Download PDF
Appl Clin Inform 2024; 15(01): 129-144
DOI: 10.1055/s-0044-1779258
Research Article

Improving Vaccine Equity: How Community Engagement and Informatics Facilitate Health System Outreach to Underrepresented Groups

Authors

  • Serena J. Xie

    1   Department of Biomedical Informatics and Medical Education, University of Washington, Seattle, Washington, United States

  • Nicholas R. Mah

    2   Department of Shared Services, Enterprise Access and Innovation, UW Medicine, University of Washington, Seattle, Washington, United States

  • Lisa Chew

    3   Department of Medicine, University of Washington School of Medicine, Harborview Medical Center, University of Washington Medicine, Seattle, Washington, United States

  • Julia Ruud

    4   Department of Performance Improvement, University of Washington, Seattle, Washington, United States

  • Jennifer Hernandez

    5   Ambulatory & Allied Care Services, Harborview Medical Center, University of Washington Medicine, Seattle, Washington, United States

  • Jessica Lowery

    5   Ambulatory & Allied Care Services, Harborview Medical Center, University of Washington Medicine, Seattle, Washington, United States

  • Andrea L. Hartzler

    1   Department of Biomedical Informatics and Medical Education, University of Washington, Seattle, Washington, United States
Further Information
Also available at
 
 PDF Download Permissions and Reprints

Abstract

Background Given the inequities in access to health care resources like COVID-19 vaccination, health systems should carefully consider how to reach underrepresented groups. Reflecting on vaccine rollout efforts holds insight on the role of community engagement and informatics support in promoting health equity.

Objectives This study aimed to assess the effectiveness of four outreach strategies deployed by University of Washington (UW) Medicine in improving vaccine equity over traditional vaccine scheduling online or by phone, we report on appointment scheduling and completion of appointments (i.e., vaccine administration) through (1) automated outreach to individuals from underrepresented groups, (2) temporary “pop-up” clinics in neighborhoods highly impacted by COVID-19, (3) vulnerable population clinics, and (4) mobile vaccine vans.

Methods We conducted a 6-month retrospective analysis of electronic health records (EHR) to describe the sociodemographic characteristics of individuals who scheduled appointments using the outreach strategies and characteristics associated with a greater likelihood of vaccine administration based on appointment completion. To help explain trends in the EHR data, we engaged 10 health system leaders and staff who spearheaded the outreach strategies in follow-up conversations to identify qualitative insights into what worked and why.

Results Compared with traditional scheduling, all outreach strategies except vulnerable population clinics had higher vaccine appointment completion rates, including automated outreach (N = 3,734 [94.7%], p < 0.001), pop-up clinics (N = 4,391 [96.0%], p < 0.001), and mobile vans (N = 4,198 [99.1%], p < 0.001); and lower cancellation rates, including automated outreach (N = 166 [1.1%], p < 0.001), pop-up clinics (N = 155 [0.6%], p < 0.001), and mobile vans (N = 0 [0%], p < 0.001). Qualitative insights emphasized ongoing community partnerships and information resources in successful outreach.

Conclusion Vaccine equity outreach strategies improved the proportion of patients who scheduled and completed vaccination appointments among populations disproportionately impacted by COVID-19. Engaging community partners and equity-focused informatics tools can facilitate outreach. Lessons from these outreach strategies carry practical implications for health systems to amplify their health equity efforts.


Keywords

health equity - COVID-19 vaccines - community–institutional relations - social vulnerability - outcome and process assessment - vulnerable populations - electronic health records and systems - consumer health informatics

Background and Significance

Health inequity is a pressing problem that coronavirus disease 2019 (COVID-19) only further exacerbated by disproportionately impacting individuals based on race, ethnicity, language, socioeconomic conditions, and other social factors like geography.[1] [2] [3] [4] In early 2021 with the rollout of COVID-19 vaccination in the United States, data monitored by Centers for Disease Control and Prevention (CDC) revealed disparities in COVID-19 vaccination rates, particularly among Black and Hispanic adults, which called for prioritizing equitable distribution of vaccines to underserved people and areas.[5]

Because many communities at disproportionate risk for COVID-19 face barriers to vaccine access and uptake, it is essential for vaccine equity strategies to include community engagement efforts that facilitate trust, provide accurate and interpretable information, and respond to the needs and values of different community groups.[6] [7] For example, Gonzalez and colleagues[8] improved vaccination rates through a community outreach program using a call/recall system that addressed vaccine hesitancy and facilitated real-time vaccine scheduling among eligible Black and Hispanic adults.

Studies have reported the use of clinical decision support in electronic health records (EHR), such as provider-facing immunization reminders[9] and alerts[10] and patient-facing personal health records,[11] showing limited impact on vaccine uptake and the need for further research. During the COVID-19 pandemic, health systems responded by prioritizing vaccine administration for at-risk groups, such as health care workers[12] and optimizing resource management for mass vaccination efforts.[13] Other informatics support, such as visualization tools,[14] automated scheduling,[15] and innovative use of Social Vulnerability Index,[3] [16] can help health systems measure the reach of vaccination to vulnerable groups. However, Diaz and colleagues[17] found the proportion of Black and Hispanic individuals receiving COVID-19 vaccine was substantially higher with direct outreach through community-based organizations (CBOs) than online self-scheduling that reserved vaccination for underserved groups in specific zip codes. Given inequities in technology that impact access to vaccination,[18] [19] health systems should carefully consider how to best employ community engagement and informatics support to reach underserved groups.

Reflecting on vaccine rollout efforts offers an opportunity to illustrate how community engagement and informatics support can amplify health system efforts to prioritize health equity. We examine the success of vaccine equity outreach at University of Washington (UW) Medicine as a case example. UW Medicine is a large health care organization in the Pacific Northwest. It was one of the first health systems to encounter and treat COVID-19 patients in the United States, which was supported by the rapid rollout of information technology services to support clinical response.[20] When COVID-19 vaccines came available in early 2021, UW Medicine responded rapidly. Mass vaccination clinics were established in which patients and the public could schedule vaccination appointments following state-eligibility guidelines.[21] Analysis of initial vaccination efforts with traditional scheduling online and by phone for the mass vaccination site, which lacked targeted strategies for vaccine appointment scheduling for underserved populations, demonstrated disparities. Although roughly 41% of UW Medicine's patient population is not White and 10% prefer a language other than English, nearly 80% of COVID-19 vaccines were scheduled by White English-speaking patients in January of 2021. Although the demographic distribution of vaccine-eligible individuals may have varied, the underrepresentation of Black, Indigenous, and People of Color and individuals with limited English proficiency (LEP) was evident.

