Objective

To evaluate whether establishing links among community-based organizations (CBOs) and between CBOs and an academic health system reduces hospital and emergency department use among residents of East and Southeast Baltimore, Maryland

Study Design

Design Element	Description
Study Design	Randomized controlled trial
Population	Staff and clients from 20 CBOs, 4,917 high-risk patients from Johns Hopkins Health System (JHHS), and inpatient and outpatient staff members from JHHS
Interventions/ Comparators	A multicomponent intervention using a website, a software platform, and in-person meetings Usual CBO operation without the multicomponent intervention
Outcomes	Primary: hospital and emergency department utilization Secondary: self-reported knowledge, experience, and referral numbers of CBO staff members, CBO clients, and JHHS staff members
Timeframe	12-month follow-up for primary outcome

The research team conducted a cluster randomized controlled trial to evaluate an intervention to link CBOs in East and Southeast Baltimore, Maryland, with the Johns Hopkins Health System (JHHS) inpatient and outpatient healthcare staff members. The research team worked with representatives from government agencies, JHHS, foundations, and local nonprofit organizations to develop the strategy for CBOs and JHHS to access and share resources.  

Researchers randomly assigned 20 CBOs that provide similar services to the intervention group or the control group. The CBOs in the intervention group participated in monthly meetings with all the intervention CBO partners, used a toolkit on a project website, and had access to a search engine to find local social and medical services. A research assistant volunteered at each CBO and acted as a liaison between the CBO and research team. Each intervention CBO partner also attended monthly meet-and-greet sessions with JHHS staff members. The control group CBOs did not participate in any of the intervention activities and provided services to their clients as usual.

Researchers assigned 4,917 high-risk JHHS patients as intervention or control patients. Researchers based the assignments on proximity of patients’ residences to intervention or control CBOs.

The primary outcome was healthcare use, measured using Medicare and Medicaid claims data and defined as the number of hospital and emergency department visits, as well as length of stay for JHHS patients. For secondary outcomes, the research team conducted baseline surveys and follow-up surveys 12 months later with:

• CBO staff members about their referrals of one or more clients to the healthcare system, receipt of one or more patient referrals from the healthcare system, levels of confidence in referring clients to other CBOs, frequency of communication with other CBOs, and number of referrals to CBOs in the previous month
• CBO clients regarding whether they received information about other CBOs, referrals to the healthcare system by CBO staff, or referrals to a CBO by a healthcare provider; surveys also assessed clients’ difficulties with finding a job, paying bills, getting transportation, getting food, or finding housing in the previous month
• JHHS inpatient and outpatient staff members regarding their perceived barriers to referring patients to community resources, referrals of patients to CBOs in the previous month, perceptions about how well healthcare organizations and community organizations work together, and levels of confidence in their knowledge about CBOs

The research team compared responses between the intervention and control groups and used difference-in-difference analyses to examine changes in responses to survey questions from baseline to follow-up.

Results

The research team found no significant differences between the intervention and control groups for healthcare use or for any outcomes measured via surveys with CBO clients and CBO staff members.

At follow-up, compared with the baseline period, JHHS staff were significantly more likely to refer a patient to an intervention CBO (3-percentage-point increase in inpatient staff referrals, p = 0.051, and 7-percentage-point increase in outpatient staff referrals, p = 0.027). JHHS outpatient staff members also reported a significant reduction in barriers to referring patients to CBOs related to a lack of available resources about CBOs (15 percentage points, p = 0.014) and up-to-date information about community resources (18 percentage points, p = 0.04). JHHS outpatient staff members reported increased confidence in knowledge about CBOs from baseline to follow-up (14 percentage points, p = 0.023).

Limitations

The research team was unable to track JHHS patients’ use of services at specific CBOs; therefore, it was not possible to relate CBO exposures to the outcomes. The team assigned patients to the intervention group or control group based on patient proximity to CBOs. The study included only 20 CBOs; the small sample size may have limited the researchers’ abilities to detect differences.

Conclusions and Relevance

The research team created a network of CBOs with links to JHHS that may have helped community members access health and social services. Although the study did not identify differences between the intervention and control groups for primary outcomes, JHHS staff members reported outcomes that suggest an increase in patient referrals to CBOs and a decrease in barriers to referring patients to CBOs.

Future Research Needs

Future research could explore how creating or strengthening links between CBOs and health networks can aid in improving the health of high-risk patients in urban settings.