We need to apply a sex, gender, race and diverse populations lens now to today’s health innovations. The male dominance in the AI and big data sectors, as well as the existing historical bias in health data research, is creating a perfect storm of biased outputs.
If we are not careful, we will again build a system that benefits only a portion of the population.
When I was the scientific director for the Canadian Institutes of Health Research’s (CIHR) Institute of Gender and Health, my team and I developed standards and curriculum to address the importance of considering sex and gender diversity in health research. We were working to correct years of research that ignored important differences in sex and gender.
In Canada and the United States we were continuing to create treatments and technologies that had primarily been tested on men (mostly white men) and we were ignoring the powerful biological effects of sex and the social effects of gender.
Why did this matter? Because of women’s smaller body sizes, their hormonal differences, and changes to metabolism due to contraceptives—these are all variables that affect how a woman will react to a course of treatment. Globally, we were approving treatments and putting women at risk.
For example, a study conducted by Ridker et al. showed that as opposed to men, Aspirin only lowered women’s risk of stroke but had no effect on the risk of heart attacks or death. In another study, the liver enzyme CY P3A4 showed a slowed drug clearance rate in women when compared to men, resulting in decreased effectiveness of antidepressants, anxiolytics, painkillers and anticonvulsive drugs.
Today in Canada, we are much better at ensuring women are not at increased risk due to uptake of policies at the federal level. However, we are still dealing with the impacts of centuries of male-dominated health research and male-biased health datasets.
Technological advances in genomic sequencing and medical imaging are helping us better understand the differences that exist among cancer patients with the same diagnosed disease. We see news stories weekly on advances in precision medicine and medical imaging in Canada.
However, equitable health care requires both ethical data collection practices and the inclusion of individuals from all backgrounds.
The reality is stark. According to another report, we see small amounts of clinical data from typically wealthy, middle-aged white males being used to inform treatment for entire populations.
This feeds the “junk in/junk out” concept that threatens our big data approaches in precision medicine. If you do not have quality representative data, your AI outcomes will be biased. The result: precision health is not as effective as it should be.
The technology sector and many areas of health research remain dominated by men. In some disciplines the disparity is as high as 87% male. We have already seen the outcomes when design in technology does not consider diversity.
Early speech recognition machines were made by all-male teams who calibrated them for their voices. The technology failed because it could not be used by women in the workplace.
The team developing the air-bag was all-male and they used the height and weight chart for the standard man. The unintended and tragic consequence was that women and children were killed when those early airbags were deployed.
We now need serious policy work to overcome these challenges. Managing policy and implementation guidelines for health innovation after the fact could result in dangerous outcomes for women and other under-represented groups.
At Simon Fraser University, we have a number of research endeavours that are overcoming barriers for data collection for harder to reach populations, including sex workers and more transient populations.
We are using strategic partnerships to access and develop unique public health data frameworks that will address a whole population approach to managing the opioid crisis.
We are creating diverse, multidisciplinary teams to look at advances in precision medicine in oncology specifically for children, who are excluded from drug testing and development. It is through partnerships and engagement that we can expand the communities we engage in health research.
Given health and technical research is funded by our tax dollars, it should serve us all and not just the privileged few. This must happen through intentional inclusion. We need to build protections, guidelines and policies to ensure health innovation in Canada supports our core value of access for all.
Dr. Joy Johnson is VP, Research and International, at Simon Fraser University.