Pandemic-Related Practice Changes and CVD Risk Management in Community Clinics

ABSTRACT

Objectives: Understanding how the COVID-19 pandemic affected cardiovascular disease (CVD) risk monitoring in primary care may inform new approaches for addressing modifiable CVD risks. This study examined how pandemic-driven changes in primary care delivery affected CVD risk management processes.

Study Design: This retrospective study used electronic health record data from patients at 70 primary care community clinics with scheduled appointments from September 1, 2018, to September 30, 2021.

Methods: Analyses examined associations between appointment type and select care process measures: appointment completion rates, time to appointment, and up-to-date documentation for blood pressure (BP) and hemoglobin A_1c (HbA_1c).

Results: Of 1,179,542 eligible scheduled primary care appointments, completion rates were higher for virtual care (VC) vs in-person appointments (10.7 percentage points [PP]; 95% CI, 10.5-11.0; P < .001). Time to appointment was shorter for VC vs in-person appointments (–3.9 days; 95% CI, –4.1 to –3.7; P < .001). BP documentation was higher for appointments completed pre– vs post pandemic onset (16.2 PP; 95% CI, 16.0-16.5; P < .001) and for appointments completed in person vs VC (54.9 PP; 95% CI, 54.6-55.2; P < .001). HbA_1c documentation was higher for completed appointments after pandemic onset vs before (5.9 PP; 95% CI, 5.1-6.7; P < .001) and for completed VC appointments vs in-person appointments (3.9 PP; 95% CI, 3.0-4.7; P < .001).

Conclusions: After pandemic onset, appointment completion rates were higher, time to appointment was shorter, HbA_1c documentation increased, and BP documentation decreased. Future research should explore the advantages of using VC for CVD risk management while continuing to monitor for unintended consequences.

Am J Manag Care. 2024;30(1):43-48. https://doi.org/10.37765/ajmc.2024.89485

Takeaway Points

Using virtual care (VC) for cardiovascular disease risk management may have several positive consequences in terms of care processes, including reduced appointment wait times and increased completion rates. Concurrently, however, there may be negative consequences (eg, reduced blood pressure documentation) that should be addressed.

• This is one of the first studies on how care processes differ between VC and in-person care in the community clinic setting.
• Findings have clinical and policy relevance by indicating the need for measures that optimize the positive outcomes of VC (such as increased access for some patients) and minimize potential negative outcomes (such as reduced reporting of biomarkers).

Cardiovascular disease (CVD) prevention is a major component of primary care, with clinical practice guidelines emphasizing the importance of monitoring and controlling multiple modifiable risk factors, including blood pressure (BP) and hemoglobin A_1c (HbA_1c) level.¹ The COVID-19 pandemic disrupted the delivery of primary care, and many health care organizations quickly moved to virtual care (VC) as a mode of care delivery.² This shift provided an opportunity to investigate how pandemic-driven changes in primary care delivery affected care process measures.³

We leveraged data from a cluster randomized trial evaluating the impact of a clinical decision support and shared decision-making tool on reversible CVD risk score and risk factor control in primary care community-based health clinics.⁴ Using these data, the presented analyses assessed how pandemic-driven changes in care delivery mode affected chronic disease risk management care processes. Specifically, we (1) describe patterns of scheduled and completed appointments before and after pandemic onset for both in-person and VC appointments and (2) assess whether appointment type was associated with appointment completion rates, time to appointment, and documentation of key risk factors (BP, HbA_1c) at completed appointments in the pre– and post–pandemic onset periods.

METHODS

Setting and Data Source

This retrospective study used data from OCHIN, a nonprofit health innovation organization that provides a single instance of the Epic electronic health record (EHR) system to community-based health clinics across the country. Its member clinics include more than 1000 sites across 45 states that provide care for more than 6 million patients.⁵ Data from this shared EHR (eg, patient demographics, insurance at appointment, patient addresses, prescriptions, diagnoses, laboratory results, and vitals) are collected and standardized, resulting in robust, longitudinal data. When the pandemic began, OCHIN supported all interested members in rapidly expanding their capacity to provide VC.

