Abstract
Objectives. We sought to examine the utilization and impact of enabling services, such as interpretation and eligibility assistance, among underserved Asian American, Native Hawaiian, and other Pacific Islander (AANHOPI) patients served at 4 community health centers.
Methods. For this project, we developed a uniform model for collecting data on enabling services and implemented it across 4 health centers that served primarily AANHOPI patients. We also examined differences in patient characteristics between users and nonusers of enabling services.
Results. Health center patients used many enabling services, with eligibility assistance being the most used service. In addition, compared with nonusers, users of enabling services were more likely to be older, female, AANHOPI, and uninsured (P < .05).
Conclusions. For underserved AANHOPI patients at community health centers, enabling services are critical for access to appropriate care. We were the first to examine uniform data on enabling services across multiple health centers serving underserved AANHOPI patients. More data on enabling services and evaluation are needed to develop interventions to improve the quality of care for underserved AANHOPI patients.
Community health centers (CHCs) are safety nets for some of the country's most vulnerable patients, but many of these patients are unable to access or use this needed medical care without enabling services.1 Major barriers to care include an inability to pay, culture and language, and insurance status.2,3 Enabling services have been identified by the National Association of Community Health Centers as key facilitators to health care delivery and are defined as nonclinical services that are specifically linked to a medical encounter or provision of medical services that aim to increase access to health care and improve health outcomes.4 Enabling services at CHCs include language interpretation, health education, and financial or insurance eligibility assistance. Enabling services have long been considered to be critical components of health care delivery for CHC patients, who are disproportionately low-income, uninsured, and of minority backgrounds. However, despite the perceived importance and potential of enabling services for improving health care for vulnerable populations, little is known about the utilization of enabling services at CHCs or the impact of these services on health care access and outcomes among medically underserved populations. In particular, no studies have examined enabling services and their impact on medically underserved Asian American, Native Hawaiian, and other Pacific Islander (AANHOPI) patients at CHCs.
Few studies have examined the effect of enabling services at CHCs on health care access and outcomes among people of color.5–9 The results of these few studies suggest that enabling services can make a significant contribution to improved access and quality of care. For example, case management has been shown to be effective at improving specific disease conditions.10,11 Interpretation services have been shown to increase both timeliness of care12 and patient satisfaction and decrease the number of emergency room visits, thereby reducing costs.5
Medically underserved AANHOPI patients at CHCs, in particular, rely more on enabling services such as interpretation and eligibility assistance for access to medical care. A lack of enabling services leads underserved AANHOPI patients and other people of color to underutilize medical services at CHCs, and causes such patients to be underrepresented in the health care system.13–15 For example, communication difficulties stemming from language or cultural issues are common reasons for AANHOPI persons to avoid health services16,17 and to feel less confident that they can get needed care as compared with non-Hispanic Whites.16 Enabling services at CHCs can help underserved AANHOPI patients obtain culturally and linguistically appropriate and effective health care.
Culturally proficient health care reduces health disparities between racial/ethnic groups.18 Culturally and linguistically appropriate enabling services can help to overcome barriers within the health care system by improving patient–provider interactions, increasing patient knowledge and understanding of treatments, and improving patient safety.19 Interpretation services can help patients navigate the health care system more easily and can improve patient–provider communication, resulting in increased medical visits and improved health outcomes. Eligibility assistance and enrollment in health insurance programs can alleviate patient financial concerns associated with care.
Federally qualified health centers are required to provide annual reports to the US Bureau of Primary Health Care as part of the Uniform Data System and to submit information on some of the enabling services provided by their health centers. However, the current Uniform Data System does not adequately capture the full scope of enabling services provided and needed by federally qualified health centers to demonstrate the critical impact of these services, nor does it document the need to adequately finance health centers to ensure full primary care access and utilization among medically underserved patients. As of 2007, the Uniform Data System report included only the number of full-time equivalent staff and encounters for case managers and education specialists and full-time equivalents only for outreach workers, transportation staff, and a category for “other enabling services.”20
Enabling services are often jeopardized during times of political and financial pressures, because the services are usually nonbillable or nonreimbursable.21 Although some CHC staffs and federal officials have indicated that enabling services improve health care access and outcomes among medically underserved patients, such services have not been adequately supported or funded.22
Our study aimed to rectify the lack of research in this area by collecting important new data on the needs for enabling services at CHCs and the impact of enabling services on the medical care and outcomes of medically underserved AANHOPI patients.