In response to this emergent disparity, UW Medicine deployed four operational tactics ([Table 1]) to improve equitable access to vaccine appointments for underrepresented groups: (1) automated outreach to individuals from underrepresented groups, (2) temporary “pop-up” clinics in neighborhoods highly impacted by COVID-19, (3) vulnerable population clinics at UW Medicine facilities, and (4) mobile vaccine vans. In compliance with Washington State vaccine phases guidelines,[21] the outreach strategies prioritized groups disproportionately impacted by the COVID-19 pandemic to receive vaccination based on principles developed by the UW Medicine Vaccine Equity Workgroup. Those principles prioritized groups with high COVID-19 positivity rates, high hospitalization/death rates, barriers to vaccine access/low vaccination rates, and CBOs that could support patient engagement and outreach. Given the emergent context, these tactics were not developed with carefully considered hypotheses, but were operationalized pragmatically in rapid response with available resources to make vaccines as immediately accessible to those in need as possible.

Table 1

Vaccine equity outreach strategies

Strategy (dates deployed)

Who the strategy was designed to reach

How the strategy was deployed

Why the strategy was chosen

Automated outreach (March 2–9, 2021)

Eligible health system patients whose race was not White, who were Hispanic/Latino, or whose preferred language was not English. Outreach included patients whose race, ethnicity or language preference was unavailable, or unknown for broad reach

Automated messages were sent to specific racial, ethnic, and language preference groups to help them schedule vaccine appointment at the UW Medicine mass vaccination site:

• Appointments were reserved for patients from specified groups

• Patients were identified by querying the EHR for the specified demographic categories

• Identified patients were contacted through automated phone calls or SMS messages

• Automated messages were sent in English to the patients' preferred method of communication (phone, SMS) and invited them to call a phone number to schedule a vaccine appointment

Strategy was chosen to address underrepresentation of BIPOC and LEP patients in initial vaccine efforts that relied on traditional scheduling by increasing awareness of vaccine availability and how to access

Pop-up clinics (March 11–June 28, 2021)

Eligible individuals in vulnerable neighborhoods in the Seattle metropolitan area that were highly impacted by COVID-19

Temporary “pop-up” clinics were implemented outside the UW Medicine mass vaccination site in neighborhoods highly impacted by COVID-19:

• Vaccination infrastructure was set up in public schools in neighborhoods with vulnerable populations. Fire department assisted with vaccinations at some schools

• Expanded vaccine access was offered through appointments at UW Medicine primary care clinics in vulnerable neighborhoods

Strategy was chosen to reduce the physical distance between the vaccine and patients in neighborhoods disproportionately impacted by COVID-19

Vulnerable population clinics (March 13–May 2, 2021)

Eligible health system patients with specific language and cultural needs, including LEP communities (e.g., Somali, Arabic, Amharic, Oromo, Spanish, Vietnamese, ASL)

Language and culture-specific clinics were offered within the UW Medicine mass vaccination site:

• Clinics designated to serve vulnerable populations by reserving appointment slots only for specific groups

• Outreach occurred through primary care and international medicine clinic staff to fill designated clinic appointments

• Enhanced support through in-person interpreters (rather than phone or video), social workers, and cultural mediators who were present throughout the clinic day to welcome and support patients in their vaccine visit

• Resource-intensive effort to identify and outreach to patients (i.e., outreach by primary care and international medicine clinic staff), and then manually track vaccinations

Strategy was chosen to address underrepresentation of LEP patients in initial vaccine efforts that relied on traditional scheduling by creating a more accessible and culturally sensitive environment

Mobile vans (March 29–June 30, 2021)

Eligible individuals from vulnerable groups and hard to reach locations, including unhoused individuals, low-income seniors, and underrepresented BIPOC groups (e.g., Pacific Islander, African American, Iraqi, and Latinx communities)

RV-style vehicles and vans were used to provide walk-up access to vaccines at diverse locations other than the UW Medicine mass vaccination site:

• Three mobile vans worked in concert with CBOs to reach unhoused, low-income, and underrepresented BIPOC people

• Deployed to a diverse range of locations to provide vaccines at shelters, churches, community organization facilities, parks, grocery stores, community festivals and events, all tailored to the needs and culture of each community

• A full-time dedicated team of medical personnel traveled with the RV at dates and times that the community was available, including evenings and weekends

• Focused on implementing a sustainable, respectful format that adapted to and reflected ongoing feedback from community partners

Strategy was chosen to reduce the physical distance between the vaccine and patient groups with low vaccination rates and barriers to vaccine access

Abbreviations: ASL, American Sign Language; BIPOC, Black, Indigenous, and People of Color; CBO, community-based organization; COVID-19, coronavirus disease 2019; EHR, electronic health records; LEP, limited English proficiency; RV, Recreational Vehicle; UW, University of Washington.


Description of strategies designed to improve equitable access to COVID-19 vaccine appointments at UW Medicine: dates deployed, who the strategy was designed to reach, how the strategy was delayed, and why the strategy was chosen.


To share lessons learned on the vital role of community engagement and informatics support in facilitating equitable outreach, we report on a large-scale EHR data analysis that describes who these outreach strategies successfully reached. However, the data tell only part of the story. It was through our conversations with health system leaders and staff that we learned qualitative insights into why certain outreach strategies were more successful in improving the proportion of appointments scheduled and completed among priority groups. These stories reflect the key roles played by community engagement and informatics.


Objective

To assess the success of vaccine equity outreach strategies, we conducted a retrospective analysis of EHR data focused on vaccine appointment scheduling and completion. We describe the sociodemographic characteristics of patients who (1) used the outreach strategies to schedule appointments and (2) had a greater likelihood of vaccine administration based on completed appointments. Follow-up conversations with health system leaders and staff who spearheaded and deployed the outreach efforts add qualitative insights into what worked and why.


Methods

Electronic Health Records Data Analysis

We extracted an EHR data sample of patients who scheduled first dose COVID-19 vaccine appointments with UW Medicine between December 17, 2020, and June 30, 2021. The data sample included vaccination appointment type (i.e., traditional scheduling online or by phone, and four outreach strategies: automated outreach, pop-up clinics, vulnerable population clinics, mobile vans; [Table 1]), appointment date, appointment status (i.e., completed, cancel, no-show, arrived, and left), and individual sociodemographic characteristics (i.e., age, race, ethnicity, preferred language, zip code).

