Peter J. Kaboli, MD, MS

 To evaluate the growth and variation of electronic consultation, or e-consult, use in the Veterans Health Administration (VHA) across regions and specialties.

 Observational cohort study using administrative data of all veterans who received an e-consult for 41 specialties across 1269 VHA medical centers and associated clinical sites from January 1, 2012, through December 31, 2018.

 Assessments included (1) the number and characteristics of all e-consults, (2) growth of e-consult use, (3) e-consults as a proportion of all consults by region and by specific specialty, (4) need for an in-person visit with the same specialty within 12 months after an e-consult, and (5) potential miles of driving saved for patients and mileage reimbursement costs avoided for VHA due to e-consult use.

 Over the 7-year study period, VHA providers completed 3,117,998 e-consults (5.5% of all specialty consults). e-Consults increased by 309% for all specialties. By 2018, for 16 of 41 specialties, e-consults accounted for greater than 10% of all consults. Overall, 21.5% of e-consults resulted in an in-person visit with the same specialty within 12 months. On average, each e-consult resulted in approximately 84.3 (SD, 89.9; interquartile range, 25.1-115.0) miles in driving saved, equating to potential driving reimbursement savings of $46 million.

 Use of e-consults in the VHA grew substantially between 2012 and 2018, with variability across specialties. In-person follow-up after an e-consult was low, suggesting that e-consults may substitute for in-person visits and reduce considerable patient travel burden.

Use of electronic consultations, or e-consults, has been implemented in many medical systems in the United States, but further evaluation of its impact on care is needed.

This work evaluates e-consult use within the Veterans Health Administration, the nation’s largest health care system. This the largest cohort of e-consults for which results are available.

For many medical and surgical subspecialties, more than 10% of total consults are now completed as e-consults. Substantial e-consult growth was seen across regions and specialties over 7 years.

Many clinical questions can be answered through e-consults without need for in-person follow-up over a 12-month period.

Electronic consultations, or e-consults, allow health care providers—primarily primary care providers (PCPs)—to receive guidance virtually from specialists about patient-specific issues. A number of health systems

 in the United States have implemented e-consults using 2 primary designs with the potential of obviating the need for in-person specialty visits: (1) a “single pathway,” where the consultant performs a preliminary evaluation and decides whether an e-consult or in-person visit is most appropriate, or (2) the “optional pathway,” where the referring provider chooses to obtain advice through an e-consult rather than referring for a traditional in-person specialty evaluation (

 The enhanced communication facilitated by e-consults between specialists and PCPs provides education, reduces patient burdens, and improves comprehensiveness, coordination, and access to care.

 Given the promise of e-consults as a potentially efficient and cost-saving care modality, there is growing interest in their use both in the US health care system and worldwide.

Since 2011, the Veterans Health Administration (VHA) has implemented and expanded the use of e-consults with the aim of improving specialty care access and coordination. VHA, the largest integrated health care system in the United States, serves a geographically and demographically diverse population of more than 6 million patients in primary care annually. The e-consult system is decentralized, allowing for each regional service network or health care system to design and implement its own e-consult process to serve its population of patients and providers. An evaluation of the VHA’s e-consult implementation from May 2011 through December 2013 demonstrated an increase in the total number of e-consults without a plateau, enhanced care access for rural veterans, and potential reductions in travel and associated costs.

 During this period, e-consults were completed at an adjusted rate of 1.9 per 100 total consults across 13 specialty clinics. The highest rates were among endocrinology (5.0/100), hematology (3.0/100), and gastroenterology (3.0/100).

 Since this evaluation, e-consults have expanded to all VHA outpatient clinics, electronically connecting them to specialty care at medical center hubs. Details of e-consult use since this expansion have not yet been reported.

The primary objective of this study was to describe e-consult use and growth across regions and specialties in the VHA from 2012 through 2018. Secondary aims were to examine potentially avoidable in-person specialty care visits within 12 months of an e-consult and to describe potential driving distance and cost savings associated with e-consults.