This project was a collaborative effort among the Association of Asian Pacific Community Health Organizations (AAPCHO), the New York Academy of Medicine, and 4 AAPCHO CHCs. Senior staff members from each CHC served on the project advisory council and were actively engaged over a 3-year period in all phases of the project, including planning, design, implementation, evaluation, and dissemination. AAPCHO is a not-for-profit national association that represents 27 community health organizations that primarily serve medically underserved AANHOPI patients. AAPCHO members, which are predominantly CHCs, are located in communities with high concentrations of medically underserved AANHOPI persons and provide comprehensive, primary health care for more than 350 000 underserved AANHOPI patients annually.
The project was implemented and data were collected at 4 AAPCHO CHC sites. The AAPCHO served as the data repository and led the data analysis and reporting. The project team developed a uniform data collection protocol,23 including a handbook for data collection, encounter forms for documenting the provision of enabling services, a manual on data file layout, and an enabling services quick reference card. After a comprehensive examination of the enabling services provided at each of the 4 participating CHCs, we developed 9 categories of enabling services with common definitions across the 4 CHCs. To facilitate data collection, we chose the 9 categories on the basis of our conceptual model as follows: (1) eligibility assistance and financial counseling, (2) interpretation, (3) health education, (4) case management assessment, (5) case management referral, (6) case management treatment, (7) transportation, (8) outreach, and (9) other (e.g., assistance with public housing, parenting workshops).24 We included only enabling services data that were linked to a medical encounter. Data submission instructions, template spreadsheets, and databases were provided to each site. During a pilot test, all providers of enabling services (e.g., community health workers, medical assistants, and interpreters) at each site were trained in data collection, and the data were evaluated for consistency and accuracy. Each of the 4 CHCs adopted its own enabling services encounter form according to the AAPCHO uniform data collection template.23 Thus, although some CHCs may have added additional categories or subcategories to their enabling service encounter forms, all 4 CHCs ultimately collected and reported the same broader categories of uniform data. Site 1 (New York City) chose to pilot test the system in its social services department, where a large number of its enabling services are provided. Thus, data from site 1 did not include all enabling services provided at that site. The other 3 sites pilot tested their enabling services data collection center-wide. In addition, each CHC collected data from patient medical records by using the site's own medical records database system.
The CHC sites were selected on the basis of their membership in AAPCHO, their geographic representation, and the diversity of their patient populations. All 4 CHCs shared similar characteristics with other CHCs nationwide, with high percentages of uninsured and low-income patients.1,25
The sites were located in New York City; Seattle, Washington; and Oahu, Hawaii (2 sites). One site in Oahu primarily served Native Hawaiians and other Pacific Islanders, whereas the other 3 sites served mostly Asian Americans.
We collected patient data from all 4 sites and general health center descriptive characteristics including enabling services from January to December 2004. The patient data included all patients who had 1 or more primary care visits between January–December 2004.
We also collected patient and encounter data from 3 of the 4 sites from June 1, 2003, through June 30, 2004, to compare the characteristics of users and nonusers of enabling services by age, gender, ethnicity, insurance, and health condition. We defined a user as a patient who had a primary care visit in June 2004 and had at least 1 enabling services visit during the previous year (June 1, 2003–June 30,2004). A nonuser was defined as a patient who had a visit in June 2004 but had not had an enabling services visit during the previous year. Ambulatory-care–sensitive conditions were designated as conditions for which primary and preventive care could help to reduce the need for hospitalizations or emergency room visits.26 We collected individual, unduplicated total patient demographic data and total patient encounter data (including 1 or more encounters per patient) for users and nonusers of enabling services by ambulatory-care–sensitive conditions across 3 of the 4 sites; 1 Hawaii site had only partially completed the data collection at the time of study.
