M. G. Janeway1, F. Pereira2, M. R. Nofal2, L. G. Paredes2, J. Galindo2, M. Poulson1, M. Neufeld1, T. A. Dechert1, S. E. Sanchez1 1Boston University,Department Of Surgery,Boston, MA, USA 2Boston University,School Of Medicine,Boston, MA, USA
Introduction: Homelessness is associated with increased overall mortality compared to the general population. Little is known about surgical need, outcomes, and follow-up in this population. Our study examined the epidemiology of surgical disease in the homeless population and the independent effect of homelessness on surgical outcomes.
Methods: In this retrospective observational cohort study, we examined data on homeless and housed patients 18-89 years admitted to surgical services at our urban, safety-net hospital from 2014 to 2017. Outcomes of interest included surgical intervention, postoperative complications, post-admission emergency department (ED) visits, and 30-day unplanned readmission. We also examined length of stay (LOS), discharge disposition, social work services, and outpatient follow-up. Patients were matched 2:1 on age, gender, and comorbidity score. Univariate analyses were used to examine associations between each independent variableand the outcomes of interest.
Results: Of the 372 patients meeting our inclusion criteria, 124 were homeless and 248 were housed. Mean age at admission was 49.1 ± 13.3 years. Of those experiencing homelessness, the majority were white (54%) and male (72%). Homeless individuals were significantly less likely to undergo surgery during their admission compared to housed individuals (42% vs 59%, p=0.002). Mean LOS was similar for homeless and housed individuals (5.22 vs 5.98 days). There was no significant difference in postoperative complications or 30-day readmission between the groups. Homeless patients, however, were significantly more likely to have a 30-day ED visit compared to housed individuals (32% vs 9%, p<0.001) and significantly less likely to make their follow-up appointments (42% vs 72%, p<0.001). While the majority of both homeless and housed patients were discharged to home/self-care, there were significant differences in discharge disposition between the two groups, with homeless individuals being more likely to leave against medical advice (13% vs 2%, p<0.001) or go to skilled nursing facilities (9% vs 3%, p<0.001). Services to address housing were provided to 24.8% of homeless patients during their visit.
Conclusion: Homeless patients matched for age, gender, and comorbidity score at our institution are less likely than housed patients to receive surgical intervention, however, they have comparable rates of postoperative complications and 30-day unplanned readmissions. This is likely secondary to systems in place at our hospital that work to improve health outcomes in this vulnerable population. Despite these services, our homeless patients have a higher ED utilization rate and are less likely to attend their follow-up appointments. Additional attention should be given to encouraging follow up in an effort to reduce ED visits. Further investigation into potential reasons for lower rates of surgical intervention is warranted.