Study Sample

All eligible scheduled primary care appointments between September 1, 2018, and September 30, 2021, at the 70 primary care clinics taking part in a cluster randomized trial assessing the impact of CV Wizard—a clinical decision support and shared decision-making tool—on reversible CVD risk were included. The details of this trial are presented elsewhere,⁴ but in brief, CV Wizard is a web-based tool that processes EHR data, calculates a 10-year reversible CVD risk estimate based on those data, and then lists risk factors and evidence-based care recommendations in the order they should be addressed to reduce risk (ie, factors with the highest potential for reducing risk if controlled are prioritized in the tool’s output). In this trial, CV Wizard was activated in 42 intervention clinics in September 2018, and the remaining 28 control clinics gained access to the tool in September 2020. Although CV Wizard use was a focus in the main trial, it is not central to the analyses presented here.

Eligible appointments (1) were scheduled to occur during the study period; (2) involved patients aged 40 to 75 years at the time of the appointment with a diagnosis of hypertension and/or type 2 diabetes during the study period; (3) had an appointment outcome of completed, no-show, or canceled; and (4) were coded as an appointment type at which the delivery of CVD/diabetes care was possible (see eAppendix Table 1 [eAppendix available at ajmc.com]). Diagnosis codes to identify hypertension and diabetes were selected based on the Charlson Comorbidity Index (see eAppendix Table 2).⁶

Outcomes

Four outcomes were selected based on care process measures of interest: appointment completion, time to appointment, up-to-date BP documentation, and up-to-date HbA_1c documentation. Scheduled appointments were classified as completed or not completed; those categorized as not completed included appointments canceled by the patient, those canceled by the provider, and no-shows. Canceled appointments were excluded if the reason for cancellation was listed as “deceased” or “error” in the EHR. Time to appointment was calculated as the number of days from the date the appointment was scheduled to the date of the appointment. Up-to-date documentation of BP was determined for all completed appointments. Appointments were considered to have up-to-date BP documentation if a BP reading was entered within a 21-day window: 7 days before through 14 days after the appointment was completed. This entry window accounted for providers requesting BP readings prior to a VC appointment, patients reporting home BP data after an appointment, and providers documenting readings after the appointment. Up-to-date documentation of HbA_1c was identified for completed appointments among patients after a diabetes diagnosis was recorded. All HbA_1c data came from laboratory results posted in the EHR; self-reported HbA_1c results were not captured in these analyses. Up-to-date documentation was considered “needed” for those with no recorded HbA_1c in the 6 months before the appointment. Patients who (1) were classified as needing updated results and (2) had a new HbA_1c laboratory result within 14 days of the completed appointment were considered to have up-to-date documentation for the appointment.

Independent Variables

Appointment type. Scheduled appointments were categorized as either in person or VC based on delivery mode. To include appointments that were not completed, we used scheduled appointment data to identify VC appointments (eg, audio, video, synchronous, asynchronous), and all other appointments were classified as in person. To validate this categorization, completed appointments were linked to Epic appointment designations and billing data; we found 94% agreement for VC appointments and 99% agreement for in-person appointments.

Time period. Scheduled appointments were categorized as either pre– or post pandemic onset. Appointments scheduled to occur between September 1, 2018, and February 28, 2020, were considered pre–pandemic onset, and those scheduled to occur between April 1, 2020, and September 30, 2021, were considered post pandemic onset. To account for rolling changes in stay-at-home orders and the time needed for the study clinics to implement the use of VC, March 2020 was designated as a washout period for regression analyses.