We developed a uniform protocol and data collection template that included the International Classification of Diseases, Ninth Revision, code that was used to extract the data, and sent the data to AAPCHO for analysis. We examined differences in patient characteristics and in the percentages of patients with chronic and acute ambulatory-care-sensitive conditions between users and nonusers of enabling services. Differences between groups were formally tested by using logistic or generalized estimating equation logistic regression,27 with enabling service use as the dependent variable, the patient characteristic as the independent variable of interest, and site as a covariate to control for site-to-site variation. Generalized estimating equation logistic regression was used when the model included patient condition (acute, chronic, routine), which varied within patient over encounters. Multivariable logistic models were fit to further evaluate the association between enabling service use and patient characteristics. In the multivariable models, AANHOPI patients were aggregated into a single category and compared with Whites because the odds ratios for the racial categories were of primary interest. Similarly, Medicaid, Medicare, and other public insurance programs were combined into a single public insurance category, because the odds ratios for the aggregated category were of primary interest; moreover, the detailed insurance categories were all similar in that they were significant in the same direction as the aggregated category.
Similar to federally qualified health centers nationwide, the 4 CHCs were located in high-need areas, were open to all residents regardless of ability to pay, were governed by community boards to ensure responsiveness to local needs, and provided comprehensive health and related services.1 In addition, more than half of the patients at each of the 4 CHCs had incomes at or below the federal poverty level, and between 17% and 46% of the patients were uninsured (Table 1). Sites varied in size and geographic location. In 2004, the number of medical encounters by site ranged from 30 652 to 145 398. In addition, the composition of the AANHOPI patient population varied by site and by geographic location. The percentage of AANHOPI patients at the 4 CHCs was 82%, compared with 3.2% at federally qualified health centers nationwide.28 In addition, all 3 states in which the CHCs were located had a higher total AANHOPI population (Hawaii, 79%; New York, 7%; Washington, 9%) proportionally in relation to the US rate (5.1%), with Hawaii having the highest proportion of AANHOPI persons nationally.29 At 3 of the sites, most patients spoke a primary language other than English; at 1 site, most patients were Native Hawaiian and spoke English as a primary language.
TABLE 1 Characteristics of the 4 Community Health Centers: New York, NY; Seattle, WA; Oahu, HI; 2004
| Community Health Center 1 | Community Health Center 2 | Community Health Center 3 | Community Health Center 4 | |
| Medical users, no. | 27 189 | 10 894 | 7691 | 23 602 |
| Medical encounters, no. | 145 398 | 35 506 | 30 652 | 102 054 |
| Enabling services, no. | 9885a | 26 847 | 7510 | 14 861b |
| Enabling services users, no. | 2410a | 11 718 | 2651 | 4803b |
| AANHOPI, % | 99 | 97 | 85 | 82 |
| Top AANHOPI subgroups, group (%) | Chinese (98%) | Chinese (54%), Vietnamese (23%) | Filipino (20%), Chinese (15%), Chuukese (15%) | Native Hawaiian (50%) |
| Family income at or below 200% of FPL, % | 90 | 89 | 87 | 76 |
| Uninsured, % | 24 | 26 | 46 | 17 |
| Enabling service users who were female, % | 69a | 62 | 65 | 61b |
| Mean age, y, of enabling service users | 27a | 43 | 38 | 30b |
| Average length, min, of enabling service per encounter | 13a | 15 | 19 | 27b |
| Most utilized type of enabling service | Case management assessment, financial counseling | Interpretation, financial counseling | Interpretation, financial counseling | Financial counseling, transportation |
Note. AANHOPI = Asian American, Native Hawaiian, and other Pacific Islander; FPL = federal poverty level.
aIncludes only enabling services provided by the Social Services Department in New York City.
bIncludes data only from April–December 2004; data from January–March 2004 were not available.
Across the 4 sites, the number of enabling services provided at each CHC in 2004 ranged from 7510 to 26 847 (Table 1). The average number of services provided per month was 1335 per site and ranged from 626 to 2237. On average, 2.7 services were provided per patient per month. The average number of users per month was 519 per site and ranged from 201 to 975. The average number of users was correlated with the number of providers and resources for providing enabling services. Patients usually required more than 1 service at each visit, according to providers. The most common services across all 4 sites were eligibility assistance or financial counseling (36%) and interpretation services (29%), followed by case management assessment (9%) and health education and supportive counseling (9%).30 Financial counseling included enrollment assistance for public insurance programs and linkage to drug discount programs.
The diversity of enabling services reflected the differing needs of the population served by each CHC. A site with patients who spoke many different primary languages provided high levels of on-site interpretation services. A site with 1 dominant primary language had more bilingual providers and thus provided fewer on-site interpretation services. A site located in a community with few public transportation systems provided a high number of transportation services using their own van service.
