Implementing a Screening Tool for Homelessness at LVHN: A Pilot Project
A Subset Analysis on Gender Differences
Mentor: Dr. Marna Greenberg, DO, MPH
Housing stability and personal health are closely correlated. A person’s health suffers if he or she identifies as homeless or at risk for being homeless due to limited access to medical care. Homeless persons frequently utilize the emergency departments (EDs) of hospitals. Dedicated to caring for all patients, Lehigh Valley Health Network (LVHN) developed the study titled Implementing a Screening Tool for Homelessness at LVHN: A Pilot Project to determine the prevalence for homelessness or at risk for homelessness in LVHN ED patients. Using a simple survey, patients in LVHN EDs were screened for homelessness or at risk for homelessness. This specific study focused on a subset analysis of gender and homelessness. Knowledge of gender differences among homeless patients allows for healthcare providers to be better prepared when caring for homeless patients due to the fact that different genders have different medical needs. It was hypothesized that there would be a significant difference in gender among homeless patients identified in the LVHN ED. While the results failed to support the hypothesis, the implications of gender differences among homeless patients remain an important healthcare topic to be explored.
Homelessness or at risk for being homeless serve as factors that negatively affect personal health. The National Coalition for Homeless (NCH) explains, "poor health is both a cause and a result of homelessness" (Health Care and Homelessness, 2009). With limited access to medical care, homeless patients often are dependent upon EDs for medical care (Ritchey, La Gory, & Mullis, 1991). Therefore, EDs often act as a primary healthcare provider to homeless patients.
The relationship between homelessness and healthcare provoked LVHN to take several initiatives. In order to assist the homeless population of the Lehigh Valley, LVHN developed the Street Medicine Team. This team focuses on providing medical care to the homeless in the Lehigh Valley. Additionally, LVHN developed the study titled Implementing a Screening Tool for Homelessness at LVHN: A Pilot Project to determine the prevalence for homelessness or at risk for homelessness in the LVHN ED population. A subset analysis on the collected was performed with a specific focus on gender. The objective of the subset analysis on gender is to determine the relationship between gender and homelessness in the LVHN ED.
According to the NCH, men are more likely than women to become homeless (Who is Homeless?, 2009). However, homeless women report more heath related symptoms than homeless men (Ritchey, La Gory, & Mullis, 1991). These conflicting reports make it difficult to determine the expected gender ratio of homeless patients. Improving the quality of medical care provided to homeless patients through the Street Medicine Program necessitates knowledge of gender ratio expectations. Gender plays a prominent role in healthcare as different genders experience different medical issues. The reproductive systems of males and females exemplify one aspect of medical care that differs between genders (Doyal, 2001). Biological aspects are not the only factors that generate different medical needs. Social construction of gender further influences the medical needs of men and women. To explain, in certain communities a male’s desire to be perceived as masculine may encourage him to put his health at risk (Doyal, 2001). Similarly, gender inequalities that have resulted from social construction of gender negatively affect women’s health (Doyal, 2001). Thus, a complex relationship arises between homelessness, health, and the biological and social factors of gender. It is hypothesized that there will be a significant difference in gender within the LVHN ED homeless population. Through examining whether gender differences exist in the LVHN ED, the LVHN Street Medicine Program will develop a better understanding of gender ratios expected in the homeless patient population, which will overall improve the medical care provided to Lehigh Valley homeless patients.