We assessed two outcomes: scheduled appointments (i.e., whether a strategy reached an individual based on appointment type) and completed appointments (i.e., whether the vaccine was administered based on the proxy of appointment status of “completed”). Due to high demand and limited vaccine at the time this dataset was collected, traditional scheduling of appointments for the mass vaccination site was online or by phone and walk-ins were not generally permitted. Patients who did not schedule vaccine appointments through any of the four outcome strategies were assigned the appointment type “traditional scheduling.” An appointment was marked “completed” only when the vaccine was administered. Based on appointment status, individuals with “completed” appointments were treated as having vaccine administered, whereas individuals with an appointment status of “cancel” and “no show” were treated as not having been administered the vaccination. Only two patients in the dataset had an appointment status “arrived” (treated as completed) and one with appointment status “left” (treated as no show).

Using the Social Vulnerable Index (SVI)[22] and population estimates from Washington State Small Area Estimate Program,[23] we computed a weighted mean SVI for each zip code tabulation area (ZCTA). SVI is a census-based metric that ranges from 0 (“low vulnerability”) to 1 (“high vulnerability”) and includes 15 socioeconomic and demographic factors assessed in an Overall score and four subscores across four themes: Socioeconomic status, Household composition and disability, Minority status and language, and Housing and transportation.[22] The Small Area Estimate Program provides annual population and housing estimates for Washington State based on residential building permits, assessor records, postal delivery statistics, and federal census data. We used the population estimates at the ZCTA level released in September 2020.[23]

We summarized characteristics of the data sample with descriptive statistics. We compared sociodemographics and SVI among the four outreach strategies and traditional scheduling using chi-square (χ 2) for categorical variables and Kruskal–Wallis (H) for continuous variables. When differences were found, we conducted post hoc pairwise comparisons between strategies with Bonferroni corrections using Wilcoxon rank sum (Z) and pairwise χ 2 tests. We report effect sizes for statistically significant pairwise differences, Phi coefficient (r) for categorical variables, and Cohen's d (d) for continuous variables. To identify factors associated with the likelihood of completing vaccine appointments, we performed a logistic regression with sociodemographic characteristics and outreach strategies as predictors.


Follow-up Conversations with Health System Leaders and Staff

A member of the research team (N.R.M.) identified health system leaders and staff who spearheaded the vaccine equity outreach strategies. For each outreach strategy, we invited these stakeholders to share their reflections on what worked and why. Two members of the research team (A.L.H., S.J.X.) met with a total of five health system leaders and five staff through Zoom meetings and email follow-up. We treated leaders/staff as key informants of whom a subset are coauthors (N.R.M., L.C., J.R., J.H., J.L.) and the remainder are acknowledged. One member of the research team (A.L.H.) led meetings that asked stakeholders as a group: (1) What patient group(s) was the strategy intended to reach and how? (2) Looking back, what went well and why? (3) If you could do it again, what would you improve? We recorded meetings, took notes, and documented email follow-up with materials stakeholders shared (e.g., screenshots of tracking dashboards). One member of the research team (A.L.H.) synthesized these data to identify emergent patterns for each outreach strategy in analytic memos.[24] We summarize these insights to help explain trends in the EHR analysis following a sequential explanatory mixed methods approach.[25]



Results

Characteristics of the Electronic Health Records Data Sample

[Table 2] summarizes characteristics of the data sample, including sociodemographics and appointment status for all patients (N = 126,236), including patients who scheduled through traditional scheduling and each of the four vaccine equity outreach strategies. The majority of patients were 18 to 65 years old, White, non-Hispanic/Latino, preferred English, and were from a total of 410 ZCTAs with most (81%) from King County. Patients had a low-to-moderate Overall SVI score and moderate-to-high SVI subscores for Minority status and language and Housing and transportation. Most patients (86%) completed their scheduled vaccine appointment.

Table 2

Patients in the data sample who scheduled vaccine appointments

All patients

N = 126,236 (100%)

Traditional scheduling

n = 113,153 (89.6%)

Automated outreach

n = 3,943 (3.1%)

Pop-up clinics

n = 4,572 (3.6%)

Vulnerable population clinics

n = 332 (0.3%)

Mobile vans

n = 4,236 (3.5%)

Difference

Age (M, SD, range)

45.7 (21.1, 2–115)

45.6 (21.1, 2–115)

61.5 (16.3, 17–97)

35.6 (22.3, 11–105)

49.2 (18.8, 12–93)

45.2 (17.1, 12–103)

H (4) = 3,292 (p < 0.001)

Age group (n, %)

Under 18

15,771 (12.5%)

13,774 (12.2%)

2 (0.05%)

1,821 (39.8%)

15 (4.5%)

159 (3.8%)

χ 2 (12) = 6,168 (p < 0.001)

18 to 65

77,944 (61.7%)

70,333 (62.2%)

1,747 (44.3%)

2,126 (46.5%)

233 (70.4%)

3,505 (82.7%)

65 and older

32,517 (25.8%)

29,046 (25.7%)

2,194 (55.7%)

621 (13.6%)

84 (25.1%)

572 (13.5%)

Race (n, %)

American Indian or Alaska Native

709 (0.6%)

580 (0.5%)

16 (0.4%)

40 (0.9%)

3 (0.9%)

70 (1.7%)

χ 2 (24) = 8,808 (p < 0.001)

Asian

17,387 (13.8%)

15,554 (13.7%)

523 (13.3%)

617 (13.5%)

54 (16.3%)

639 (15.1%)

Black or African American

6,315 (5.0%)

5,225 (4.6%)

233 (5.9%)

203 (4.4%)

69 (20.8%)

585 (13.8%)

Native Hawaiian or Other Pacific Islander

1,162 (0.9%)

555 (0.5%)

28 (0.7%)

31 (0.7%)

4 (1.2%)

544 (12.8%)

White

72,747 (57.6%)

66,379 (58.7%)

2,322 (58.9%)

2,030 (44.4%)

107 (32.2%)

1,909 (45.1%)

Two or more races

3,688 (2.9%)

3,397 (3.0%)

50 (1.3%)

137 (3.0%)

5 (1.5%)

99 (2.3%)

Unavailable or unknown

24,228 (19.2%)

21,463 (15.0%)

771 (16.9%)

1,514 (33.1%)

90 (27.1%)

390 (9.2%)

Ethnicity (n, %)

Hispanic or Latino

7,357 (5.8%)

5,811 (5.1%)

172 (4.4%)

339 (7.4%)

155 (46.7%)

880 (20.8%)

χ 2 (16) = 4,285 (p < 0.001)

Not Hispanic or Latino

89,359 (70.8%)

80,968 (71.6%)

2,870 (72.8%)

2,419 (52.9%)

