Abstract
This study uses electronic health record data to describe primary care services offered by US community health centers in March through May 2020, including SARS-CoV-2 testing, well-child visits, HbA1c testing, and cancer screening.
The effect of the coronavirus disease 2019 (COVID-19) pandemic on the primary care community health centers (CHCs) in the US has not been well described. CHCs serve approximately 30 million people,1 including high proportions of patients susceptible to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and adverse outcomes.2,3 Surveys indicate that many primary care practices, including CHCs, lacked access to SARS-CoV-2 testing, personal protective equipment, and emergency financial resources, and that in-person visits and preventive and chronic illness care have been reduced.4 We describe SARS-CoV-2 testing, individuals with positive test results, and changes in select primary care services in a multistate network of CHCs in the first weeks of the pandemic.
Methods
We used electronic health record (EHR) data from OCHIN, a network of CHCs in 21 states that share a linked EHR.5 We included patients aged 79 years or younger with a clinic visit between January 1, 2019, and May 31, 2020. We examined patients who underwent SARS-CoV-2 testing and had results positive for SARS-CoV-2 in their EHR by patient characteristics (Table 1) from March 1, 2020, to May 31, 2020, using bivariable generalized estimating equation Poisson regression accounting for clustering of patients within health systems. Patients were considered “tested” once their first test was recorded; subsequent tests were included only in counts of positive test results. We also compared the clinic-level delivery of common primary care services indicated for varying ages/sexes (relative rates of face-to-face visits with a medical professional, well-child visits, Papanicolaou testing, hemoglobin A1c testing, and mammogram orders) in January 2019 to February 2019, March 2019 to May 2019, January 2020 to February 2020, and March 2020 to May 2020 using generalized estimating equation Poisson regression. Statistical significance was defined as a 95% CI that did not include 1.0. Analyses were performed with SAS Enterprise Guide, version 7.15. This study was considered exempt by the Advarra Institutional Review Board.
Table 1. Characteristics of Primary Care Patients in 132 Community Health Centers (CHC), January-May 2020.
| Characteristic | Patient sample, No. (%)a | SARS-CoV-2 tests per 1000 patientsb | Results positive for SARS-CoV-2 per 1000 tested patientsc | ||||
|---|---|---|---|---|---|---|---|
| No. | Rate (95% CI) | Relative rate (95% CI) | No. | Rate (95% CI) | Relative rate (95% CI) | ||
| Total patients | 1 982 954 | 33 266 | 15.1 (10.9-20.8) | 9348 | 200 (165-243) | ||
| Patient statusd | |||||||
| Established | 1 907 901 (96.2) | 28 216 | 13.2 (9.6-18.1) | 1 [Reference] | 7570 | 194 (159-237) | 1 [Reference] |
| New patient | 75 053 (3.8) | 5050 | 62.7 (44.1-89.1) | 4.76 (3.84-5.89) | 1778 | 238 (189-299) | 1.22 (1.01-1.49) |
| Sexe | |||||||
| Male | 879 565 (44.4) | 15 049 | 14.8 (10.5-20.9) | 1 [Reference] | 4660 | 215 (177-262) | 1 [Reference] |
| Female | 1 102 822 (55.6) | 18 200 | 15.3 (11.2-20.9) | 1.03 (0.92-1.15) | 4684 | 188 (154-230) | 0.88 (0.81-0.95) |