Potential confounders that were considered in the analysis included age at appointment, sex (female or male), race (White, Black, Asian, all other races, or missing or unknown race), ethnicity (Hispanic, non-Hispanic, or missing or unknown ethnicity), insurance at the appointment (Medicaid, Medicare, private, uninsured, or other insurance), number of scheduled appointments during the study period, rurality of patient residence (urban [population of ≥ 50,000 and 30%-49% commuting flow to an urban area], large rural [population 10,000-49,999 with < 30% commuting flow to an urban area], or small rural [population 2500-9999 with < 30% commuting flow to an urban area]),⁷ and chronic conditions documented in the EHR at the time of the appointment (hypertension only, diabetes only, or hypertension and diabetes). Models estimating associations with appointment completion also adjusted for number of days from when the appointment was scheduled to when it occurred.

Statistical Methods

Descriptive analyses included all eligible scheduled appointments at study clinics during the study period (including the 1-month washout period imposed for regression analyses). We examined changes in the number of scheduled appointments by describing the total number of monthly scheduled appointments during the study period as well as the number of monthly appointments stratified by appointment type, appointment completion, and appointment type by appointment completion. Appointment-level demographics for all scheduled appointments were reviewed. The clinics included in the descriptive analyses are in the following census regions: Midwest (n = 5; 7.1%), Northeast (n = 1; 1.4%), South (n = 11; 15.7%), and West (n = 53; 75.7%)

Regression models excluded appointments from clinics that had 10% or less of all study-eligible appointments conducted via VC (n = 9) or less than 50 VC appointments after pandemic onset (n = 2) due to their very low VC adoption rates. A total of 11 clinics were excluded, resulting in data from 59 clinics in the final models; descriptive analyses comparing included and excluded clinics are presented in eAppendix Table 3. Additionally, appointments from the 1-month washout period were excluded. The clinics included in the regression analyses are in the following census regions: Midwest (n = 5; 8.5%), Northeast (n = 1; 1.7%), South (n = 5; 8.5%), and West (n = 48; 81.4%).

To examine the association between appointment type (in person vs VC) and time period (pre– vs post pandemic onset) for each process measure outcome, multilevel logistic and linear regression models were used. These models contain fixed and random effects as well as an unstructured covariance structure. Average marginal effects were obtained, and the models estimated associations between appointment type and time period with (1) appointment completion, (2) BP documentation, and (3) HbA_1c documentation. Multilevel linear regression was used to estimate associations between appointment type and time period with time to appointment. To adjust for confounders, all models included a random effect for patient and adjusted for age at appointment, sex, race, ethnicity, insurance at the appointment, number of scheduled appointments during the study period, rurality of patient residence, and chronic conditions documented in the EHR at the time of the appointment. A clinic-level fixed effect was included due to models not converging when a random effect for clinic was included.

SAS Enterprise Guide 8.3 (SAS Institute Inc) and SQL Server Management Studio (Microsoft) were used for all data extraction, cleaning, and descriptive results. Stata 15.1 (StataCorp LLC) was used for all modeling.

Ethics Approval

This study (project No. 1394220) was approved by the Kaiser Permanente Interregional Institutional Review Board.

RESULTS

Scheduled Appointment Characteristics

There were 1,179,542 scheduled appointments during the 3-year study period. Of those, 74.8% (882,110) were scheduled as in-person appointments, 69.2% (816,603) were completed, and the mean (SD) number of days between when the appointment was made and when it was scheduled to occur was 24 (41.8) days. See Table 1; additional information is presented in the Descriptive Results subsection.

Participant Characteristics

Of the eligible appointments, 57.4% (677,092) were for female patients, 24.0% (283,583) were for individuals of Hispanic ethnicity, 68.2% (804,799) were for White individuals, and 74.5% (878,370) were for those with English listed as their primary language. Additionally, 85.1% (1,003,456) were for individuals with insurance coverage, 62.2% (733,726) were for patients living in urban areas, and 40.7% (479,491) were for those with both type 2 diabetes and hypertension (Table 1).