The average length per enabling service encounter was 19.5 minutes across all 4 sites and services and varied by type of service. Health education took the longest service time per encounter on average across all 4 sites (data not shown). At 1 site, eligibility assistance took the longest service time per encounter, on average.
Between 61% and 69% of the users of enabling services were women, a proportion similar to the composition of the CHC patients overall; the average age of the users varied (Table 1). Sites also varied in composition of AANHOPI subgroups and in primary language, and these characteristics reflected the characteristics of the overall CHC population. All AANHOPI subgroups used all types of enabling services. Some AANHOPI subgroups, including Chinese, Vietnamese, Korean, and Filipino patients, utilized interpretation more frequently than did other subgroups.
Users of enabling services were predominantly covered by Medicaid (20%–67%) or another type of public insurance, which included state insurance programs. Many users of enabling services were uninsured (18%–40%).
Using unduplicated, total individual patient demographic data (June 1, 2004, to June 30, 2004) and total patient encounter data (June 1, 2003, to June 30, 2004) for 3 of 4 sites, we conducted analyses of the characteristics of individual users and nonusers of enabling services and found significant (P < .05) differences in age, gender, ethnicity, and insurance type (Table 2). Users of enabling services were significantly more likely than nonusers to be older, female, AANHOPI, and uninsured. Multivariable logistic regression yielded similar associations between patient characteristics and enabling services utilization after control for other factors (Table 3). After we controlled for age, gender, insurance, and ethnicity, the proportion of total patient encounters by enabling services users and nonusers with chronic and acute conditions was not significantly different (Table 3). For both groups by total patient encounters, diabetes was the most common chronic condition, and ear, nose, and throat infections were the most common acute condition (Table 4). Patients with diagnoses of acute conditions were more likely to use eligibility assistance services than were patients with chronic conditions (P < .05; data not shown).
TABLE 2 Patient Demographics of Users and Nonusers of Enabling Services at Community Health Centers: New York, NY; Seattle, WA; Oahu, HI; 2004
| Users of Enabling Services, Frequency (%) | Nonusers of Enabling Services, Frequency (%) | Total, Frequency (%) | P valuea | |
| Total | 2656 (100) | 2190 (100) | 4846 (100) | |
| Gender | < .001 | |||
| Female | 1809 (68) | 1255 (57) | 3064 (63) | |
| Male | 847 (32) | 935 (43) | 1782 (37) | |
| Ethnicity | < .001 | |||
| AANHOPI | 2430 (92) | 1870 (86) | 4300 (88) | |
| Chinese | 1150 (43) | 779 (36) | 1929 (40) | |
| Filipino | 165 (6) | 231 (11) | 396 (8) | |
| Korean | 107 (4) | 38 (2) | 145 (3) | |
| Vietnamese | 307 (12) | 307 (14) | 614 (13) | |
| Other Asianb | 137 (5) | 120 (5) | 257 (5) | |
| Native Hawaiian | 469 (18) | 318 (15) | 787 (16) | |
| Other Pacific Islander | 95 (4) | 77 (4) | 172 (3) | |
| White | 132 (5) | 138 (6) | 270 (6) | |
| Other race/ethnicityc | 92 (3) | 180 (8) | 272 (6) | |
| Insurance carrier | < .001 | |||
| Medicaid | 1004 (38) | 976 (45) | 1980 (41) | |
| Medicare | 337 (13) | 251 (11) | 588 (12) | |
| Other public | 505 (19) | 272 (12) | 777 (16) | |
| Private | 285 (11) | 358 (16) | 643 (13) | |
| Self-pay | 525 (20) | 326 (15) | 851 (18) | |
| Age, y | < .001 | |||
| 0–14 | 400 (16) | 630 (29) | 1030 (21) | |
| 15–24 | 390 (15) | 240 (11) | 630 (13) | |
| 25–44 | 687 (26) | 488 (22) | 1175 (24) | |
| 45–64 | 687 (26) | 501 (23) | 1188 (25) | |
| ≥ 65 | 492 (19) | 331 (15) | 823 (17) |
Note. AANHOPI = Asian American, Native Hawaiian, and Other Pacific Islander. Patient demographic data were derived from June 1–30, 2004 because of potential changes in insurance status for community health center patients. Data are from 3 of the 4 sites because of limited data available from the fourth site.
aP value from logistic regression of enabling service use on covariate, with control for project site.
bIncludes Japanese and Asian Indian.
cIncludes Black, mixed-Asian American, mixed-Native Hawaiian and other Pacific Islander, and mixed-other.
