The symposium featured speakers from around the globe, including France, Brazil, and Zimbabwe.

Krieger delivered a presentation on the English system for gathering health statistics, with information provided by John Fox, director of customer and stakeholder engagement of the Health and Social Care Information Center in the United Kingdom. (Fox had to cancel his trip to Harvard at the last minute.)

The United Kingdom has taken the need for health statistics seriously, establishing the Health and Social Care Information Center in April 2005. The center will seek opportunities to strengthen data collection efforts and match those efforts with changes in policies and the health care agenda.

In France, health statistical data often leaves out information on ethnicity, creating a void of information on whether disparities exist among the nation's ethnic groups, according to François Heran, director of the Institut National D'Etudes Demographiques in Paris.

Heran said the failure to gather statistical information by ethnic groups stems from an administrative law blocking the gathering of such information in an effort to stem discrimination. The law, which prohibits the gathering of ethnic information for governmental administrative purposes, does not bar the gathering of ethnic information for statistical purposes, but Heran said many researchers act as if it does.

"The block is more intellectual than legal," Heran said.

French studies tend to focus on socio-economic information such as poverty and economic insecurity rather than race and ethnicity.

In the United States, such data is collected, but University of California, Los Angeles, Professor of Clinical Psychology Vickie Mays said the system is far from ideal.

American health statistics have many gaps and inconsistencies, she said, ranging from changing self-identification reflecting the presence of a large number of multiracial and multiethnic people in the population to the need for better training for people who record racial information.

"Racial identity and race identification can be fluid over time and across circumstances, especially among people with heterogeneous ancestries," Mays said.

Mays, who chaired a government advisory committee on health statistics, said racial and ethnic information often reflects the preconceptions of the person recording it.

"Who I am depends on who you are," Mays said. "We have to start teaching that a statistical decision carries ethical considerations. [One must consider] who you chose as a reference group and who that choice disadvantages or advantages."

A major gap in the system of health data collection is the lack of information about the large numbers of men from particular racial and ethnic groups in prisons across the United States, Mays said.

Godfrey Woelk, associate professor of the Department of Community Medicine at the College of Health Sciences of the University of Zimbabwe, said the main challenge of gathering health statistics is to make disparities visible to the general public so that action can be taken. What data is gathered and how it is used is critical but it also needs to be flexible to take into account different situations.

In Zimbabwe, for example, he said that data comparing the rich with the poor would leave out an important trend in recent years that would show the increasing impoverishment of the middle class.

"What indicators to use, this is a crucial decision," Woelk said. "Theory informed by values helps us decide who should get interventions."