193 (41.9%)

2,963 (69.9%)

Unavailable or unknown

29,520 (23.4%)

26,374 (23.3%)

901 (22.9%)

1,814 (39.7%)

38 (11.4%)

393 (9.3%)

Preferred language (n, %)

English

116,580 (92.4%)

106,062 (93.7%)

3,590 (91.0%)

3,964 (86.7%)

49 (14.8%)

2,915 (68.8%)

χ 2 (8) = 9,894 (p < 0.001)

Not English

7,666 (6.1%)

5,540 (4.9%)

321 (8.1%)

220 (4.8%)

280 (84.3%)

1,305 (30.8%)

Unavailable or unknown

1,990 (1.6%

1,551 (1.4%)

32 (0.8%)

388 (8.5%)

3 (0.9%)

16 (0.4%)

SVI (M, SD)

Overall SVI

0.35 (0.18)

0.34 (0.17)

0.36 (0.17)

0.37 (0.17)

0.56 (0.18)

0.54 (0.19)

H (4) = 3,818 (p < 0.001)

Socioeconomic status

0.24 (0.15)

0.24 (0.15)

0.25 (0.15)

0.25 (0.14)

0.42 (0.16)

0.41 (0.17)

H (4) = 1,951 (p < 0.001)

Household composition and disability

0.21 (0.15)

0.21 (0.15)

0.22 (0.15)

0.23 (0.13)

0.34 (0.16)

0.32 (0.18)

H(4) = 3,237 (p < 0.001)

Minority status and language

0.55 (0.15)

0.55 (0.15)

0.56 (0.16)

0.58 (0.13)

0.72 (0.14)

0.68 (0.15)

H (4) = 2,605 (p < 0.001)

House type and transportation

0.61 (0.16)

0.60 (0.16)

0.60 (0.16)

0.61 (0.16)

0.71 (0.13)

0.73 (0.16)

H(4) = 4,182 (p < 0.001)

Appointment status (n, %)

Completed

108,415 (85.8%)

95,829 (84.7%)

3,734 (94.7%)

4,391 (96.0%)

263 (79.2%)

4,198 (99.1%)

χ 2 (8) = 1,522 (p < 0.001)

Canceled

14,684 (11.8%)

14,297 (2.7%)

166 (1.1%)

155 (0.6%)

28 (12.3%)

0 (0%)

No show

3,137 (2.4%)

3,027 (12.7%)

43 (4.2%)

26 (3.4%)

41 (8.4%)

38 (0.9%)

Abbreviations: SD, standard deviation; SVI, Social Vulnerable Index.


Characteristics of patients with scheduled vaccine appointments, comparing traditional scheduling and four outreach strategies deployed at University of Washington Medicine.



Who Scheduled Appointments Using the Outreach Strategies?

Of the 126,236 individuals in the data sample, 13,083 (10.4%) scheduled through one of the four outreach strategies ([Fig. 1]). Of those, the majority scheduled through automated outreach (N = 3,943, 30.1%), pop-up clinics (N = 4,572, 35.0%), and mobile vans (N = 4,236, 32.4%). Sociodemographic characteristics differed significantly among traditional scheduling and the four outreach strategies, including age, race, ethnicity, preferred language, and SVI ([Table 2]). [Fig. 2] shows the distribution of appointments scheduled through each outreach strategy by race, ethnicity, and preferred language.

Zoom
Fig. 1 Number of appointments made through each outreach strategy by date.
Zoom
Fig. 2 Distribution of patients who scheduled appointments for each outreach strategy across the four outreach strategies by race (top), ethnicity (middle), and preferred language (bottom).

Automated Outreach

Compared to traditional scheduling, automated outreach reached a higher proportion of individuals who are Black or African American (N = 233 [5.9%], χ 2 [1] = 14.0, p < 0.001, r = 0.01) and prefer languages other than English (N = 321 [8.1%], χ 2 [1] = 83.7, p < 0.001, r = 0.03). Despite its goal to prioritize Hispanic or Latino individuals, automated outreach did not reach a higher proportion of those individuals than traditional scheduling. Compared with traditional scheduling, automated outreach reached individuals in more socially vulnerable areas, reflected in a higher Overall SVI (Z = 4.3, p < 0.001, d = 0.07), and higher SVI subscores for Socioeconomic status (Z = 3.1, p = 0.002, d = 0.05), Household composition and disability (Z = 5.3, p < 0.001, d = 0.08), and Minority status and language (Z = 4.0, p < 0.001, d = 0.06).


Pop-up Clinics

Pop-up clinics had the highest proportion of individuals with unavailable or unknown race (N = 1,514, 33.1%), ethnicity (N = 1,814, 39.7%), and preferred language (N = 388, 8.5%) among all outreach strategies and traditional scheduling. Yet compared to traditional scheduling, pop-up clinics still reached higher proportions of individuals who are American Indian or Alaska Native (N = 40 [0.9%], χ 2 [1] = 10.3, p = 0.001, r = 0.01) and Hispanic or Latino (N = 399 [7.4%], χ 2 [1] = 45.6, p < 0.001, r = 0.02). Compared with traditional scheduling, pop-up clinics also reached individuals with higher Overall SVI (Z = 13.0, p < 0.001, d = 0.2) and higher SVI subscores for Socioeconomic status (Z = 5.7, p < 0.001, d = 0.07), Household composition and disability (14.6, p < 0.001, d = 0.1), Minority status and language (Z = 13.9, p < 0.001, d = 0.2), and Housing and transportation (Z = 3.6, p < 0.001, d = 0.04).


Vulnerable Population Clinics

Like pop-up clinics, vulnerable population clinics reached a relatively large proportion of individuals with unavailable or unknown race (N = 90, 27.1%). Compared to traditional scheduling, a higher proportion of individuals reached by vulnerable population clinics were Black or African American (N = 69 [20.8%], χ 2 [1] = 190.9, p < 0.001, r = 0.04), Hispanic or Latino (N = 155 [46.7%], χ 2 [1] = 1,139.1, p < 0.001, r = 0.1), and preferred languages other than English (N = 280 [84.3%], χ 2 [1] = 4,277, p < 0.001, r = 0.2). Vulnerable population clinics also reached a higher proportion of individuals with preferred languages other than English than that group reached by automated outreach (N = 321 [8.1%], χ 2 [1] = 4,277, p < 0.001, r = 0.2), pop-up clinics (N = 220 [4.8%], χ 2 [1] = 2,129, p < 0.001, r = 0.66), and mobile vans (N = 1,305 [30.8%], χ 2 [1] = 387, p < 0.001, r = 0.3). Individuals scheduled through vulnerable population clinics had the highest Overall SVI compared to traditional scheduling (Z = 18.8, p < 0.001, d = 1.3), automated outreach (Z = 17.3, p < 0.001, d = 1.2), pop-up clinics (Z = 16.4, p < 0.001, d = 1.1), and mobile van (Z = 2.3, p = 0.02, d = 0.1). Efforts to reach LEP individuals through vulnerable population clinics was also reflected in the highest SVI subscore for Minority status and language, which was higher than traditional scheduling (Z = 18.7, p < 0.001, d = 1.1), automated outreach (Z = 17.2, p < 0.001, d = 1.1), pop-up clinics (Z = 16.9, p < 0.001, d = 1.0), and mobile vans (Z = 5.3, p < 0.001, d = 0.3).