This is an observational, retrospective analysis of e-consult use from January 1, 2012, through December 31, 2018. The initial run-in period of e-consult implementation of May through December 2011 was excluded. We examined data across the VHA’s 18 regional Veteran Integrated Service Networks (VISNs), including all 170 Veterans Affairs Medical Centers (VAMCs) and their 1269 associated clinical sites. The study population included all VHA patients receiving primary care during the observation period. Data were obtained from the VA Service Support Center, VHA’s Corporate Data Warehouse, and the VA Planning Systems Support Group.

We identified specialty type, date, and location of all e-consults by clinical “stop code,” which included 41 unique clinical care specialties, categorized within 3 subgroups as medical (n = 16), surgical (n = 12), and other (n = 13) specialties. e-Consults for mental health services were not included. e-Consult originating site was designated as the VAMC or an associated clinical site.

We identified date and type of all in-person specialty consults occurring across the same 41 medical, surgical, and other specialties as e-consults. For patients with an e-consult, we identified any in-person visit to that specialty clinic occurring within 3, 6, and 12 months after the e-consult was placed. Individual patients could have more than 1 e-consult in any given year. For each specialty and specialty category, we determined the annual proportion of e-consults performed of the total electronic and in-person consults.

Demographics including age, gender, and rurality are described. Patient rurality was defined by the Rural-Urban Commuting Area classification, which creates 30 mutually exclusive categories representing population density and proximity to urban centers using the patient’s residential zip code.

 These categories were condensed to urban, rural, and highly rural.

Geocoded addresses of patients’ home addresses and VHA locations of specialty care were used to calculate the shortest drive distance using Bing maps.

All analyses were performed using SAS version 7.1 (SAS Institute) and R software. The study was reviewed by the University of Iowa Institutional Review Board and the Iowa City VA Healthcare System Research and Development Committee.

Between January 1, 2012, and December 31, 2018, we identified a total of 3,230,566 e-consults, representing 3.9% of 83,271,378 total electronic and in-person consults. We narrowed our analysis to 41 specialties of interest, which encompassed 96.5% of all e-consults. Over the 7-year period, e-consults across these specialties increased by 309% from 151,984 e-consults in 2012 to 621,2013 e-consults in 2018, representing 2.2% and 7.0% of total electronic and in-person consults, respectively (

e-Consult use increased within 17 of the 18 regional networks (ie, VISNs). Across VISNs, e-consults as a proportion of all consults increased from a range of 0.2% to 7.0% (mean, 2.4%) in 2012 to 3.8% to 14.1% (mean, 5.0%) in 2018 (

). The majority of this increase was experienced by 2014. However, VISN 9, serving primarily Kentucky, Tennessee, and northern Mississippi, experienced rapid and sustained e-consult growth, starting in 2013 (5.2%) and continuing through 2018 (14.1%).

Over the period of analysis, the number of facilities ordering e-consults increased from 553 (57%) to 774 (61%). By 2018, 179 (98.8%) of the 181 VAMCs, 55.5% of associated outpatient clinics, and 8.4% of other facilities ordered e-consults. With the expansion of 296 additional outpatient clinics over this time period, the percentage of clinics using e-consults increased from 49.8% to 55.5%. The distribution of e-consults across clinical sites remained similar, with 83% from VAMCs and 17% from associated outpatient clinics. However, the proportion of e-consults relative to in-person consults increased across all sites, with the greatest increase among associated outpatient clinics (2.4% in 2012 to 8.9% in 2018) followed by VAMCs (2.2% to 6.6%).

The number of unique patients receiving at least 1 e-consult increased from 131,403 in 2012 to 494,253 in 2018, representing 6.7% of the enrolled veteran population and 84 e-consults per 1000 enrolled patients. By 2018, veterans receiving an e-consult had a mean (SD) age of 63.1 (14.7) years, 90.1% identified as male, and 31.7% and 1.1% resided in rural and highly rural areas, respectively (Table).

e-Consults for medical specialties accounted for 59.5% (n = 1,855,521) of all e-consults and 10.0% of all medical consults. Over the 7 years of observation, the number of medical e-consults increased 4.6-fold, with the relative percentage of e-consults of all medical consults increasing from 3.9% to 12.8%. e-Consult use grew across all medical subspecialties (