TABLE 3 Patient Characteristics Independently Associated With the Use of Enabling Services at Community Health Centers: New York, NY; Seattle, WA; Oahu, HI; 2004
| ORa (95% CI) | P | |
| Ageb | 1.00 (1.00, 1.01) | < .001 |
| Female gender | 1.77 (1.60, 1.95) | < .001 |
| Insurance | ||
| Private (Ref) | 1.00 | < .001 |
| Public | 1.45 (1.24, 1.69) | |
| Self-pay | 3.25 (2.69, 3.92) | |
| Ethnicity | ||
| White (Ref) | 1.00 | < .001 |
| AANHOPI | 1.92 (1.47, 2.50) | |
| Other | 0.33 (0.23, 0.46) | |
| Patient condition | ||
| Acute (Ref) | 1.00 | .647 |
| Chronic | 1.00 (0.97, 1.02) | |
| Other | 1.00 (0.98, 1.02) | |
| Routine | 0.99 (0.96, 1.01) | |
Note. AANHOPI = Asian American, Native Hawaiian, and Other Pacific Islander; CI = confidence interval; OR = odds ratio. Generalized estimating equation logistic regression model with enabling service use as the dependent variable, and age, gender, insurance, ethnicity, and patient condition as independent variables, with control for project site. The data for this analysis were patient diagnosis data from 3 of the 4 sites, because of the limited data available from the fourth site.
aOR for comparison of a given category to the referent is statistically significant when the CI does not include 1.
bMultivariable model with age as a categorical variable yielded similar results. Age categories were (in years): < 1, 1–4, 5–14, 15–24, 25–44, 45–64, and > 64.
TABLE 4 Chronic and Acute Ambulatory-Care-Sensitive Conditions of Patient Encounters by Users and Nonusers of Enabling Services: New York, NY; Seattle, WA; Oahu, HI; 2004
| Encounters by Users, Frequency (%) or Average | Encounters by Nonusers, Frequency (%) or Average | Total, Frequency (%) or Average | |
| Chronic conditions | |||
| Diabetes | 324 (5) | 209 (4) | 533 (5) |
| Pulmonary disease | 174 (3) | 135 (3) | 309 (3) |
| Cellulitis | 160 (3) | 95 (2) | 255 (2) |
| Asthma | 127 (2) | 144 (3) | 271 (2) |
| Hypertension | 108 (2) | 109 (2) | 217 (2) |
| Congestive heart failure | 71 (1) | 26 (1) | 97 (1) |
| Other chronic conditions | 29 (0.5) | 23 (0.4) | 52 (0.5) |
| Total chronic conditions | 993 (17) | 741 (15) | 1734 (16) |
| Acute conditions | |||
| Ear, nose, and throat infections | 667 (11) | 693 (14) | 1360 (12) |
| Kidney and urinary infections | 163 (3) | 89 (2) | 252 (2) |
| Gastroenteritis and dehydration | 36 (1) | 81 (2) | 117 (1) |
| Other acute conditions | 31 (0.6) | 31 (0.5) | 62 (0.6) |
| Total acute conditions | 902 (15) | 896 (18) | 1798 (16) |
| Routine care | 874 (15) | 773 (15) | 1647 (15) |
| Other reasons | 3224 (54) | 2676 (53) | 5900 (53) |
| Total conditions | 5993 (100) | 5086 (100) | 11 079 (100) |
| Average no. conditions per patient | 1.74 | 1.76 | … |
Note. Data for chronic and acute conditions were derived from patient medical encounters from June 1, 2003–June 30, 2004, for those patients who had a primary care visit from June 1-30, 2004. Note that patients may have had multiple encounters. Data are from 3 of the 4 sites, because of the limited data available from the fourth site.
