Characteristics of Patients Using Online Asynchronous E-Visits for Low-Acuity Conditions

Many retail health clinics, insurers, and concierge providers are now offering virtual health services to consumers. Increased choice could mean less incentive for patients to seek ambulatory care for low-acuity conditions at academic medical centers (AMCs) or other established ambulatory practices, leading to loss of business and reduced patient loyalty. The business of virtual health is growing, spanning a continuum of services that promises increased access and efficiency.(1) Patient preference studies have shown that patients are receptive to the idea of virtual healthcare, citing benefits such as convenience and lower costs.(2,3) Practices can address consumer preferences, become more competitive, and achieve potential cost savings through implementation of virtual care.

Virtual health programs can achieve cost avoidance by replacing costly, in-person, low-acuity visits with lower-cost virtual visits. For example, insurer Anthem, Inc. found that in-person retail health clinic visits, urgent care visits, office visits, and emergency department visits were, on average, $36 to $1735 more expensive than virtual visits.(4) At Mayo Clinic Rochester, it was found that e-visits can lower costs at a family medicine clinic by $44 per visit by replacing same-day acute visits with online visits.(5) Finally, HealthPartners saved an average of $88 per encounter using online visits compared with in-person care.(6) It is noteworthy that cost avoidance may be achieved only if virtual care visits are used as substitutions for in-person visits.

Previous literature has demonstrated that e-visit programs at AMCs can offer improved efficiency, access, and continuity of care.(7) Although research comparing quality among virtual health providers is limited, some evidence shows that variation in quality of care between virtual health providers may be similar to variations in quality that exist between in-person providers within traditional care settings.(8) Other studies assessing the quality of virtual health programs have focused on metrics such as antibiotic prescription rates and need for follow-up visits.

Researchers found a higher likelihood to prescribe antibiotics for urinary tract infection when conducting an e-visit.(9) For some patients, the convenience of e-visits may lead to increased utilization instead of the desired substitution for in-person visits. For example, a large payer study with Teladoc and Blue Cross Blue Shield found that net annual spending on acute respiratory illness increased by $45 per telehealth user.(10) Furthermore, Kaiser Permanente found that patients who used a patient portal also had more visits to a provider and more telephone encounters compared with those who did not access the portal.(11) On the other hand, Mayo Clinic found that although 34% of e-visit patients had an additional interaction with a provider within 30 days of the e-visit for the same health concern, most e-visits were completed without further interaction with the healthcare system.(12) At the Medical University of South Carolina, 92% of e-visits replaced in-person visits.(13) Future research is needed to understand these trends across institutions, noting that there are opportunities to replace costly low-acuity in-person visits while recognizing that some patients may increase utilization.

Researchers have begun to define the patient characteristics of e-visit users. Findings suggest early adopters tend to be younger, white, and female.(13-15) The representation of female patients using e-visits may be explained by the high use of e-visits for female-specific conditions (e.g., urinary tract infections and yeast infections) and because women are more likely to access primary care.(6) It also has been shown that higher engagement with a patient portal increases the patient’s likelihood to use an e-visit.(15) Time is also an important factor in e-visit use: research at an AMC found that younger patients were more likely to use e-visits for reasons of convenience.(16) Importantly, patients who are educated about virtual care may be more likely to use it. An employee health study found that among all employees offered access to a virtual care program, those who were 30 to 49 years of age and attended an orientation to the program were more likely to use virtual health services.(17) Furthermore, CVS Health found that a patient’s understanding of telehealth services was a significant predictor of likelihood to rate high satisfaction with a telehealth visit.(3) These findings underscore the importance of marketing and outreach for the growth of virtual care programs.

In May 2018, a large Midwestern AMC launched an online asynchronous e-visit platform for ambulatory care patients through the existing online patient portal. Nine low-acuity conditions were deemed eligible for treatment via an e-visit: 1) back pain; 2) cough or cold; 3) headache; 4) red eye; 5) heartburn; 6) sinus issues; 7) diarrhea; 8) painful urination; and 9) vaginal irritation. Patients eligible for this new service must have had at least one outpatient visit within the previous three years, and have an active online patient portal account. An e-visit is initiated through the completion of a condition-specific questionnaire through the patient portal. Results are reviewed by a clinician to assess clinical appropriateness to receive treatment virtually.