Descriptive Results

The number of scheduled appointments was stable across both time periods (eAppendix Figure). Most appointments prior to pandemic onset were scheduled to occur in person (Figure 1). After the pandemic onset in March 2020, this pattern shifted. From April to September 2020, most appointments were scheduled to take place virtually; between October 2020 and February 2021, scheduled appointment type varied by month; and starting in March 2021, the rate of appointments scheduled as in person exceeded the rate scheduled as virtual, a pattern that continued through the end of the study period.

Figure 2 displays monthly rates of completed appointments over the study period, both overall and by appointment type. The rate of completed appointments was stable prior to pandemic onset, with a mean completed appointment rate of 66.8% for September 2018 through February 2020. This rate increased after the onset of the pandemic to a mean completion rate of 72.1% between April 2020 and September 2021.

Prior to the pandemic, the mean completion rate was 66.7% for in-person appointments and 73.1% for VC appointments. After pandemic onset, the mean completion rates were 66.1% and 78.6%, respectively. Mean time to appointment before the pandemic was 26.1 days for in-person appointments and 14.9 days for VC appointments. After pandemic onset, the mean times to appointment were 28.0 days and 15.6 days, respectively. BP documentation was updated in 95.5% of in-person appointments before the pandemic; after pandemic onset, it was 86.0% for in-person appointments and 21.2% for VC appointments. HbA_1c documentation was updated in 14.4% of in-person appointments before the pandemic; after the pandemic, it was 8.7% for in-person appointments and 20.0% for VC appointments.

Regression Results

The average marginal effects from the regression models are summarized in Table 2. Completion rate was higher for (1) appointments scheduled before pandemic onset compared with those scheduled after (1.1 percentage points [PP]; 95% CI, 0.9-1.3; P < .001) and (2) VC compared with in-person appointments (10.7 PP; 95% CI, 10.5-11.0; P < .001). Time to appointment for scheduled appointments was shorter (1) after pandemic onset compared with before (–2.9 days; 95% CI, –3.0 to –2.6; P < .001) and (2) for VC compared with in-person appointments (–3.9 days; 95% CI, –4.1 to –3.7; P < .001). Updated BP documentation was significantly higher for (1) completed appointments before pandemic onset compared with after (16.2 PP; 95% CI, 16.0-16.5; P < .001) and (2) appointments completed in-person compared with VC appointments (54.9 PP; 95% CI, 54.6-55.2; P < .001). Updated HbA_1c documentation was significantly higher for (1) completed appointments after pandemic onset compared with before (5.9 PP; 95% CI, 5.1-6.7; P < .001) and (2) completed VC appointments compared with in person (3.9 PP; 95% CI, 3.0-4.7; P < .001).

DISCUSSION

The findings of this study demonstrate the relationships among the COVID-19 pandemic, care process measures, and the management of select CVD risk factors in community-based clinics. Results show that although the mean number of scheduled appointments per month did not change after the onset of the pandemic, a finding observed in other settings,⁸ the completion rate of scheduled appointments increased overall in the post–pandemic onset period. Additionally, the number of scheduled VC appointments increased after pandemic onset, and although the magnitude of this increase diminished over time, rates remained much higher than before the pandemic.

When appointment type was accounted for, appointments scheduled before the pandemic were more likely to be completed than those scheduled after pandemic onset. This suggests that the overall increase in completion rates seen in the descriptive analysis for that period is driven by the high completion rate for VC appointments. This finding is consistent with other studies showing higher appointment completion rates and lower no-show rates for VC appointments.^9,10 Furthermore, VC addresses some of the reasons why in-person appointments may not be completed (eg, illness, COVID-19 exposure concerns, transportation barriers).

The time between when an appointment was made and when it was scheduled to occur was shorter for VC appointments and those scheduled after pandemic onset. This is likely because the increase in VC appointments post pandemic meant wait time for an appointment was shorter for VC appointments. These results are consistent with previous research¹¹ and suggest that offering both VC and in-person care options may improve clinic operational efficiency and patient experience; more research is needed to explore how to optimize the use of varying care modality options.