Mobile Vans

Compared to traditional scheduling, mobile vans reached a higher proportion of individuals who are American Indian or Alaska Native (N = 70 [1.7%], χ 2 [1] = 94, p < 0.001, r = 0.03), Asian (N = 639 [15.1%], χ 2 [1] = 6, p = 0.01, r = 0.01), Black or African American (N = 585 [13.8%], χ 2 [1] = 731, p < 0.001, r = 0.08), Native Hawaiian and Other Pacific Islander (N = 544 [47.5%], χ 2 [1] = 6,704, p < 0.001, r = 0.24), and Hispanic or Latino (N = 880 [20.8%], χ 2 [1] = 1,855, p < 0.001, r = 0.13). Nearly half (N = 544, 47.5%) of all Native Hawaiian and other Pacific Islanders in the overall data sample scheduled their vaccine appointments through mobile vans. Compared to traditional scheduling, mobile vans reached a higher proportion of individuals with preferred languages other than English (N = 1,305 [30.8%], χ 2 [1] = 4,988, p < 0.001, r = 0.21). This strategy also reached individuals with a higher Overall SVI compared to traditional scheduling (Z = 61.1, p < 0.001, d = 1.1) and a higher SVI subscore for Housing and transportation than traditional scheduling (Z = 13.7, p < 0.001, d = 0.8), automated outreach (Z = 34.6, p < 0.001, d = 0.8), pop-up clinics (Z = 34.8, p < 0.001, d = 0.8), and vulnerable population clinics (Z = 2.7, p = 0.01, d = 0.1), reflecting one goal of mobile vans to reach the unhoused.



Who was Most Likely to be Administered the Vaccine Based on Appointment Completion?

[Table 3] summarizes characteristics of patients who completed appointments that were scheduled through traditional scheduling and outreach strategies, which were similar to those who scheduled appointments. [Fig. 3] shows the distribution of appointments completed through each outreach strategy by race, ethnicity, and preferred language.

Table 3

Completed vaccine appointments

All patients

N = 108,415 (100%)

Traditional scheduling

n = 95,829 (88.4%)

Automated outreach

n = 3,734 (3.4%)

Pop-up clinics

n = 4,391 (4.1%)

Vulnerable population clinics

n = 263 (0.2%)

Mobile vans

n = 4,198 (3.9%)

Difference

Age (M, SD, range)

45.7 (20.8, 11–105

45.6 (20.7, 11–104)

61.4 (16.3, 17–97)

35.2 (22.4, 11–105)

48.9 (18.8, 12–93)

45.2 (17.1, 12–103)

H (4) = 3,303 (p < 0.001)

Age group (n, %)

Under 18

11,914 (11.0%)

9,929 (10.4%)

2 (0.1%)

1,812 (41.3%)

14 (5.3%)

157 (3.7%)

χ 2 (12) = 6,891 (p < 0.001)

18 to 65

69,559 (64.2%)

62,244 (65.0%)

1,665 (44.6%)

1,988 (45.3%)

185 (70.3%)

3,477 (82.8%)

65 and older

26,941 (24.8%)

23,656 (24.7%)

2,067 (55.4%)

590 (13.4%)

64 (24.3%)

564 (13.4%)

Race (n, %)

American Indian or Alaska Native

627 (0.6%)

500 (0.5%)

16 (0.4%)

39 (0.9%)

3 (1.1%)

68 (1.6%)

χ 2 (24) = 8,066 (p < 0.001)

Asian

14,697 (13.6%)

12,934 (13.5%)

488 (13.1%)

601 (13.7%)

39 (14.8%)

635 (15.1%)

Black or African American

5,466 (5.0%)

4,429 (4.6%)

212 (5.7%)

196 (4.5%)

52 (19.8%)

577 (13.7%)

Native Hawaiian or Other Pacific Islander

1,082 (1.0%)

481 (0.5%)

27 (0.7%)

31 (0.7%)

4 (1.5%)

539 (12.8%)

White

63,141 (58.2%)

56,994 (59.5%)

2,206 (59.1%)

1,955 (44.5%)

93 (35.4%)

1,893 (45.1%)

Two or more races

3,150 (2.9%)

2,866 (3.0%)

47 (1.3%)

134 (3.1%)

5 (1.9%)

98 (2.3%)

Unavailable or unknown

20,252 (18.7%)

17,625 (18.4%)

738 (19.8%)

1,435 (32.7%)

67 (25.5%)

387 (9.2%)

Ethnicity (n, %)

Hispanic or Latino

6,560 (6.2%)

5,060 (5.4%)

163 (4.6%)

331 (7.7%)

130 (50.2%)

876 (20.9%)

χ 2 (16) = 3,995 (p < 0.001)

Not Hispanic or Latino

77,018 (72.4%)

68,937 (73.3%)

2,711 (76.6%)

2,329 (54.3%)

105 (40.5%)

2,936 (69.9%)

Unavailable or unknown

22,737 (21.5%)

20,033 (21.3%)

667 (18.9%)

1,627 (38.0%)

24 (9.3%)

386 (9.2%)

Preferred language (n, %)

English

100,172 (92.4%)

90,004 (93.9%)

3,415 (91.5%)

3,822 (87.0%)

43 (16.3%)

2,888 (68.8%)

χ 2 (8) = 8,829 (p < 0.001)

Not English

6,659 (6.1%)

4,642 (4.8%)

289 (7.7%)

215 (4.9%)

218 (82.9%)

1,295 (30.8%)

Unavailable or unknown

1,584 (1.5%)

1,183 (1.2%)

30 (0.8%)

354 (8.1%)

2 (0.8%)

15 (0.4%)

SVI (M, SD)

Overall SVI

0.36 (0.18)

0.35 (0.17)

0.35 (0.17)

0.37 (0.17)

0.55 (0.18)