Previous literature on e-visits and virtual health largely addresses operations and quality, focusing on cost savings and patient safety. Asynchronous e-visits are still a new and understudied service. Practice administrators may benefit from understanding which patients are early adopters of e-visits to target marketing. Our study hypothesized that demographics and utilization history can predict likelihood to choose an e-visit over an office visit for treatment of low-acuity conditions.

Methods

Study Sample and Data Collection

This was a retrospective, cross-sectional study design. Data were collected from patients accessing care for low-acuity conditions at a large Midwestern AMC. Sample selection was based on four criteria: time period; ICD-10 code; patient age; and encounter type (Figure 1). The study period included the first eight months after the launch of the e-visit program, from May 1, 2018, to December 31, 2018. Encounters were included if patients were treated for any of the nine conditions eligible for an e-visit, as determined by ICD-10 codes. Patients aged 18 years and older were considered eligible for study participation. Finally, patients who had office visits had to be seen at one of fifteen primary care clinics. A four-year history (January 1, 2015—December 31, 2018) of telephone encounters, patient portal secure messaging, inpatient admissions, and office visits for each study participant also was collected. Telephone encounter history excluded administrative calls. Eight eligible cases were excluded due to missing data.

Figure 1. Eligibility consort diagram.

Demographics and utilization metrics were summarized using descriptive statistics. Analysis of variance (ANOVA) and Pearson chi-square (χ-squared) tests were used to determine bivariate associations between patient characteristics and care pathway (e.g., office visit only, e-visit only, or both during the eight-month study period). A multinomial logistic regression was used to test the association between patient characteristics and care pathway. The study was approved by the institutional review board.

Results

Study Sample Description

Based on ICD-10 codes and visit location, 5826 patients were identified for inclusion in the study. Of these, 334 patients had an e-visit(s) only, 5432 had an office visit(s) only, and 60 patients had both office and e-visit(s). The average age was 47 years old. The e-visit only and both visits categories had a younger average age, of 41 and 38, respectively. The overall sample was 72% female. Similarly, those that used e-visits, or had both office and e-visits, were predominantly female (87% and 80%, respectively). The overall sample was 44% white, 34% black or African-American, 4% Asian, and 18% other. The e-visit only and both e-visit and office visit categories were disproportionately white, at 65% and 58%, respectively. Patients sent an average of 6 secure patient portal messages, made 17 clinical phone calls, and had 16 office visits to an outpatient clinic between 2015 and 2018. The average number of inpatient admissions from 2015 to 2018 among all patients was less than one, suggesting a low-acuity sample. Descriptive summaries can be found in Table 1.

Results of Analysis

Bivariate results are found in Table 1. Age (F 2, 5283 = 31.002, p < .001), patient portal messaging (F 2, 5823 = 38.45, p < .001), telephone encounters (F 2, 5823 = 21.76, p < .001), inpatient admissions (F 2, 5823 = 5.57, p = .004), number of office visits (F 2, 5823 = 17.85, p < .001) and sex (χ-squared (2) = 40.409, p < .001) and race (χ-squared (6) = 75.308, p < .001) were all significantly associated with care pathway. Follow-up pairwise comparisons using Tukey’s HSD (honestly significant difference) were done on all one-way ANOVAs. E-visit use was associated with a younger age, being female, having fewer telephone encounters, fewer office visits, and fewer hospital admissions. All pairwise comparison results can be found in Table 1.

Multivariate results can be found in Table 2 and Figure 2. Our overall multinomial regression model was significant (x-squared (22) = 329.691, p < .001). Comparing e-visit only to office visit only, age, patient portal message frequency, and telephone encounter frequency were found to be statistically significant predictors. For each one-year increase in age, the patient has 0.988 less odds to use an e-visit compared with an office visit (OR = 0.988, p = .002). Compared with males, females were found to have 3.161 times higher odds of using an e-visit (OR = 3.161, p <.001). Patients who identified as black or African-American, Asian, or other race were significantly less likely than white patients to use an e-visit as compared with an office visit (OR = 0.397, p < .001; OR = 0.561, p = .046; OR = 0.600, p = .017, respectively). Hispanic or Latino participants had 0.672 less odds than non-Hispanic or Latino participants to use an e-visit compared with an office visit (OR = 0.672, p = .039). For each additional patient portal message, the patient has 1.034 more odds to use an e-visit compared with an office visit (OR = 1.034, p <.001). For each additional telephone encounter, the patient has 0.960 less odds to use an e-visit compared with an office visit (OR = 0.960, p <.001).