). The top 5 specialties performing e-consults were cardiology (n = 332,216), gastroenterology (n = 243,020), sleep medicine (n = 209,073), endocrinology (n = 204,509), and pulmonary (n = 181,430). For 11 of 16 medical specialties, more than 10% of all consults within the given specialty were completed as e-consults: hematology (38.7%), infectious diseases (23.5%), endocrinology (21.1%), nephrology (19.5%), pulmonary (17.4%), oncology (16.4%), gastroenterology (16.2%), geriatrics (12.7%), sleep medicine (11.8%), neurology (10.8%), and immunology (10.4%). By 2018, more than 1 in 5 consults for hematology, infectious diseases, endocrinology, pulmonology, nephrology, and gastroenterology were e-consults.

Surgical specialties accounted for 20.0% (n = 624,879) of all e-consults and 2.0% of all surgical consults. From 2012 through 2018, the number of surgical e-consults grew 8.0-fold, with the relative percentage of e-consults increasing from 0.8% to 5.8% of all surgical consults. The top 5 surgical specialties performing e-consults were preoperative evaluation (n = 144,838), urology (n = 99,584), orthopedics (n = 97,803), neurosurgery (n = 65,790), and vascular (n = 58,872). Four surgical specialties completed greater than 10% of all consults as e-consults: preoperative evaluation (22.0%), neurosurgery (20.4%), vascular surgery (14.6%), and cardiothoracic surgery (14.6%) (Figure 3 [B]).

Other specialties accounted for 20.4% (n = 637,698) of all e-consults. By 2018, the number of other consults grew by 2.0-fold. The most e-consults performed among other specialties were pharmacy (n = 369,413), physiatry (n = 108,654), and pain (n = 81,611). Two specialties (genomic medicine, 19.3%; pain medicine, 11%) completed greater than 10% of all consults as e-consults (Figure 3 [C]).

Overall, 56.2% of e-consults were requested by physicians, 23.4% by nonphysician providers (NPPs; ie, physician assistants or nurse practitioners), 8.9% by residents or fellows, 3.3% by pharmacists and other providers, 0.8% by nurses, and 7.5% by unknown provider type. With regard to providers completing e-consults, 53.4% were completed by physicians, 21.8% by NPPs, 11.4% by pharmacists, 4.7% by other, 4.7% by residents or fellows, and 3.2% by nurses; 0.8% were completed by missing provider type. For medical specialties, 65.1% of e-consults were completed by physicians and 21.9% by NPPs. Over time, the proportion of e-consults completed by physicians decreased from 67.0% to 63.6% and by NPPs increased from 18.9% to 23.9%. For surgical specialties, 50.9% of e-consults were completed by physicians and 37.2% by NPPs. For other specialties, pharmacists completed 56.0% of the e-consults, followed by physicians (21.7%).

For all 41 specialties over the 7 years of observation, 13.1% of e-consults resulted in an in-person visit within 3 months, 16.7% within 6 months, and 21.5% within 12 months. Across categories, in-person follow-up after e-consults was lowest among surgical specialties (3 months, 6.0%; 6 months, 8.1%; 12 months, 10.7%) followed by medical (7.0%, 9.4%, and 12.5%, respectively) and other specialties (23.1%, 26.9, and 32.3%) (

We tracked only 1-year in-person follow-up visits for e-consults until the end of 2018. From 2012 to 2017, in-person specialty visits within 12 months after an e-consult increased from 16.4% to 23.0% for all specialties, increased from 15.6% to 23.9% for medical specialties, decreased from 18.8% to 9.8% for surgical specialties, and increased from 12.1% to 18.6% for other specialties.

The mean (SD) drive distance to the nearest tertiary care facility among patients who did not have an in-person visit with the same specialty within 12 months of an e-consult was 84.3 (89.9) miles. Over the study time period, 170,658,048 total miles of travel for in-person specialty care visits were potentially “saved.” At the VHA mileage reimbursement rate of 41.5 cents per mile, this equates to an estimated savings of $70,823,090 in travel expenses. Considering that approximately 65% of veterans are eligible for travel reimbursement, e-consults potentially saved VHA more than $46 million in travel costs.