0.54 (0.19)

H (4) = 3,975 (p < 0.001)

Socioeconomic status

0.25 (0.15)

0.24 (0.15)

0.25 (0.15)

0.25 (0.14)

0.41 (0.17)

0.41 (0.17)

H (4) = 3,632 (p < 0.001)

Household composition and disability

0.21 (0.15)

0.21 (0.15)

0.22 (0.15)

0.23 (0.13)

0.34 (0.16)

0.32 (0.18)

H (4) = 1,984 (p < 0.001)

Minority status and language

0.55 (0.15)

0.55 (0.15)

0.56 (0.15)

0.58 (0.13)

0.71 (0.14)

0.68 (0.15)

H (4) = 3,150 (p < 0.001)

House type and transportation

0.61 (0.16)

0.60 (0.16)

0.60 (0.16)

0.61 (0.15)

0.71 (0.13)

0.73 (0.16)

H (4) = 2,393 (p < 0.001)

Abbreviations: SD, standard deviation; SVI, Social Vulnerable Index.


Characteristics of patients with completed vaccine appointments, comparing traditional scheduling and four outreach strategies deployed at University of Washington Medicine.


Zoom
Fig. 3 Distribution of patients who completed appointments for each outreach strategy across the four outreach strategies by race (top), ethnicity (middle), and preferred language (bottom).

Appointment status varied significantly among traditional scheduling and outreach strategies (χ 2 [8] = 1,522 [p < 0.001]). Compared with traditional scheduling, all outreach strategies except vulnerable population clinics had higher vaccine appointment completion rates, including automated outreach (N = 3,734 [94.7%], χ 2 [1] = 299, p < 0.001), pop-up clinics (N = 4,391 [96.0%], χ 2 [1] = 446, p < 0.001), mobile vans (N = 4,198 [99.1%], χ 2 [1] = 672, p < 0.001) and lower cancellation rates, including automated outreach (N = 166 [1.1%], χ 2 [1] = 249, p < 0.001), pop-up clinics (N = 155 [0.6%], χ 2 [1] = 348, p < 0.001), and mobile vans (N = 0 [0%], χ 2 [1] = 524, p < 0.001; [Table 2]).

Regression analysis found a higher likelihood of appointment completion when scheduled through mobile vans (p < 0.001), pop-up clinics (p < 0.001), and automated outreach (p < 0.001; [Table 4]). Moreover, higher completion rates were associated with younger age, English as preferred language, lower social vulnerability regarding Household composition and disability, and higher social vulnerability regarding Minority status and language and Housing and transportation.

Table 4

Results of logistic regression with patient sociodemographic characteristics and outreach strategies as predictors for completing vaccination appointments at University of Washington Medicine

Predictor

Coefficient

Odds ratio

p-Value

Age

−0.02

0.98

<0.001[a]

American Indian or Alaska Native

−0.06

0.94

0.78

Asian

−0.08

0.92

0.07

Black or African American

0.08

1.09

0.22

Native Hawaiian or Other Pacific Islander

0.49

1.63

0.04

Two or more races

−0.32

0.72

0.005[b]

Unavailable or Unknown race

−0.20

0.82

<0.001[a]

Hispanic or Latino

0.16

1.17

0.06

Non-English language preference

−0.19

0.83

<0.001[a]

SVI 1: Socioeconomic status

−0.20

0.82

0.24

SVI 2: Household composition and disability

−1.19

0.30

<0.001[a]

SVI 3: Minority status and language

0.79

2.19

<0.001[a]

SVI 4: Housing and transportation

0.48

1.62

<0.001[a]

Outreach strategy: Automated outreach

1.44

4.24

<0.001[a]

Outreach strategy: Pop-up clinics

1.15

3.15

<0.001[a]

Outreach strategy: Vulnerable population clinics

0.01

1.01

0.97

Outreach strategy: Mobile vans

3.01

20.38

<0.001[a]

Abbreviation: SVI, Social Vulnerability Index.


a p < 0.001.


b p < 0.01.



Qualitative Insights from Health System Leaders and Staff about Outreach Strategies

Qualitative insights about what strategies worked and why from conversations with health system leaders and staff helped expand on findings from the EHR analysis. In particular, these stakeholders surfaced community engagement activities and information resources as key ingredients for success. For example, stakeholders described factors contributing to the higher proportion of completion rates among outreach strategies compared to traditional scheduling, including collaboration between community partners and clinical staff in tracking individuals in need of vaccination through shared resources (e.g., informal lists of pending vaccine recipients).

When asked what worked well for automated outreach, stakeholders described using the EHR to identify UW Medicine patients for large-scale outreach based on race, ethnicity, and language preference. Although the inclusion of individuals with unavailable or unknown race, ethnicity, and language preference was intended to reduce disparities, it may have inadvertently resulted in some vaccine appointments for individuals from well-represented groups. Care should be taken when defining specific groups for outreach when EHR data are incomplete. Although EHR helped identify individuals for outreach, there were insufficient tools for tracking who received and responded to automated messages, which made it difficult to monitor the impact of automated outreach. Community members were not engaged in planning for automated outreach, but could have guided increased receptiveness to outreach messages through tailoring languages with cultural sensitivity (e.g., Spanish messages might have improved the proportion of Hispanic/Latino individuals who scheduled and completed appointments).

When asked what worked well for pop-up clinics, stakeholders described how this strategy brought vaccines to people in their own neighborhoods, rather than requiring people to travel to a medical center. Stakeholders described the use of social media and local news announcements to inform community members about pop-up events, with messaging early and often to improve turnout. The stakeholders described their community engagement efforts focused largely on partnerships with public schools, which facilitated vaccination of children, parents, families, and teachers. While stakeholders perceived many community members to appreciate the convenience of local vaccine access, they described pop-up clinics as vital to breaking down access barriers, such as transportation and time off work that could have otherwise prevented vaccination. In contrast to automated outreach, pop-up clinics served many individuals without existing patient data in the EHR. Stakeholders described challenges associated with collecting complete demographic data for individuals who were new to UW Medicine “in the field” during vaccination events that resulted in high rates of missing race, ethnicity, and language preference data.

When asked what worked well for vulnerable population clinics, stakeholders described guidance provided by insider knowledge of existing community advocates, including interpreters and cultural mediators who were employed by the health system and had a history of working with LEP populations. For example, stakeholders described efforts to create a welcoming environment where individuals saw increased presence from members of their community, and received language access through in-person interpreters and vaccination regardless of insurance coverage. Stakeholders describe the use of risk stratification models using EHR data to identify priority groups for outreach. They also emphasized scheduling clinics at accessible times and with consideration of events of significance to specific communities (e.g., Ramadan).