Figure 2. Care pathway.

In the both vs. office visit only analysis, age, race, patient portal messages, telephone encounters, and office visits were found to be significant predictors. For each one-year increase in age, participants have 0.967 less odds to be in the both category compared to office visit only (OR = 0.967, p <.001). Black or African-American participants have 0.408 less odds than white participants to use both an e-visit and office visit compared with an office visit only (OR = 0.408, p = .018). No other race or ethnicity associations were found to be statistically significant in the both vs. office visit only analysis. For each additional patient portal message, the patient has 1.043 times more odds of using an e-visit and office visit compared with an office visit only (OR = 1.043, p < 001). For each additional telephone encounter, the patient has 1.029 times more odds of falling in the both category compared with office visit only (OR = 1.029, p < .001). Finally, as office visits increase, participants have 0.938 less odds of falling in the both category as opposed to office visit only (OR = 0.938, p <.001).

Discussion

This study found that use of an e-visit platform for low-acuity conditions can be predicted based on age, sex, race, ethnicity, history of patient portal messages, history of telephone encounters, and, in some cases, history of office visits. Our findings related to younger age, white race, and female sex predicting e-visit use are supported by existing literature on demographics of e-visit program users.(13,14,16) Our findings related to secure patient portal messages and telephone encounters as predictors of e-visit use contribute to the literature on utilization behavior of e-visit users. Increased patient portal messaging was associated with an increased likelihood of using an e-visit, which is consistent with findings at a health system in Pennsylvania.(15) It may be that patients already highly engaged in a patient portal are more comfortable contacting their provider remotely through virtual methods, indicating an increased likelihood to use an e-visit. Telephone encounters were significant, but in the opposite direction. This finding suggests that e-visit only patients show a stronger preference for digital methods when contacting their provider remotely as compared to the participants who accessed both e-visit and office visits, or office visits only where increased telephone encounters were used. This suggests a market of consumers who will consistently choose digital methods over telephone or in-person contact when given the choice. A previous study from Kaiser Permanente found that patients highly engaged with a patient portal are more likely to have increased office visits.(11) We found that a history of office visits was not significant in the e-visit only analysis and was negatively associated for those with both an e-visit and office visit. Our findings may be explained by the limitations in our data, as all patients were already established patients and therefore somewhat already accustomed to coming to the clinic.

Limitations for this study stem from the nature of a new program, as the organization’s e-visit program is still in its infancy. The e-visit program launched in May 2018 and was offered only to established patients for treatment of nine conditions. Because all patients in the sample were already engaged in care, the findings may not be applicable to e-visit programs that are accepting new patients, so the research may not be entirely generalizable. Future studies should consider a sample of new and existing patients to help e-visit programs better understand adopters of these programs.

This study examined demographic and utilization characteristics of established patients to guide marketing efforts for an e-visit program. A previous large-scale payer study anticipated that making care too convenient might create more demand and lead to increased cost.(10) However, academic medical centers have expressed interest in e-visits to replace in-person low-acuity visits with lower-cost e-visits.(5,9) Results from this study suggest that ambulatory settings looking to grow e-visit programs among established patients to achieve cost avoidance should anticipate early adopters who are younger, female, and non-Hispanic white who display high engagement with a patient portal but make fewer clinical phone calls. Future studies may consider investigation of demographic and utilization characteristics of new patients using an e-visit program to help AMCs and ambulatory care settings remain competitive among patients not already engaged in care.

As asynchronous e-visits often are unscheduled, practices will need to prepare for managing unpredictable demand.