This project was among the first to examine uniform enabling services data across CHCs in multiple states and was the first such study to examine the impact of enabling services on medically underserved AANHOPI patients. AAPCHO, CHCs, and the National Association of Community Health Centers have been collaborating to standardize enabling services data collection as part of a patient-centered “medical home” movement. Enabling services are an integral part of CHCs nationally, and a large proportion of CHC patients need enabling services to access care effectively.1 Our results suggested that users of enabling services are more likely to be people of color (AANHOPI), uninsured, female, and older than were nonusers.22 It is not surprising that we found financial counseling to be the most common enabling service, given that patients without insurance coverage need eligibility assistance to be linked to affordable health services.31 In addition, individuals who are publicly insured also need assistance with navigating enrollment requirements, including managed care choices, understanding their health coverage and rights, and obtaining provider referrals.31 People of color with limited English proficiency and who are foreign born often need interpretation services and cultural liaisons to the US health care system.5 In addition, the elderly population is more likely to have public insurance coverage, to have greater language and cultural barriers, and often to have comorbidities and complex medical conditions that require case management.32,33 Furthermore, the results of our study suggest that patients with acute conditions are more likely to be uninsured and may delay seeking services and coverage. It is also possible that patients with acute conditions require more eligibility services than do patients with chronic conditions, because an acute condition often brings a patient into the clinic for the first time.
We implemented a multisite model of data collection at 4 CHCs across 3 states. This model showed that it is possible to use uniform definitions and coding for enabling services and encounters and to collect uniform patient data across CHCs. We aggregated the data for analysis, assessed differences between CHCs and their patient populations, and aimed to communicate best practices in provision of enabling services, particularly for underserved AANHOPI patients.
Our data had several limitations. Our results reflected only the 4 CHCs that served predominantly underserved AANHOPI patients and were not nationally representative. Thus, it is not clear whether the enabling service protocol may be appropriate for CHCs serving a small number of AANHOPI patients and other underserved ethnic populations. The enabling services data collection varied across sites, and only 1 year of data was obtained. Because enabling services are not reimbursed by encounter, many CHCs are understaffed and find it difficult to find the time to document all the services provided. In addition, enabling services at 1 site were collected in only a single service department, not centerwide. Thus, the utilization of enabling services in our study was likely to be underestimated. Our estimates of actual enabling services may also have been understated as a function of our specific definitions. For example, patients utilizing financial counseling services who did not complete an application for a sliding fee or health insurance program were not counted.
Continuing CHC collaborations and collection of enabling services data over several years will allow the data collection to become more consistent and will facilitate assessment of trends over time. Documentation of encounters and services was found to be more complete and accurate after training sessions were held and feedback provided to enabling services staff. Annual training sessions will be important next steps. However, the patterns of services that were undocumented were not found to be systematic; thus, the characteristics of users of enabling services are likely to be representative.
Furthermore, for the purposes of this study, enabling services were linked only to medical encounters. We plan to expand future data collection to include all patients, including dental patients and mental health patients. We also plan to continue our data collection with current and new CHC partners to build on these pioneering efforts and to develop a larger, more comprehensive demonstration project. Further longitudinal data collection and analysis of types of enabling services by AANHOPI subgroups, types of insurance, and specific acute or chronic conditions will also be important next steps.
Despite these limitations, we have provided important new multisite data across 3 states about enabling services at CHCs for medically underserved AANHOPI patients. Uniform enabling services data will make it possible to better understand the integral role of these services at CHCs and to examine the impact of enabling services in improving access to and quality of medical care. Our data can provide CHC managers with the tools needed to allocate their limited resources to meet the needs of their patients, as well as to inform the development of new strategies and interventions. In fact, the commitment and prioritization by CHC senior management to enabling services provision and data collection were critical to the success of this project. The utilization of uniform data collection tools nationally across the CHC system could provide valuable information for policymakers, managers, and providers in reducing health disparities and improving the quality of care for the most vulnerable populations.
Acknowledgments
This project was funded in part by grants from the US Department of Health and Human Services Office of Minority Health (MPCMP051003) and the Agency for Healthcare Research and Quality (HS13401).
The authors acknowledge the federally qualified health center Project Site Coordinators, Mary Oneha, Lynn Sherman, Monique van der Aa, and Janelle Jacobs, for their valuable collaboration on the Enabling Services Accountability Project, and Katherine Chen for her research and administrative support.
Human Participant Protection
Institutional review board approval was obtained from the New York Academy of Medicine.