When asked what worked well for mobile vans, stakeholders described strong partnerships they forged with community organizations and leaders. This strategy was characterized by stakeholders as a deep engagement over time that was guided through community conversations to understand community needs, barriers, and community-built strategies for vaccination. For example, stakeholders described listening sessions held with CBOs prior to launching mobile vans. One health system leader recounted a listening session in which a CBO provided strong input on how well-resourced health systems should recognize and fund the labor, effort, expertise, and space that CBOs provide within community partnerships, which shaped the strategy moving forward. In addition to providing CBO partners with funding for each vaccination event, stakeholders described the ongoing feedback sought from community partners to iterate and adapt the outreach strategy. They used this feedback to make changes that ensured sufficient time for planning, recurring presence, relationship-building over time, and a respectful partnership that tailored events to community needs at convenient times and local spaces. Stakeholders described the vaccine equity dashboard UW Medicine implemented, which monitored daily vaccine patient volumes, LEP status, age, race, ethnicity, gender, housing status, region, and payor in the EHR. Health Insurance Portability and Accountability Act (HIPAA)-and Memorandum of Understanding (MOU)-compliant dashboard data were shared with community partners and external funders to track the ongoing impact of mobile vans. Other key reflections stakeholders shared about mobile vans included the ability to adapt to diverse needs and spaces (e.g., staff intentionally changed the site look and feel depending on the group served), institutional flexibility to allow community peer interpreters, alternative forms of identification, use of proper pronouns and preferred names, and a strong sense of partnership. As one health system leader told us, “The most successful events were those where the community partners were engaged, because those are the people that patients trusted.”



Discussion

Compared to traditional scheduling, the outreach strategies improved vaccine equity for vulnerable populations disproportionately impacted by COVID-19. The four outreach strategies reached intended groups at a higher proportion of scheduled appointments than traditional scheduling, and three were associated with a higher likelihood of vaccine completion, higher completion rates, and lower cancellation rates: automated outreach, pop-up clinics, and mobile vans. While all strategies reached groups that historically experience social vulnerabilities, each strategy excelled in reaching a higher proportion of different groups than traditional scheduling. Based on scheduled appointments, automated outreach reached Black/African American individuals who prefer languages other than English. Pop-up clinics reached American Indian/Alaska Native and Hispanic/Latino individuals. Vulnerable population clinics reached individuals who prefer languages other than English. Mobile vans effectively reached unhoused individuals, people who are not White, and those who prefer languages other than English. Qualitative insights from health system leaders and staff emphasized ongoing community partnerships and information resources that fostered success. Based on these insights, we offer recommendations to strengthen outreach strategies through ongoing partnerships with community representatives and informatics tools ([Table 5]).

Table 5

Recommendations for health systems to improve health equity outreach to vulnerable groups

A. Partner with community representatives and organizations to

 ● Tailor outreach messages with appropriate language and cultural sensitivity

 ● Breakdown access barriers (e.g., transportation) by leveraging local community spaces that are welcoming, trusted, and community-owned (e.g., public schools, libraries)

 ● Direct social media and news announcements about outreach events to the appropriate community groups early and often to improve turnout

 ● Leverage insider knowledge of health system advocates who work closely with vulnerable groups that outreach is designed to reach (e.g., interpreters, cultural mediators, community health workers)

 ● Create a welcoming environment where people see presence of other people from their community with shared language and culture

 ● Schedule outreach around events of significance to the community (e.g., Ramadan)

 ● Hold community conversations to listen to community needs, barriers, and community-built solutions, rather than pushing health system solutions

 ● Share philanthropic support for the labor, expertise, space, and effort required to hold successful outreach events

 ● Seek ongoing feedback to iterate and adapt outreach strategies that flexibly match community needs and preferences

B. Leverage equity-focused informatics tools to facilitate community outreach efforts

 ● Query the electronic health records to identify priority populations for large scale outreach, while taking care in defining groups when patient data are incomplete (e.g., unknown race)

 ● Create tools that track which individuals receive and respond to outreach, and which groups may be missed and remain underrepresented

 ● Create lightweight tools that facilitate “in the field” collection of accurate and complete data from individuals who are new to the health system during outreach efforts

 ● Implement risk stratification models that identify high-risk patient groups for outreach

 ● Develop dashboards and reports that monitor the daily impact and reach of outreach efforts, enabling iteration and adaptation of outreach strategies

 ● Leverage social referral platforms to facilitate secure exchange of care recipient data between community-based organizations and the health systems

Study findings carry implications for health systems to improve health equity outreach to vulnerable groups, which could translate beyond vaccination to improve access for screening, prevention, chronic disease management, and emergency care.[26] Our recommendations offer practical advice for clinical informatics practitioners to implement community-based participatory and informatics approaches that align with research-based principles for outreach to underserved populations.[27] Because these recommendations are based on lessons from an academic health system, it is important to consider their applicability to lower resourced organizations and the need for future work. For example, mobile vans can be expensive. Leveraging local infrastructure, such as public libraries, for pop-up health events might be less cost prohibitive for many organizations. In contrast, other recommendations may be more generalizable, such as scheduling outreach around holidays significant to the community.

Community partnerships take shape over time. Early on, COVID-19 vaccines were in high demand but limited supply, which required careful prioritization based on State-eligibility guidelines. Later when vaccine supply improved, the convenience of walk-up appointments may have facilitated high completion rates for pop-up clinics and mobile vans. For example, school-based pop-up clinics may have reduced barriers for working parents and those without transportation. Concurrently, vaccine hesitancy became a stronger focus of outreach, with increased emphasis on community relationships and trust. This is where investment in community engagement paid off most. Early conversations regarding community needs facilitated partnership, mutual respect, and opportunities for ongoing feedback, process improvement, and sustained engagement. Communities are different and health systems should listen to what each group needs and tailor their approach over time to meet those needs.

Furthermore, data matter. Collecting accurate and complete data on race, ethnicity, language, and other markers of social advantage and disadvantage is foundational to enabling health systems to reach communities and identify who is at risk. Missing data pose challenges to outreach and monitoring. The high rate of missing sociodemographic data, particularly for pop-up clinics is a clear limitation of our analysis. Traditional registration forms were found to create literacy barriers and unease in some individuals (e.g., citizenship). There is a need for lightweight, secure methods to streamline collection of sociodemographic data “in the field,” such as mobile and quick response technology.[28] Informal systems for sharing vaccine recipient information between CBOs and health systems could be supported with social referral platforms integrated with EHR.[29] Despite these opportunities, our findings illustrate the value of EHR-based tools to track vaccination among priority populations. Through our follow-up conversations with health system leaders and staff, we learned that although staff for all strategies sought data to monitor the impact of their effort (e.g., most data “hand gathered”), only the mobile van had the support of a vaccine equity dashboard that tracked daily vaccine patient volumes and sociodemographic data, and reflected the lowest rate of missing data. Such EHR-based informatics tools with proper staff training could have benefited all strategies with collection of data needed to drive health equity efforts.