BP risk factor documentation was more likely to be updated for completed appointments before pandemic onset compared with after and for in-person compared with VC appointments. This is likely because almost all prepandemic appointments were conducted in person, during which it is standard to obtain BP readings. This concerning potential consequence of VC indicates the need for further study of effective strategies to improve patient self-reporting of BP and use of home BP measurement to augment VC.

In contrast, HbA_1c risk factor documentation was more likely to be updated for completed appointments after pandemic onset and for completed VC appointments. This finding is consistent with recent research suggesting that VC facilitates care continuity, which resulted in increased HbA_1c testing.¹² Another potential explanation is that HbA_1c monitoring may have been more vigilant in the post–pandemic onset period given the association between type 2 diabetes and COVID-19 severity¹³ or possibly due to increased patient motivation and engagement in self-care.¹⁴ The increased likelihood of HbA_1c documentation for VC appointments could also be related to an associated decrease in burden or barrier removal. For example, in-person primary care appointments may require additional time or travel to separate locations (1 trip for the appointment and 1 for the laboratory), whereas VC appointments may only require 1 trip for the patient (laboratory only).

Limitations

Analyses included only data from community-based clinics participating in the parent study, which may limit the generalizability of these findings. Furthermore, clinics included in the regression analyses were limited to those meeting a specific threshold for VC appointments. Because VC appointments occurred in only a handful of study clinics before pandemic onset and for different reasons from those that occurred after onset, VC appointments before and after pandemic onset could not be meaningfully compared; therefore, we did not explore the interaction between appointment type and time period in the regression analyses. This study was limited to select process measures of interest; additional research is needed to assess the degree of BP and HbA_1c control associated with appointments in different time periods and VC vs in-person care. It is possible that other factors (eg, disease severity, CV Wizard tool use, the need for BP and/or HbA_1c data, new vs established patient status) influenced scheduled appointment type. Finally, documentation in the EHR for scheduled appointments can vary by clinic, and it is possible that some appointment cancellations were misclassified as clinics shifted visits from in person to VC.

CONCLUSIONS

The findings show an upside of the COVID-19 pandemic in stimulating the use of VC appointments, which were associated with an increase in appointment completion, a decrease in time to appointment, and an increase in updated HbA_1c documentation. The potential advantages of VC for CVD risk management appear worthy of continued exploration and enhancement while also monitoring for unintended negative consequences (eg, reduced BP documentation). These findings also show that community-based health clinics were able to nimbly navigate the shift to VC and maintain care for patients needing chronic disease management following the onset of the pandemic.

Acknowledgments

The authors would like to thank Nadia Yosuf, MPH, and Jenny Hauschildt, MPH, for their contributions to this project. This work was conducted with the Accelerating Data Value Across a National Community Health Center Network (ADVANCE) Clinical Research Network. OCHIN leads the ADVANCE network in partnership with Health Choice Network, Fenway Health, and Oregon Health & Science University. ADVANCE is funded through the Patient-Centered Outcomes Research Institute, contract No. RI-CRN-2020-001.

Author Affiliations: Kaiser Permanente Center for Health Research (CRS, RG), Portland, OR; OCHIN (AEL, DB, NC, BMM, RG), Portland, OR; HealthPartners Institute (PJO), Minneapolis, MN; Case Western Reserve University (KCS), Cleveland, OH.

