The increasing diversity of the population of the United States presents many challenges to conducting health research that is representative and informative. In their article on the importance of subpopulation research in cancer control, Srinivasan and colleagues1 said that “it is critical to ensure that all segments of the U.S. population benefit from [health] research and from the latest technologic advances in cancer care services and delivery.” They expressed concern regarding underrepresented groups and the issue of “small data,” which arises when “the size, dispersion, or accessibility of the population of interest makes it difficult to obtain adequate sample sizes to test specific research questions” (Srinivasan et al., 2015, p. 1).
Dispersion and accessibility issues can increase logistical costs; populations for which it is difficult to obtain adequate sample size are also likely to be expensive to study. Hence, even if it is technically feasible to study a small population, it may not be easy to obtain the funding to do so. This argues for increased efforts to document the needs, potential benefits, and methods for enhancing the efficiency of studying small and hard-to-reach populations.
In some situations, there may be no good sampling frames for a subgroup because there is no consensus among researchers regarding how to
1 Shobha Srinivasan and her coauthors are representatives of the National Cancer Institute, one of the primary sponsors of this workshop. Authors are cosponsors of this workshop. Srinivasan, S., Moser, R.P., Willis, G., Riley, W., Alexander, M., Berrigan, D., and Kobrin, S. (2015). Small is essential: Importance of subpopulation research in cancer control. American Journal of Public Health, 105(S3):S371-S373.
define or identify the population in question or because people are reluctant to self-identify as a member of the subgroup in administrative records or on national surveys because of perceived stigma. These “so-called” hidden populations are small by the above definition but raise more fundamental questions in health research and require additional data source work to measure their health disparities.
Researchers at the National Institutes of Health (NIH) and especially within the National Cancer Institute (NCI) have been contemplating the steps needed to improve health research in small populations for a number of years. Srinivasan et al. (2015) is one milestone of their deliberations. Their paper describes many of the issues in small or subpopulation research, noting that solutions will require interdisciplinary collaborations. In describing the challenges associated with finding the right people to study, the paper addresses issues related to identifying and collecting information from small, hard-to-reach, and hidden populations and issues related to recruitment and retention—key to most studies, but critical for small populations. It also documents the need to develop new intervention designs for use with small populations/samples, and the need for new or enhanced analytical techniques.
In the spring of 2017, NIH—through a collaboration among NCI, the National Institute on Minority Health and Disparities, and the Office of Behavioral and Social Sciences Research—requested that the Committee on National Statistics (CNSTAT), in collaboration with the Health and Medicine Division, of the National Academies of Sciences, Engineering, and Medicine convene a steering committee to organize a workshop to address the issues associated with improving health research of small populations. The statement of task is shown in Box 1-1. Support for this workshop was also provided, in part, by the Robert Wood Johnson Foundation. The views expressed here do not necessarily reflect the views of the sponsors.
PLANNING THE WORKSHOP
CNSTAT, in collaboration with the Board on Health Care Services, recruited and gained approval for the steering committee: five individuals with expertise in research design, statistical analysis, epidemiological and surveillance analysis, research methodologies, and approaches for recruitment, retention, and participation, all with a focus on small and hard-to-reach populations (see Appendix B for biographical sketches). The steering committee organized the workshop around the various phases of health research studies: using available data, collecting new data, designing experimental studies, recruitment and retention, and analysis of the data.
The Workshop on Improving Health Research on Small Populations took place at the National Academies of Sciences, Engineering, and Medi-
cine in Washington, D.C., January 18-19, 2018. The audience included not only NCI leadership and technical staff, but also the broader health research and statistical communities who have an interest in funding, conducting, and evaluating research on small population groups. It was important to the sponsors to include diverse perspectives in the workshop, including input from those with broad academic technical expertise and from researchers with practical field experience studying small and rare populations.
MOTIVATION AND OBJECTIVES FOR THE WORKSHOP
Robert Croyle (NCI), in introductory remarks at the workshop, noted the issue of health research for small populations is relevant to many disciplines and many organizations beyond NCI and the other workshop funders. He invited the audience to raise cross-cutting methodological scientific issues in order to support broader work in social behavioral sciences, public health,
medicine, and other disciplines, noting that one of the sponsors’ goals was receiving feedback from a variety of different content domains.
Croyle said that for decades, funders gave many reasons why they did not fund more research relative to small populations, which is perhaps why not much is known about health issues relevant to American Indian and Alaska Native populations, sexual and gender minority populations, and people who live in extreme rural areas or more than 150 miles from an NCI-designated cancer center. However, with evolutions in statistical methods, sampling, and new data sources, he said it is time to support this research.
Croyle noted the traditional solution has often been to “increase the denominator”—deepen the sampling pool. For example, if NCI tripled the budget for an epidemiological cohort studying a rare cancer to $1 billion per year, the researcher could guarantee an adequate sample size of that rare cancer. However, he said, this method is not practical. In addition to studying rare cancers or dividing groups by race, income, or geography, researchers may also want to target members of a rare population who have been exposed to a carcinogen. Each time another molecular subtype of cancer is identified, another subpopulation is also identified that merits its own clinical trial, drug, survivorship care plan, and psychosocial support strategy. This creates challenges with Food and Drug Administration approvals.
Another challenge, Croyle said, is facilitating discussions across disciplines that do not often communicate with one another. For example, NCI is working with the geography community to conduct geospatial analysis of population health. Statistical sampling and methodology is another area in which different disciplines have their own traditional approaches to addressing this problem. In many cases, the approaches are shared among disciplines, but Croyle said other knowledge should likewise be shared.
Croyle talked about how market research consultants, large corporations, and major political parties and campaigns often microtarget very small subpopulations—blue collar, low income, or other groups. They can also draw inferences by linking data from many different sources and then making broader generalizations and conclusions about strategies to approach those populations. He expressed hope that the health research community will follow suit and become better about targeting and engaging small populations.
STRUCTURE OF THIS PROCEEDINGS
After the welcoming remarks, there were six main sessions and a concluding panel discussion among steering committee members. The remainder of this proceedings reflects the six main workshop sessions and
the concluding panel discussion. Chapter 2 provides examples that illustrate progress in studying small subpopulations and where work remains. Chapter 3 addresses the challenges of using existing data to study small populations. Chapter 4 covers a variety of sampling methods that can be applied to small population research, and Chapter 5 highlights applicable design, intervention, and implementation strategies. Chapter 6 explores recruitment and retention strategies for small populations, and Chapter 7 identifies survey design and data analysis techniques. Chapter 8 reports on the final panel discussion. The workshop agenda is in Appendix A, and Appendix B includes biographical sketches of steering committee members and invited speakers.
This publication was prepared by two rapporteurs as a factual summary of what occurred at the workshop. The steering committee’s role was limited to planning and convening the workshop. The views contained in the report are those of individual workshop participants and do not necessarily represent the views of nonparticipants, other workshop participants, the steering committee, the sponsors, or the National Academies of Sciences, Engineering, and Medicine.
This page intentionally left blank.