A homelessness screening tool consisting of a 5 question survey was developed and subsequently approved by LVHN IRB. The survey is based on the U.S. Department of Housing and Urban Development (HUD) definition of homelessness, which fortifies its ability to identify patients that are homeless or patients that are at risk for homelessness. The time, location, and ED section (pod) where the survey was to be given were scheduled to eliminate screening bias and to ensure a wide-ranging sample of patient demographics. The surveying times were either 7:00 a.m. to 4:00 p.m. or 4:00 p.m. to 1:00 a.m. These shifts aligned with the shifts of physicians; this design enabled students to work alongside of the same physician for the entire surveying shift. With the same physician accessible throughout the surveying shift, the students could easily verify if a patient had capacity to participate in the survey if necessary. The surveying locations included the EDs at Lehigh Valley Hospital Cedar Crest, Muhlenberg, and 17th and Chew Street. The survey was given verbally, and the results were recorded electronically using the Homelessness Screening Tool webpage. USF students, Research Scholar students, and Bridging the Gap students administered the survey to patients in accordance with the inclusion and exclusion criteria. The inclusion criteria states that patients must be 18 years old or older; patients must speak English; patients must have capacity to answer survey questions; patients must not be critically ill, and patients are willing to participate. The exclusion criteria excludes patients that are less than 18 years old; patients who do not speak English; patients who do not have capacity to answer survey questions; patients who are critically ill, and patients who are unwilling to participate (Implementing a Screening Tool for Homelessness at LVHN: A Pilot Project Protocol; See Attachment 1). The students administering the survey emphasized to the patients that completion of the survey is completely voluntary, and the survey will not affect the care that the patients are receiving. A Patient Screening Log was kept at each ED location. The Patient Screening Log allowed for students to record additional information that the Homelessness Screening Tool webpage could not record. The Patient Screening Log recorded the date and time that the patients were put into their ED rooms, which accounted for the patients who were admitted between 1:00 a.m. and 7:00 a.m. The Patient Screening Log also recorded
Patient gender, patient age, any reasons a patient did not participate in the survey, and the initials of the student that administered the survey. Lastly, the log allowed for students to indicate if a consult for the Street Medicine Program was the result of a survey. Research Scholar students conducted Patient Screening Log data entry using Excel. The Homelessness Screening Tool data as well as the Patient Screening Log data were analyzed together using either a two-tailed Fischer’s Exact test or Chi-Square test. Since this study is a subset analysis focused on gender, only the results pertaining to gender were applicable in this study. The gender analysis results were compared to the p value of 0.05 to determine if a significant statistical difference existed.
Table 1. The gender differences in positive survey question responses. N=113.
Figure 1. The gender differences in the number of total patients who participated in the survey. N=1615.
Figure 2. The gender differences in the number of total patients who screened positive for homelessness. N=113.
Figure 1 displays that 42% of the participants were male and 58% were female out of 1615 total participants. Therefore, more women participated in the survey than men. The p value of 0.02 indicates that the gender differences among participating patients is statistically significant. Out of 1615 total participants, 113 patients identified as homeless. Figure 2 illustrates that out of the 113 patients that identified as homeless, 46% were male and 54% were female. While the percentages representing homeless gender differences correspond to the percentages representing total participant gender differences, these two data sets are not as similar as they appear. The p value representing homeless gender differences (Figure 2) is 0.29, proving that there is no significant difference in gender of LVHN ED homeless patients.
Table 1 represents the gender differences among the responses to each survey question. A separate analysis was conducted for each question. Although Table 1 shows that males and females responded differently to the survey questions, the responses did not differ enough to be considered significant. There were no p values that indicated any significant statistical difference in question responses. Table 1 further confirms that there is no significant gender difference among the LVHN ED homeless patient population.
The conclusion that there is no significant gender difference among homeless patients of the LVHN ED fails to support the statement that men are more likely than women to be homeless (Who is Homeless?, 2009). Although this statement supported the hypothesis by suggesting that a significant difference would exist, the results did not support the hypothesis. Therefore, it was incorrect to use the statement from the NCH to formulate a hypothesis for the LVHN ED homeless population, which is significantly smaller than the national homeless population. However, the lack of a significant gender difference among LVHN ED homeless patients remains unexpected based on the literature review of the article titled “Gender differences in health risks and physical symptoms among the homeless” (Ritchey et al., 1991). This article suggests that homeless women are more likely to report health related symptoms than homeless men (Ritchey et al., 1991). The results confirmed that homeless women utilized LVHN EDs more than homeless men. While this correlates with the hypothesis that there would be a significant gender difference among the LVHN ED homeless population, it did not fully support the hypothesis, due to the fact that the gender differences were not significant.