| Age, yf | |||||||
| <12 | 325 850 (16.4) | 1437 | 3.6 (2.1-6.1) | 0.21 (0.15-0.30) | 315 | 169 (132-217) | 0.85 (0.71-1.01) |
| 12-19 | 234 033 (11.8) | 1497 | 5.6 (3.6-8.6) | 0.33 (0.26-0.41) | 427 | 214 (176-260) | 1.08 (0.97-1.19) |
| 20-39 | 598 692 (30.1) | 11 583 | 17.0 (12.5-23.1) | 1 [Reference] | 3269 | 199 (163-142) | 1 [Reference] |
| 40-59 | 514 716 (26.0) | 12 606 | 21.4 (16.0-28.8) | 1.26 (1.13-1.41) | 3603 | 199 (164-242) | 1.00 (0.93-1.08) |
| ≥60 | 309 279 (15.6) | 6086 | 16.8 (11.9-23.7) | 0.99 (0.80-1.23) | 4406 | 197 (152-255) | 0.99 (0.82-1.19) |
| Ethnicityg | |||||||
| Non-Hispanic | 1 166 088 (58.8) | 19 434 | 8.2 (3.4-19.8) | 1 [Reference] | 4328 | 63 (25-159) | 1 [Reference] |
| Hispanic | 673 182 (34.0) | 11 099 | 6.2 (2.0-18.9) | 0.76 (0.18-3.23) | 4307 | 174 (93-326) | 2.78 (1.35-5.73) |
| Unknown | 143 684 (7.25) | 2733 | 5.1 (1.7-15.2) | 0.63 (0.44-0.90) | 713 | 122 (58-259) | 1.95 (0.92-4.15) |
| Race | |||||||
| White | 1 186 540 (59.8) | 17 414 | 16.4 (11.9-22.6) | 1 [Reference] | 3819 | 182 (146 -228) | 1 [Reference] |
| American Indian | 18 876 (1.0) | 349 | 17.6 (11.8-26.5) | 1.08 (0.83-1.39) | 81 | 202 (145-281) | 1.11 (0.85-1.44) |
| Asian | 117 656 (5.9) | 1206 | 9.1 (5.6-14.8) | 0.56 (0.41-0.76) | 268 | 186 (145-241) | 1.02 (0.80-1.31) |
| Black or African | 368 532 (18.6) | 8482 | 13.8 (9.0-21.2) | 0.84 (0.64-1.11) | 3211 | 218 (167-285) | 1.20 (0.92-1.57) |
| Multiple race | 21 628 (1.1) | 234 | 13.2 (8.9-19.5) | 0.81 (0.65-1.00) | 50 | 231 (171-311) | 1.26 (1.01-1.59) |
| Native Hawaiian/Pacific Islander | 14 051 (0.7) | 376 | 17.2 (10.5-28.2) | 1.05 (0.72-1.53) | 182 | 259 (200-335) | 1.42 (1.06-1.91) |
| Other/unknown | 255 671 (12.9) | 5205 | 14.7 (9.0-24.0) | 0.89 (0.61-1.31) | 1789 | 253 (210-304) | 1.39 (1.31-1.61) |
| Primary languageh | |||||||
| English | 1 369 936 (69.1) | 23 146 | 16.0 (11.5-22.1) | 1 [Reference] | 4983 | 162 (127-207) | 1 [Reference] |
| Spanish | 433 037 (21.8) | 7242 | 16.4 (11.3-23.8) | 1.03 (0.82-1.30) | 3247 | 304 (253-365) | 1.87 (1.40-2.52) |
| Other | 179 595 (9.1) | 2530 | 8.0 (4.6-13.9) | 0.50 (0.32-0.77) | 984 | 274 (209-358) | 1.69 (1.26-2.26) |
| Charlson Comorbidity Indexi | |||||||
| 0 | 954 598 (48.1) | 13 738 | 12.9 (8.9-18.9) | 1 [Reference] | 4499 | 230 (191-277) | 1 [Reference] |
| 1-3 | 672 847 (33.9) | 11 055 | 14.8 (10.8-20.2) | 1.14 (0.98-1.34) | 3074 | 203 (168-245) | 0.88 (0.79-0.99) |
| ≥4 | 355 476 (17.9) | 8473 | 20.5 (15.7-26.7) | 1.59 (1.29-1.95) | 1775 | 157 (120-205) | 0.68 (0.55-0.84) |
| Insurancej | |||||||
| Private | 288 909 (14.6) | 6823 | 18.4 (12.8-26.5) | 1 [Reference] | 1735 | 171 (134-218) | 1 [Reference] |
| Medicaid/other public | 1 037 795 (52.3) | 10 535 | 8.4 (5.5-12.9) | 0.46 (0.36-0.58) | 2748 | 177 (142-220) | 1.03 (0.88-1.20) |
| Medicare | 155 611 (7.9) | 2419 | 13.1 (8.8-19.6) | 0.71 (0.59-0.87) | 492 | 159 (114-222) | 0.93 (0.76-1.13) |
| Uninsured | 497 738 (25.1) | 13 263 | 26.9 (19.0-38.0) | 1.46 (1.02-2.10) | 4151 | 229 (191-276) | 1.34 (1.12-1.61) |
Patients with CHC contact from January 2019 through May 2020 at any of the 431 US study clinics (in AK, CA, CO, CT, GA, ID, IN, LA, MA, MN, MO, MT, NC, NJ, NY, OH, OR, SC, TX, WA, and WI).
