Mortality Trends Among Adults Experiencing Homelessness in Boston, Massachusetts (2024)

Research Letter

Kirsten A.Dickins,PhD, AM, APRN, FNP-C1,2,3; Danielle R.Fine,MD, MSc3,4; Logan D.Adams,MD4; et al Nora K.Horick,MS5; ElizabethLewis,MBA6,7; Sara E.Looby,PhD, APN-BC2,3,8; Travis P.Baggett,MD, MPH3,4,6

Author Affiliations Article Information

Homelessness is a growing public health concern in the US, with an estimated 580 445 people experiencing homelessness on any given night.¹ Prior studies have documented mortality disparities with the general population^2,3; however, temporal mortality trends among people experiencing homelessness remain understudied.

This cohort study included adults (aged ≥18 years) who had an in-person encounter at Boston Health Care for the Homeless Program (BHCHP) between January 1, 2003, and December 31, 2017. We used Match*Pro, version 1.6.3 (IMS Inc), to link the BHCHP cohort with Massachusetts death occurrence files spanning 2003 to 2018 using previously described methods.^3,4 We used SAS version 9.4 (SAS Institute Inc) to conduct all data analyses from March 16, 2021, through May 12, 2022. The Mass General Brigham Human Research Committee Institutional Review Board approved this study with a waiver of informed consent because the study posed minimal risk to the individuals involved.

We used Poisson regression models to estimate annual age- and sex-standardized all-cause mortality rates in the BHCHP cohort and to compare the temporal trend in these rates with the all-cause mortality trend in the urban Northeast US. We repeated this procedure for the 6 leading causes of death observed in the BHCHP cohort (eTable in Supplement 1). Detailed methods are provided in the eMethods in Supplement 1.

The cohort consisted of 60 092 adults experiencing homelessness, yielding 520 430 person-years of follow-up. The mean (SD) age at cohort entry was 40.4 (13.1) years, 63.4% were male, 26.5% Black, 17.9% Latinx, and 43.9% White.⁴ Race and ethnicity were patient self-reported at initial care enrollment.

Overall, 7130 individuals (11.9%) died during the study period, generating a crude mortality rate of 1370 deaths per 100 000 person-years. Between 2003 and 2018, the standardized all-cause mortality rate did not significantly change in the BHCHP cohort (0.16% annual growth; 95% CI, −0.36% to 0.68%; P = .54), whereas the all-cause mortality rate decreased 1.21% annually (95% CI, −1.25% to −1.17%; P < .001) in the urban Northeast population (P < .001 for interaction) (Figure 1).

Drug overdose (23.7%), cardiovascular disease (16.5%), cancer (15.2%), psychoactive substance use disorder (6.6%), liver disease (5.7%), and chronic lower respiratory disease (2.7%) were the 6 leading causes of death. Between 2003 and 2018, drug overdose mortality increased on average 9.35% annually (95% CI, 7.90%-10.79%; P < .001), compared with an average annual increase of 7.57% (95% CI, 7.34%-7.80%; P < .001) in the urban Northeast population (P = .02 for interaction) (Figure 2). Cardiovascular and cancer mortality decreased on average 2.92% annually (95% CI, −4.18% to −1.67%; P < .001) and 2.52% annually (95% CI, −3.94% to −1.10%; P < .001), respectively, at rates that did not differ significantly from the urban Northeast population. Liver disease mortality decreased on average 3.78% annually (95% CI, −6.33% to −1.23%; P = .004) and remained stable in the urban Northeast population (P = .004 for interaction). Chronic lower respiratory disease mortality increased on average 3.65% annually (95% CI, 0.63%-6.68%; P = .02) and decreased on average 0.65% annually (95% CI, −0.88% to −0.41%; P < .001) in the urban Northeast population (P = .006 for interaction).

In this study, the disparity in all-cause mortality between this large cohort of people experiencing homelessness and the general population widened over 16 years. Drug overdose mortality was a factor in this widening gap, illustrating that public health epidemics are often magnified among people experiencing homelessness.⁵ In addition, increases in chronic lower respiratory disease mortality underscore the need for measures to reduce the high prevalence of tobacco use among people experiencing homelessness.⁶ Temporal reductions in death from cardiovascular disease, liver disease, and cancer could reflect improvements in diagnostic and treatment strategies, although competing risk from increasing drug overdose mortality is another possible explanation. Although these specific trends may not be generalizable to homeless populations in other geographic locations, these findings highlight that policies tailored to meet the distinct needs of people experiencing homelessness are essential to reducing long-standing mortality disparities in this population.

Accepted for Publication: December 21, 2022

Published Online: March 13, 2023. doi:10.1001/jamainternmed.2022.7011

Corresponding Author: Danielle R. Fine, MD, MSc, Division of General Internal Medicine, Massachusetts General Hospital, 100 Cambridge St, 16th Floor, Boston, MA 02114 (drfine@mgh.harvard.edu).

Author Contributions: Drs Dickins and Fine had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Dickins and Fine contributed equally as co–first authors.

Concept and design: Dickins, Fine, Adams, Looby, Baggett.

Acquisition, analysis, or interpretation of data: Dickins, Fine, Adams, Horick, Lewis, Baggett.

Drafting of the manuscript: Dickins, Fine, Adams.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Fine, Horick.

Obtained funding: Dickins, Looby.

Administrative, technical, or material support: Dickins, Fine, Adams, Lewis, Baggett.

Supervision: Baggett.

Conflict of Interest Disclosures: Dr Baggett reported receiving author royalties from UpToDate outside the submitted work. No other disclosures were reported.

Funding/Support: The Yvonne L. Munn Center for Nursing Research, the William F. Connell Family, Connell Postdoctoral Fellowship in Nursing Research, the National Institutes of Health (K12DA043490), and the Division of General Internal Medicine at Massachusetts General Hospital.

Role of the Funder/Sponsor: The funding sources played no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 2.

Additional Contributions: We would like to acknowledge Derek J. Smolenski, PhD, MPH (Psychological Health Center of Excellence, Defense Health Agency, US Department of Defense) for contributing to the statistical design of the study and Jessie Gaeta, MD (Boston Health Care for the Homeless Program and Boston University School of Medicine) for revising the manuscript for important intellectual content. Neither of these individuals received compensation for their contributions and both agreed to be included in the Acknowledgment section.