Despite the fact that the results failed to confirm a significant gender difference among the LVHN ED homeless patients, the implications of gender differences among homeless patients remain an important topic. A detectable significant difference would allow healthcare providers in the Street Medicine Program to be prepared for gender related medical issues. As previously stated, both biological and social factors affect gender while influencing health (Doyal, 2001). Therefore, it is essential for healthcare providers to develop an awareness of the interaction between gender and health among homeless patients. This would ultimately help LVHN to sustain the goal of better cost, better health, and better care.
There were several limitations to this study. A major limitation was that the survey excluded non-English speaking patients and psychiatric patients. Non-English speaking patients are frequently encountered in LVHN EDs and could potentially provide data that would alter the current results. Psychiatric patient exclusion also eliminates valuable data from the data collection due to the fact that mental health and housing stability are correlated. Additionally, the summer screening season remains incomplete. Several weeks of data have yet to be collected and analyzed for the complete Summer 2015 screening. The screening will also be repeated again in Winter 2016 to determine if data analysis results vary due to season. Lastly, the gender options available in the homeless screening tool are not inclusive. The options for gender include male and female, which automatically excludes patients who identify as intersex or transgender. With gender being a dynamic topic in today’s society, it may be necessary in the future to modify the gender choices, allowing the survey to be representative of all genders.
I would like to thank my mentor Dr. Marna Greenberg for the opportunity to work with the Street Medicine Team this summer, and the opportunity to contribute to this meaningful data collection.
Doyal, L. (2001). Sex, gender, and health: the need for a new approach. BMJ : British Medical
Journal, 323(7320), 1061–1063.
Health Care and Homelessness. (2009) Retrieved from National Coalition for the Homeless Website: http://www.nationalhomeless.org/factsheets/health.html
Kushel, M. B., Vittinghoff, E., & Haas, J. S. (2001). Factors associated with the health care u tilization of homeless persons. Jama, 285(2), 200-206.
Ritchey, F. J., La Gory, M., & Mullis, J. (1991). Gender differences in health risks and physical symptoms among the homeless. Journal of Health and Social Behavior, 33-48.
Who is Homeless? (2009). Retrieved from National Coalition for the Homeless Website:
Implementing a Screening Tool for Homelessness at LVHN:
A Pilot Project
Social determinants of health have become recognized as some of the most influential factors affecting personal wellness—and few material possessions are as significant in this regard as housing. According to the National Health Care for the Homeless Council, 1.5 million Americans experience homelessness each year, with over 600,000 experiencing homelessness or housing instability on any given night (NHCHC, n.d.). Approximately one-third of the homeless are families with children, with another three percent representing unaccompanied minors (Homelessness in Osteopathic Pre-Doctoral Education [Project HOPE], 2014). The Lehigh Valley is not immune to these national trends, with an estimated 10,500 individuals qualifying as “homeless” within Lehigh and Northampton Counties based on local shelter census data—a figure that is on the rise.
The impact of housing situation on health outcomes cannot be overstated, as evidenced by an average life expectancy of 45-47 years within the homeless population (as compared to 78 for non-homeless individuals), and a mortality rate for those experiencing chronic homelessness of four to nine times higher than for the general population (NHCHC, n.d.; Wilkins & Elliot, 2010). The need to provide quality primary care for the homeless population is great given the research that has found the high level of disease burden and health care utilization among people who are homeless, as well as data that clearly support the critical role of primary care with regard to wellness promotion and disease management (World Health Organization [WHO], 2008). Numerous studies from major US cities demonstrate that large numbers of the homeless population access hospital ER’s as a place for care on a regular basis (Kushel, Perry, Bangsberg, Clark, & Moss, 2002; Morris & Gordon, 2006; Bharel et al., 2013). When admitted, the average Length of Stay (LOS) for these individuals was estimated at an extra 4 days regardless of diagnosis, and the 30 day readmission rate was ten times greater than a citizen living in poverty but housed (Hwang, Weaver, Aubry, & Hoch, 2011). In a study using data from the ER components of the 2005 and 2006 National Hospital Ambulatory Care Survey (NHAMCS-ED), Ku, Scott, Kertesz, and Pitts (2010) found that homeless persons who were seen at an ER were three times more likely to have been seen in the same ER during the previous 3 days than non-homeless persons; and more than twice as likely to be seen at an ER within one week following a hospitalization. The NHCHC (n.d.) estimates that 80% of ER visits by homeless individuals can be prevented by adequate primary care—a staggering figure when considered within the context of healthcare utilization, outcomes, and cost.
