The persistent and pervasive lack of diversity within medical research obstructs the pursuit of equitable healthcare. For too long, the medical establishment has operated with a significant blind spot, primarily focusing its research and data collection on a narrow demographic—white males. This historical status quo has resulted in critical gaps in our understanding of how diseases manifest and progress across the US population, especially women and people of color. To best advance medical science and ensure effective treatments for all, a fundamental shift in approach is necessary. Fortunately, the strategic integration of artificial intelligence (AI) offers a promising pathway forward.
Despite commendable efforts over the past few decades to prioritize diversity in clinical trials, such as the 1993 National Institutes of Health (NIH) Revitalization Act requiring inclusion of women and minorities in federally funded research, women and people of color remain consistently underrepresented in clinical trials. For example, within cardiovascular research, women and non-white populations are still routinely excluded despite cardiovascular disease being the number one killer for women. This detriment in representation not only hinders the development of effective medical interventions specialized to these groups but also impedes the innovation of new treatments. According to a commentary from the Penn Capital-Star, the promising concept of ‘precision medicine’ aims to provide individualized medical care “that is tailored to their own biology,” thereby moving away from the potentially dangerous generalization of clinical research findings across the entire U.S. population. Per Scientific American, certain medications may be ineffectxive or even harmful in specific genetic groups. For instance, carbamazepine, a drug commonly prescribed for epilepsy, can trigger a serious skin reaction in individuals with a specific gene variant found among people of Asian descent. Precision medicine holds the key to more effective treatments because it is designed to apply to a wide range of diverse bodies, rather than an abstract, generic human that fails to represent the entirety of the U.S. population. However, the successful implementation of precision medicine hinges on the availability of representative data, which is currently missing from many clinical trials and pre-existing medical research. This perpetuates a cycle: medical challenges and disparities continue to disproportionately affect underrepresented populations in medical research, with progress often frustratingly slow.
What are the implications of this issue for Andover’s future leaders and innovators in the medical world? The time has come to strategically harness the power of new technologies to accelerate medical research in a way that truly reflects the diverse demographics that comprise the United States. Traditionally, ensuring the safety and efficacy of new medications relies on linear and sequential clinical trials, considered the ‘gold standard,’ as noted in research from the National Institutes of Health. However, these trials are often hampered by challenges such as under-enrollment, participants who drop out mid-study, unforeseen side effects, and contradictory data. As clinical trial phases advance, the required sample size of patients also increases. Inefficient patient screening and recruitment methods, along with difficulties in monitoring participants’ adherence to treatment protocols, can contribute to high trial failure rates and escalating research and development expenses. While the successful identification of appropriate participants significantly boosts a trial’s potential effectiveness, inefficient or resource-intensive recruitment efforts can directly lead to a failure of a study. However, this risk of failure serves as a powerful motivator for using AI throughout the earliest phases of clinical trials. AI possesses the ability to sift through vast quantities of data to identify specific subsets of patients who are most likely to respond positively to a particular clinical study. Furthermore, analyzing social media data can help pinpoint geographical hotspots for specific diseases or disorders, thereby enabling a more focused recruitment effort. By efficiently assessing hospital medical records and proactively notifying both physicians and potentially eligible patients about relevant clinical trial opportunities, AI has the potential to significantly accelerate the process of finding suitable participants.
However, AI is not immune to the inherent bias present in existing research datasets. Given that historically underrepresented groups constitute a significant portion of the population, AI systems trained predominantly on these skewed datasets will inevitably possess an insufficient understanding of the nuances of health and disease within these populations. This can lead to biased findings that may not be applicable or effective for these groups or for patient pools with different demographic compositions.
Understanding this critical limitation, the onus falls on us, as future leaders and innovators, to champion a more inclusive and effective model in the medical workforce; therefore, it is crucial for us to be aware of these disparities and the transformative potential of AI. A hybrid model consisting of both AI and humans will be necessary in order to develop drugs and testing that will truly reflect the diversity of the United States. The potential of AI is boundless, and with determined, forward-thinking researchers leading the way, we can realize its promise of more accurate diagnoses and effective treatments. By advocating for and supporting responsible and ethical AI development in medical research, we can contribute to a future where healthcare truly serves “Youth from Every Quarter” and beyond, leading to a more equitable and healthier society for everyone.