For AMCs and other ambulatory practices interested in growing volumes of established patients in an e-visit program as a replacement for office visits, the e-visit only versus office visit only results suggest that marketing efforts should anticipate early patient adopters who are younger, female, of white and non-Hispanic descent, more engaged in a patient portal, and with fewer telephone encounters. In addition to marketing, e-visit programs also must consider management implications related to operations and provider satisfaction. As asynchronous e-visits often are unscheduled, practices will need to prepare for managing unpredictable demand. E-visit services often are open during evenings and weekends, further affecting provider schedules. More time spent by providers on e-visits also means a loss of human connection in daily clinical practice. Managers must maintain an open dialogue with providers about their engagement in virtual health programs to maintain provider satisfaction. In regard to policy implications, the future of reimbursement for direct-to-consumer telehealth will influence demographics of patients accessing the service.(18) For example, if payment models become more favorable to asynchronous e-visits, the financial assistance with e-visit co-pays may enable new demographics of patients for whom the self-pay fee is prohibitive to access the service. All in all, ambulatory care practices that are considering offering e-visits, either synchronous or a-synchronous, must consider the demographic, payer, and medical characteristics of their patient panels to better understand who may use and benefit from such offerings.

Acknowledgment: The research team thanks Justine Humber, Project Manager in the Office of Transformation at Rush University Medical Center, for her role in this project.

References

1. Mehrotra A. The convenience revolution for treatment of low-acuity conditions. JAMA. 2013;310:35-36.

2. Powell RE, Henstenburg JM, Cooper G, Hollander JE, Rising KL. patient perceptions of telehealth primary care video visits. Ann Fam Med. 2017;15:225-229.

3. Polinski J, Barker T, Gagliano N, Sussman A, Brennan T, Shrank W. Patients’ satisfaction with and preference for telehealth visits. J Gen Intern Med. 2015;31:269-275.

4. Gordon AS, Adamson WC, DeVries AR. Virtual visits for acute, nonurgent care: a claims analysis of episode-level utilization. J Med Internet Res. 2017;19(2):e35.

5. Rohrer JE, Angstman KB, Adamson SC, Bernard ME, Bachman JW, Morgan ME. Impact of online primary care visits on standard costs: a pilot study. Popul Health Manag. 2010;13(2):59-63.

6. Courneya PT, Palattao KJ, Gallagher JM. HealthPartners’ online clinic for simple conditions delivers savings of $88 per episode and high patient approval. Health Aff. 2013;32:385-392.

7. Hickson R, Talbert J, Thornbury WC, Perin NR, Goodin AJ. Online medical care: the current state of ‘‘evisits’’ in acute primary care delivery. Telemed J E Health. 2015;21(2):90-96.

8. Schoenfeld AJ, Davies JM, Marafino BJ, et al. Variation in quality of urgent health care provided during commercial virtual visits. JAMA Intern Med. 2016;176:635-642.

9. Mehrotra A, Paone S, Martich GD, Albert SM, Shevchik GJ. A Comparison of care at e-visits and physician office visits for sinusitis and urinary tract infection. JAMA Intern Med. 2013;173:72-74.

10. Ashwood JS, Mehrotra A, Cowling D, Uscher-Pines L. Direct-to-consumer telehealth may increase access to care but does not decrease spending. Health Aff. 2017;36:485-491.

11. Palen TE, Ross C, Powers JD, Xu S. Association of online patient access to clinicians and medical records with use of clinical services. JAMA. 2012;308:2012-2019.

12. Penza K, Murray M, Pecina J, Myers J, Furst J. Electronic visits for minor acute illnesses: analysis of patient demographics, prescription rates, and follow-up care within an asynchronous text-based online visit. Telemed J E Health. 2018;24:210-215.

13. Player M, O’Bryan E, Sederstrom E, Pinckney J, Diaz V. Electronic visits for common acute conditions: evaluation of a recently established program. Health Aff. 2018;37:2024-2030.

14. Jung C, Padman R. Virtualized healthcare delivery: understanding users and their usage patterns of online medical consultations. Int J Med Inform. 2014;83:901-914.

15. Jung C, Padman R, Shevchik G, Paone S. Who are portal users vs. early e-visit adopters? A preliminary analysis. AMIA Annu Symp Proc. 2011;2011:1070-1079.

16. Mehrotra A, Paone S, Martich GD, Albert SM, Shevchik GJ. Characteristics of patients who seek care via evisits instead of office visits. Telemed J E Health. 2013;19:515-519.

17. Edgerton S. A pilot study investigating employee utilization of corporate telehealth services. Perspect Health Inf Manag. 2017;14(Fall):1g.

18. Shrank W. Primary care practice transformation and the rise of consumerism. J Gen Intern Med. 2017;32:387-391.