In light of these realities—extracted from national data, yet not dissimilar to the Lehigh Valley experience—an LVHN Street Medicine Team was created, utilizing an integrative, interdisciplinary mobile team approach for the care of homeless individuals and families in the Lehigh Valley. Basic medical and preventive services are provided free of charge, for acute and chronic illnesses, to people who are homeless at multiple points of service. This is administered through LVHN’s 17th Street Community Practices and coordinated by its Mark J. Young Community Health and Wellness Center. Patients accurately identified as homeless during an Emergency Room visit or inpatient hospitalization are referred to a Street Medicine Consult Service, to provide safe discharge planning and rapid outpatient follow up to prevent readmissions.
Still, LVHN is currently unable to report actual utilization rates or costs of caring for patients who are homeless due to the challenges in documenting homelessness and housing instability. Cost estimates of caring for this population yield relatively crude figures, as they often rely on data related to self-pay patients, which include all individuals who do not have insurance; and indeed, while the majority of homeless individuals have no health coverage, some people who are homeless do have government provided health insurance. In addition, the demographics and healthcare utilization habits of housed, uninsured patients are different from those of homeless patients.
Thus, annual cost totals become impossible to generate due to this incomplete data—although it is clear that the status quo in caring for this patient demographic is quite unsustainable, as evidenced by the following: In FY2012, each LVHN admission of a self-pay patient cost the hospital an average of $20,000, and each ER visit cost a minimum of $150. With a total of $327 million in uncompensated care provided to the community in 2013, adequate attention to this particular subset of (largely) uninsured patients could provide significant cost savings and improved health outcomes for our Network (Lehigh Valley Health Network [LVHN], 2013).
Predicting the need for community services and the rate of service utilization has always been a challenge for homelessness advocates and healthcare professionals. A significant component of this uncertainty is the aforementioned dearth of reliable data, due primarily to a lack of network-wide screening initiatives that are integrated into our care models. While certain practices and clinics maintain comprehensive records regarding their patients’ socioeconomic standing (e.g., AAO), the majority of our inpatient and outpatient care settings have not standardized an approach to screening for—and responding to—housing instability, despite its profound effects on health outcomes.
At a Network level, it is hoped that continuous collection of valid data related to rates of homelessness within the LVHN patient population can be attained. This prevalence estimate would ultimately allow for projections of utilization patterns and costs of caring for this subgroup, and would provide an invaluable springboard for the Street Medicine Program in the way of operational planning, funding acquisition, outcomes evaluation, and prediction of future trends—thus providing an instrumental resource for clinical care teams (both inpatient and outpatient) as well as Network administrators. Creating such an opportunity for a population that is too often marginalized will be of value to the beneficiaries themselves, to the Network as their “provider,” and to the Lehigh Valley as a whole.
In order to begin addressing this complex issue, we have devised a simple survey method for prospectively capturing the needed data. In addition to demographic data (age and gender), a brief screening tool, comprised of five “Yes or No” questions will be administered by USF medical students and/or EM residents. The screening tool was derived from the US Department of Housing and Urban Development (HUD) 2012 definition for homelessness, which increases its reliability.
This study’s goal is to determine the prevalence of homelessness or at risk for homelessness in the LVHN ED population.
1) The prevalence of homelessness or at risk for homelessness in the LVHN ED population will be different between the summer and winter months.
2) The prevalence of homelessness or at risk for homelessness in the LVHN ED population will be different for men and women.
3) The prevalence of homelessness or at risk for homelessness in the LVHN ED population will be different at our ED sites (Cedar Crest, Muhlenberg, and 17th Street)
4) The prevalence of homelessness or at risk for homelessness in the LVHN ED population will be different by time of day and day of week.