Source of Funding: Research reported in this publication was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health (NIH) under award No. R01HL133793 and the National Institute on Minority Health and Health Disparities of the NIH under award No. R01MD016389. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Author Disclosures: Dr O’Connor reported receiving research funding from NIH. Dr Stange reported being a paid affiliate investigator with OCHIN and being a principal investigator on 2 related NIH grants. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (CRS, AEL, DB, PJO, NC, RG); analysis and interpretation of data (CRS, AEL, DB, NC, BMM, KCS, RG); drafting of the manuscript (CRS, AEL, DB, RG); critical revision of the manuscript for important intellectual content (CRS, AEL, PJO, NC, BMM, KCS, RG); statistical analysis (AEL, BMM); obtaining funding (PJO, RG); administrative, technical, or logistic support (CRS, RG); supervision (RG); and linkage with a related study (KCS).

Address Correspondence to: Rachel Gold, PhD, MPH, Kaiser Permanente Center for Health Research, 3800 N Interstate Ave, Portland, OR 97227. Email: Rachel.gold@kpchr.org.

REFERENCES

1. Arnett DK, Blumenthal RS, Albert MA, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: a report of the American College of Cardiology/American Heart Association Task Force on clinical practice guidelines. Circulation. 2019;140(11):e596-e646. doi:10.1161/CIR.0000000000000678

2. The adoption of telehealth. American Telemedicine Association. May 10, 2021. Accessed February 8, 2023. https://www.americantelemed.org/wp-content/uploads/2021/05/Adoption-of-Telehealth.pdf

3. Types of health care quality measures. Agency for Healthcare Research and Quality. Updated July 2015. Accessed March 22, 2023. https://www.ahrq.gov/talkingquality/measures/types.html

4. Gold R, Larson AE, Sperl-Hillen JM, et al. Effect of clinical decision support at community health centers on the risk of cardiovascular disease: a cluster randomized clinical trial. JAMA Netw Open. 2022;5(2):e2146519. doi:10.1001/jamanetworkopen.2021.46519

5. OCHIN. Accessed December 20, 2022. https://ochin.org/

6. Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol. 1994;47(11):1245-1251. doi:10.1016/0895-4356(94)90129-5

7. RUCA data: using RUCA data. Rural Health Research Center. Accessed July 31, 2023. https://depts.washington.edu/uwruca/ruca-uses.php

8. Tu K, Sarkadi Kristiansson R, Gronsbell J, et al; INTRePID. Changes in primary care visits arising from the COVID-19 pandemic: an international comparative study by the International Consortium of Primary Care Big Data Researchers (INTRePID). BMJ Open. 2022;12(5):e059130. doi:10.1136/bmjopen-2021-059130

9. Haggerty T, Stephens HM, Peckens SA, et al. Telemedicine versus in-person primary care: impact on visit completion rate in a rural Appalachian population. J Am Board Fam Med. 2022;35(3):475-484. doi:10.3122/jabfm.2022.03.210518

10. Drerup B, Espenschied J, Wiedemer J, Hamilton L. Reduced no-show rates and sustained patient satisfaction of telehealth during the COVID-19 pandemic. Telemed J E Health. 2021;27(12):1409-1415. doi:10.1089/tmj.2021.0002

11. Mishra K, Edwards B. Cardiac outpatient care in a digital age: remote cardiology clinic visits in the era of COVID-19. Curr Cardiol Rep. 2022;24(1):1-6. doi:10.1007/s11886-021-01618-9

12. Tierney AA, Payán DD, Brown TT, Aguilera A, Shortell SM, Rodriguez HP. Telehealth use, care continuity, and quality: diabetes and hypertension care in community health centers before and during the COVID-19 pandemic. Med Care. 2023;61(suppl 1):S62-S69. doi:10.1097/MLR.0000000000001811

13. Norouzi M, Norouzi S, Ruggiero A, et al. Type-2 diabetes as a risk factor for severe COVID-19 infection. Microorganisms. 2021;9(6):1211. doi:10.3390/microorganisms9061211

14. Duffy E, Chilazi M, Cainzos-Achirica M, Michos ED. Cardiovascular disease prevention during the COVID-19 pandemic: lessons learned and future opportunities. Methodist Debakey Cardiovasc J. 2021;17(4):68-78. doi:10.14797/mdcvj.210