Rates of testing from March through May 2020 determined by generalized estimating equation (GEE) Poisson models. More than 99% of reported severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) tests were reverse transcriptase–polymerase chain reaction tests via nasophargyngeal swabs. Tests were counted if the result was available directly in the electronic health record; most tests were likely ordered/collected in the clinic and performed at a lab offsite. Results were unique patients tested; multiple tests were not counted.
Rates of positive results determined by GEE Poisson models.
Status of new patient if they had no contact with the CHC before March 2020.
A total of 567 patients reported other or unknown sex status and were not reported in this table.
On January 1, 2020. A total of 264 patients did not have a documented age and were not reported in this table.
Race, ethnicity, and income were collected via by clinics via voluntary self-report. Categories were United States Office Management and Budget categories, and inclusion was necessary to evaluate testing and positive results in groups vulnerable to poor health outcomes.
A total of 386 patients had an unknown primary language and were not reported in this table.
As of last visit, determined from historical problem list and encounter diagnosis codes; 33 patients had an unknown index and were not reported in this table.
As reported at last visit; 2901 patients had unknown insurance status and were not reported in this table.
Results
Among 1 938 736 patients seen at 431 primary care clinics in 132 CHCs, 33 266 patients (1.7%) underwent SARS-CoV-2 testing, of whom 9348 (28% of all tests) had positive results (Table 1). New patients (n = 75 053) seen between March 1, 2020, and May 31, 2020, had higher SARS-CoV-2 testing rates compared with established patients. Being Asian, speaking a language other than English or Spanish, and being publicly insured (Medicare or Medicaid/other) were associated with lower testing rates. Having a Charlson Comorbidity Index6 of at least 4 and being uninsured were associated with higher testing rates. Speaking Spanish, being Hispanic, being uninsured, and speaking a language other than English or Spanish were associated with higher rates of positive test results than reference comparators. Being female and having a Charlson Comorbidity Index of 1 to 3 or greater than or equal to 4 (vs 0) were associated with lower rates of positive test results. There were no differences in rates of positive test results across age groups.
In March 2020 to May 2020, compared with March 2019 to May 2019, rates of face-to-face visits (relative rate [RR], 0.58 [95% CI, 0.51-0.66]), well-child visits (RR, 0.47 [95% CI, 0.41-0.54]), Papanicolaou tests (RR, 0.34 [95% CI, 0.30-0.39]), hemoglobin A1c tests (RR, 0.51 [95% CI, 0.46-0.57]), and mammograms (RR, 0.44 [95% CI, 0.38-0.51]) declined. No such decline was observed comparing January to February in 2019 vs 2020 (Table 2).
Table 2. Rates of Services in Primary Care Clinics in Community Health Centers in January Through May of 2019 and 2020.
| Primary care service | Rate of monthly services delivered per 1000 patients (95% CI)a | Monthly service relative rate (95% CI) | ||||
|---|---|---|---|---|---|---|
| January-February 2019 | March-May 2019 | January-February 2020 | March-May 2020 | January-February 2020 vs 2019 | March-May 2020 vs 2019 | |
| Completed face-to-face visits with a medical professionalb | 246 (228-265) | 258 (242-276) | 249 (232-266) | 151 (131-173) | 1.01 (0.97-1.05) | 0.58 (0.51-0.66) |
| Completed well-child visitsc | 232 (199-272) | 213 (179-253) | 238 (205-276) | 101 (80-126) | 1.03 (0.97-1.08) | 0.47 (0.41-0.54) |
| Papanicolaou testsd | 15.8 (13.9-18.1) | 16.5 (14.8 -18.3) | 15.0 (13.6-16.6) | 5.6 (4.8-6.6) | 0.95 (0.88-1.03) | 0.34 (0.30-0.39) |
| Hemoglobin A1c testse | 141 (134-148) | 147 (140-154) | 147 (140-153) | 75 (68-83) | 1.04 (1.00-1.08) | 0.51 (0.46-0.57) |
| Mammogram ordersf | 31.9 (28.8-35.4) | 34.9 (32.0-38.0) | 32.6 (29.7-35.8) | 15.4 (13.2-18.1) | 1.02 (0.96-1.10) | 0.44 (0.38-0.51) |
Community health center (CHC) clinics were limited to 431 clinics from 132 community health centers on OCHIN electronic health records by January 1, 2019. Patient sample size for each service was determined by the number of patients in the specified age and sex category with a primary care visit between January 1, 2019, and March 31, 2020. Rates of monthly delivery and CIs were determined by generalized estimating equation models of monthly service counts specifying a Poisson distribution, a log link, and an offset equal to log(patient sample size/ adult_1000). We adjusted for within–health system correlation using robust sandwich estimators.