Research Ethics and Recruitment
This study will be reviewed and approved by LVHN’s Institutional Review Board prior to any study surveys being administered. Participation in completing the surveys will be voluntary. There will be little to no risk to study subjects. If a subject chooses not to participate or takes the survey and changes their mind, they may ask not to have their information included and it will not affect their care or their relationships with any of the health care team. Consent will be implied by survey completion.
Population and Sample Studied
All adult patients who present to the EDs at CC, MHC, or 17th and Chew during scheduled survey times.
Patients must be 18 years or older, must speak English, have capacity to answer survey questions, not critically ill, and are willing to participate.
Patients must not be less than 18 years old, do not speak English, do not have capacity to answer survey questions, critically ill, or are unwilling to participate.
A simple five-question survey (see Appendix 2) will be administered in the three LVHN ED settings on a scheduled basis. All patients within an ED pod who meet inclusion and exclusion criteria will be approached during the scheduled period.
Despite its brevity and ease of use, this screening tool may not be readily adopted by clinicians and/or administrators, whose time is often in great demand and short supply, and whose priorities may include their own QI initiatives and research efforts. Thus, it is our aim to recruit volunteers (e.g., USF medical students and EM residents, who will be assigned to this study) specifically for this project with screening as their only assigned duty to ensure data is collected appropriately.
In light of the documented trends related to ED utilization by homeless individuals—as well as the relative ease with which ED patients can be divided into screening sub-groups based on location—it is our intention to initiate the first round of survey administration within LVHN’s three Emergency Departments (LVHN Cedar Crest, Muhlenberg, and 17th Street locations). The screening protocol will be uniform throughout all sites, but will adapt to the unique layout and patient flow of each setting: Within LVHN-CC ED, a single pod will be selected for screening purposes, and all patients within it will be eligible for voluntary involvement; at LVHN-17th Street, it is expected that a single volunteer will be able to administer the questionnaire to all eligible and willing patients within the ED on a given shift; and at LVHN-Muhlenberg, volunteers will rotate between the Rapid Assessment Unit (RAU) and various pods over the course of multiple weeks, allowing for comprehensive sampling of patient demographics and problem acuity. Patients are regularly assigned randomly to different sections of the Emergency Department, so by screening all patients in these sections we can eliminate screening bias.
Volunteers will be assigned shifts, so that over the course of the screening period, a reasonable collection representing available hours and days of the week will have coverage. Survey administration will occur electronically through the use of iPads and/or laptop computers, utilizing a secure online interface, which will store only anonymous and de-identified data. All patients will be afforded the opportunity to receive technical support from volunteers in order to complete the survey, although their input will otherwise be self-reported and fully anonymous, and not be linked to any patient record; a patient may choose to decline participation in the screening at any point in the interaction, which will be documented as such. Patients with positive screens for homelessness (or “at risk for homelessness”) will be offered a Street Medicine Consult at the healthcare providers’ discretion. Screening will initially occur over the course of 3 months--in Summer, 2015--and repeated in the winter months of 2016, in an attempt to capture seasonal variation. Coordination of volunteers, data collection, and other resources will be the responsibility of the research study team leadership.
Data Management and Confidentiality
The data for this study will come from electronically-captured survey responses, which will be automatically de-identified and anonymous. These responses will be transmitted to the ED Research study coordinator, who will assign each survey a study ID number. All databases will be password-protected and saved in a restricted-access electronic research folder on the LVHN “X” drive. These databases will be maintained by ED Research. Only CITI-trained study team members will have access to the study data.
Analysis of data will be straightforward and assess for homelessness (questions 2-5 of the administered survey) and at risk for homelessness (question 1 of the administered survey). Continuous demographic parameters (patient age) will be summarized as descriptive statistics, and categorical demographic parameters such as gender will be summarized as a proportion of the subject group and compared using a two-tailed Fischer’s Exact test (or Chi-Square if the prevalence is greater than 5 subjects). Prevalence (site, season [summer or winter], time of day, and day of week) will be calculated based on the number of occurrences (any subject in which one or more survey questions are answered “yes”) and the size of the population studied. P-values will be used to report differences with significance set at 0.05. Analyses will be performed using Statasoftware v.12.1 (Stata Corporation, College Station, TX).