This work is not without limitations. First, roughly one-fifth of the data sample had missing race or ethnicity, which could have biased the results in reflecting the true proportion of scheduled and completed vaccine appointments among priority groups the outreach strategies were designed to reach. Second, we used the outcome of completed vaccination appointments as a proxy for actual vaccine administration, which could have overestimated the results. Although an appointment was marked “completed” only when the vaccine was administered, a completed appointment could have incorrectly documented patients who declined vaccination after receiving education on side effects. Third, considering the large sample sizes in our analysis, some statistically significant differences with small effect sizes may not be clinically meaningful. Yet, some groups reached by outreach strategies have small numbers but consist of relatively high proportions for that outreach strategy (e.g., 280 appointments scheduled through vulnerable population clinics by LEP patients consist of 84% of all appointments scheduled through this strategy). Finally, the data sample reflects individuals who responded to outreach. We lack data on people who were offered the option to schedule, but did not respond or refused to schedule appointments. For example, automated outreach aimed to reach a broad demographic. Groups that traditionally have low response rates could still have low response rates to automated outreach. Similarly, we were unable to ask patients about their motivations to schedule and complete vaccinations nor their experience with outreach strategies. Understanding patients' perspectives on what outreach strategies work and why is a critical area for future research.

Despite these limitations, this study has several strengths to inform future work, including the use of patient zip code to incorporate social vulnerability[22] in the analysis. Alternative sources of geocoded data could improve representation in population-specific vulnerability profiles, such as the income,[30] wages,[31] school free or reduced-price lunch eligibility,[32] crime,[33] and other measures of community well-being from the American Community Survey,[34] Area Deprivation Index,[35] and Opportunity Index.[36] Future work should explore the use of geocoded data to map the geospatial distribution of individuals touched by health equity outreach using CDC SVI interactive map[37] or Area Deprivation Index maps.[38] Such maps could enrich EHR-based visualizations for monitoring the need and impact of health equity initiatives that account for social drivers of health. Another strength was coupling EHR analysis with firsthand insights from health system stakeholder, which led to practical lessons about what works. Future efforts should guide health systems in meaningful and measurable methods for patient and community engagement. Graffigna et al[39] provide a roadmap for operationalizing measurable patient engagement at multiple levels from self-care to service codesign, health technology assessment, and advocacy. Equity-focused informatics tools, combined with qualitative input from diverse stakeholder groups, can guide health systems in demonstrating “return on engagement” in amplifying health equity.


Conclusion

Given persistent health care disparities, health systems must carefully consider how to improve health equity through outreach to underrepresented groups. Vaccine equity outreach strategies improved the proportion of patients who scheduled and completed vaccination appointments among vulnerable populations disproportionately impacted by COVID-19. Vaccine equity strategies that were found most effective engaged community members as partners in flexible, adaptable outreach that leveraged tools in the EHR to identify vulnerable groups and track the reach of outreach efforts on an ongoing basis. These findings carry practical implications for how community engagement and informatics support can amplify health system efforts to prioritize health equity by reaching vulnerable groups.


Clinical Relevance Statement

Given persistent health care disparities, it is more important than ever for health systems to improve health equity. With a focus on COVID-19 vaccination, findings from this study indicate that outreach efforts using informatics tools and partnerships with community representatives can improve access to equitable health care services among vulnerable populations. Examples include automated outreach, pop-up clinics, mobile vans, and vulnerable population clinics designed in partnership with underrepresented communities and monitored through EHR-based tools, like health equity dashboards.


Multiple Choice Questions

  1. Which of the following outreach strategies was deployed through SMS messages and phone calls?

    • Mobile vans

    • Pop-up clinics

    • Automated outreach

    • Vulnerable population clinics

    Correct answer: c. Of the four outreach strategies described and compared in this paper, automated outreach was the only strategy deployed through SMS messages and phone calls. For automated outreach, patients from intended populations were first identified by querying the EHR and then contacted through automated SMS messages or phone calls to schedule a vaccine appointment.

  2. Compared to traditional scheduling, vulnerable population clinics reached a higher proportion of which population?

    • People from areas with low social vulnerability

    • People who prefer languages other than English

    • Black or African American

    • Asian

    Correct answer: b. Of the four outreach strategies described and compared in this paper, vulnerable population clinics reached more people who prefer languages other than English. Outreach efforts focused on reaching individuals with LEP by offering language- and culture-specific clinics with in-person interpreters.

  3. Which of the following outreach strategies provided funding to CBOs to support their substantial role in the outreach effort?

    • Automated outreach

    • Pop-up clinics

    • Vulnerable population clinics

    • Mobile vans

    Correct answer: d. The mobile van strategy provided funding to CBOs to support their effort for outreach associated with each vaccination event. In listening sessions with CBOs prior to launching the mobile vans, the CBOs shared strong input on how well-resourced health systems should recognize and fund the labor, effort, expertise, and space that CBOs provide within community partnerships, which shaped the strategy moving forward.



Conflict of Interest

None declared.

Acknowledgements

We wish to thank Martine Pierre-Louis, Tricia Madden, Jenny Brackett, Nicholas Postiglione, Naomi Matana Shike, 10 CBOs, and many other community partners in this effort.

Protection of Human Subjects

This analysis was reviewed by the Institutional Review Board at University of Washington and determined exempted (IRB ID STUDY00013550).



Address for correspondence

Andrea Hartzler, PhD
850 Republican St. Building C. Seattle, WA 98109
United States   

Publication History

Received: 19 September 2022

Accepted: 22 December 2023

Article published online:
14 February 2024

© 2024. Thieme. All rights reserved.

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany


  Permissions and Reprints
Zoom
Fig. 1 Number of appointments made through each outreach strategy by date.
Zoom
Fig. 2 Distribution of patients who scheduled appointments for each outreach strategy across the four outreach strategies by race (top), ethnicity (middle), and preferred language (bottom).
Zoom
Fig. 3 Distribution of patients who completed appointments for each outreach strategy across the four outreach strategies by race (top), ethnicity (middle), and preferred language (bottom).