Excludes visits coded as mental or behavioral health visits. Medical professionals were defined as physicians, advanced practice clinicians, or registered nurses. Patient sample size was 768 810 adult patients with primary care visits aged 19 to 79 years.
Includes all visits by patients younger than 18 years with preventive care evaluation and management codes or diagnosis codes indicating well-child visits (n = 289 117 children with primary care visits).
Resulted papanicolaou tests for women aged 21 to 65 years (n = 354 911 women with primary care visits).
Resulted hemoglobin A1c tests in patients with diabetes aged 21 to 79 years (n = 107 100 patients with diabetes and primary care visits).
Mammogram orders in women aged 50 to 70 years (n = 134 627 women with primary care visits).
Discussion
Despite limited test availability,4 CHCs reported thousands of SARS-CoV-2 tests, underscoring their important role in serving vulnerable populations. The 75 053 patients seen for the first time after the pandemic started suggests that CHCs were an access point in the midst of the crisis. Small differences in testing and positive rates by race, and larger differences by ethnicity, preferred language, and insurance status, suggest ongoing need for targeted, language-concordant testing strategies.
In the pandemic’s initial weeks, delivery of common services in CHCs declined, possibly due to in-person care reductions. Although these changes may have been necessary and unavoidable, the potential consequences are concerning because reductions in preventive/chronic disease care may affect population health.
Study limitations include not having clinical presentation data of tested patients or additional clinic-level factors, such as testing capacity, geography, local disease rates, and infrastructure. CHC patients may have received tests that were not recorded in the networked EHR; therefore, the results may underestimate true population testing rates. Results may not be generalizable to all primary care clinics; however, the network has similar patient characteristics as CHCs nationwide.1,5 Further research should assess differences in clinical presentation in various demographic groups, the effect of remote (telephone/video) visits on care quality, and the potential ongoing role of CHCs in mitigating pandemic-related inequity.
Section Editor: Jody W. Zylke, MD, Deputy Editor.
References
- 1.Community Health Center Chartbook. National Association of Community Health Centers ; 2019. Accessed April 24, 2019. http://www.nachc.org/wp-content/uploads/2019/01/Community-Health-Center-Chartbook-FINAL-1.28.19.pdf
- 2.Coronavirus (COVID-19). Centers for Disease Control and Prevention website. Accessed May 27, 2020. https://www.cdc.gov/coronavirus/2019-ncov/index.html
- 3.Price-Haywood EG, Burton J, Fort D, Seoane L. Hospitalization and mortality among black patients and white patients with COVID-19. N Engl J Med. 2020;382(26):2534-2543. doi: 10.1056/NEJMsa2011686 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Quick COVID-19 survey. The Larry Green Center website. Accessed June 17, 2020. https://www.green-center.org/covid-survey
- 5.2019 Annual Report. OCHIN ; 2020. Accessed July 15, 2020. https://ochin.org/2019-annual-report
- 6.Huntley AL, Johnson R, Purdy S, Valderas JM, Salisbury C. Measures of multimorbidity and morbidity burden for use in primary care and community settings: a systematic review and guide. Ann Fam Med. 2012;10(2):134-141. doi: 10.1370/afm.1363 [DOI] [PMC free article] [PubMed] [Google Scholar]