The initial phase of determining the prevalence of homelessness within LVHN’s patient population will occur via this screening protocol through mid to late 2015. By initiating this questionnaire in smaller cohorts, it may suggest a need to screen for homelessness Network-wide. It will be important to eventually identify all at-risk patients to connect them with much needed resources, including the Street Medicine Program. After both screening rounds are complete, we will closely analyze the mechanisms of data collection and reporting, and enhance them as necessary in order to optimize their accuracy and efficiency. It is hoped that this will only be the beginning of a more comprehensive effort, catalyzing processes network-wide (e.g. incorporating the questionnaire, and identifiers tied to positive screens, into Epic) that will carry us forward in eliminating health disparities within our community.
Bharel M., Lin W.C., Zhang J., O’Connell E., Taube R., and Clark R.E. Health Care Utilization Patterns of Homeless Individuals in Boston: Preparing for Medicaid Expansion Under the Affordable Care Act. AJPH, 2013; 103(S2): S311-S317.
Hwang S.W., Weaver J., Aubry T., and Hoch J.S. Hospital Costs and Length of Stay Among Homeless Patients Admitted to Medical, Surgical, and Psychiatric Services. Medical Care, 2011; 49(4)350-354.
Ku B.S., Scott K.C., Kertesz S.G., and Pitts S.R. Factors Associated with Use of Urban Emergency Departments by the U.S. Homeless Population. Public Health Reports, 2010; 125: 398-405.
Kushel M.B., Perry S., Bangsberg D., Clark R., and Moss A.R. Emergency Department Use Among the Homeless and Marginally Housed: Results from a Community-Based Study. AJPH. 2002; 92(5): 778-784
Life on the streets is hard. (n.d.). Retrieved from NHCHC Website: www.nhchc.org.
LVHN 2013 Annual Report - Community Benefit FY 2013. (2013). Retrieved from LVHN Website: http://www.lvhn.org/2013report/downloads/LVHN_Community_Benefits.pdf.
Morris D.M., & Gordon J.A. The Role of the Emergency Department in the Care of Homeless and Disadvantaged Populations. Emerg Med Clin N Am, 2006; 24: 839-848
Project HOPE. (2014). Retrieved from Nova Southeastern University College of Osteopathic Medicine Website: http://medicine.nova.edu/epr/project-hope.html
What is the official definition of homelessness? (2015). Retrieved from NHCHC Website: http://www.nhchc.org/faq/official-definition-homelessness/
Wilkins C., & Elliot J. (2010, June). Background Paper - Chronic Homelessness. Retrieved from USICH Website: http://usich.gov/resources/uploads/asset_library/BkgrdPap_ChronicHomelessness.pdf
There are no sources in the current document.
Definitions of homelessness (adapted directly from NHCHC website):
U.S. Department of Health and Human Services (HHS) definition:
A homeless individual is defined in section 330(h)(5)(A) as “an individual who lacks housing (without regard to whether the individual is a member of a family), including an individual whose primary residence during the night is a supervised public or private facility (e.g., shelters) that provides temporary living accommodations, and an individual who is a resident in transitional housing.” A homeless person is an individual without permanent housing who may live on the streets; stay in a shelter, mission, single room occupancy facilities, abandoned building or vehicle; or in any other unstable or non-permanent situation. [Section 330 of the Public Health Service Act (42 U.S.C., 254b)]
An individual may be considered to be homeless if that person is “doubled up,” a term that refers to a situation where individuals are unable to maintain
Published In/Presented At
Glovas, S. (2015, July 31) Implementing a Screening Tool for Homelessness at LVHN: A Pilot Project A Subset Analysis on Gender Differences. Poster presented at LVHN Research Scholar Program Poster Session, Lehigh Valley Health Network, Allentown, PA.
Research Scholars (Acknowledgements and Co-authored Publications), Research Scholars - Posters