From 2012 to 2018, VHA clinicians adopted e-consults for specialty care, resulting in more than 3 million e-consults and potentially avoiding hundreds of thousands of in-person clinic visits, saving more than 170 million driving miles (and associated costs), and enhancing convenience for patients and family members. By 2018, e-consults represented 12.8% of medical, 5.8% of surgical, and 3.0% of all “other” specialty consults. For the majority of medical specialties, at least 10% of total consults are now e-consults. With the substantial growth of e-consults, a few findings and considerations are worth highlighting.

First, the reach and adoption of e-consults have been maintained. Our findings identified that e-consults reached a substantial portion of VHA clinics and maintained growth beyond the first years of implementation.

 Over the observation period, all but 1 region experienced growth in e-consults. The maintained use of e-consults highlights the value to both primary care providers and specialists to enhance communication and timeliness to specialty care.

 The integrated electronic health record (EHR) and e-consult platform within the VHA overcomes barriers experienced in other e-consult programs,

 easing access to specialist input. Across various settings, other programs experience similar rapid adoption and maintained use.

 Our evidence further supports e-consults as a sustainable model to potentially improve the comprehensiveness of and access to care.

Second, the differential growth of e-consult use across regions and specialties signifies further opportunity for expansion. Use of e-consults across most regions plateaued at around 6% of total in-person consults. However, we identified 1 region (VISN 9) where clinical champions promoted the implementation of the single pathway, requiring a preemptive e-consult to determine appropriateness and type of consultation. This region used e-consults for more than 25% of medical consults and more than 15% of surgical consults, with specific specialties exceeding 50%, including preoperative evaluations, hematology, neurosurgery, and sleep medicine. Although most locations have implemented the optional pathway, the single pathway has resulted in highest e-consult use at care facilities within

 the VHA. By changing the e-consult pathway or supporting physician champions and engaged management teams, regions and clinics may be able to further expand the reach and adoption of e-consults.

e-Consult expansion may be particularly valuable and achievable among specialties that depend more on interpretation of laboratory testing and imaging and less on physical examination. Across the 41 specialties examined, e-consults for hematology, preoperative evaluation, neurosurgery, endocrinology, and infectious diseases were the most frequently used relative to in-person consults among all regions and the regions of highest e-consult use. e-Consults for these specialties

 serve as examples of best practice dissemination.

 In certain regions, clinical champions have responded to unique local needs for improved access and care efficiency by expanding specific e-consult programs. Examples include programs for geriatric services,

 In addition, e-consults can be offered as a templated electronic interface to reduce unneeded in-person consults and enhance access among the most commonly ordered specialty consults (eg, cardiology).

Third, follow-up in-person visits to the same specialty up to 1 year after an e-consult were low and in line with those found in prior studies.

 However, we did observe an increase in follow-up at 3, 6, or 12 months for medical specialties over time. Low overall follow-up suggests that e-consults can substitute for in-person visits; however, the availability and ease of e-consults may result in a “lower bar” for consultation, resulting in e-consults for conditions that otherwise would be handled independently by the requesting provider. This study did not examine content or appropriateness of e-consults; however, in a recently published evaluation within a single health care system, more than 70% of e-consults were considered appropriate.

 The recent increase in in-person follow-up may be due to increased use of the single pathway. Facilities may mandate or providers may choose to place a preemptive e-consult to clarify proper evaluation (eg, imaging, laboratory studies) and facilitate an efficient in-person specialty consultation.

By averting in-person consultations, e-consults reduce unneeded travel to in-person specialty care. This travel for patients and caregivers incurs significant costs in time, money, and inconvenience. Across the 41 specialties, we estimated that, on average, an averted visit would save 84 miles of travel, a total exceeding 30 million miles annually, with significant potential savings to patient and the VHA. This is a benefit viewed by both providers and patients.

The benefit of e-consults also translates to providers and care delivery, offering opportunities to improve education, communication, and care coordination across disciplines. In this study, only half of e-consults were requested and completed by staff physicians. VHA has a history of incorporating a teaching mission

 as well as advancing the integration of NPPs into care.

 e-Consults provide requesting providers, NPPs, and trainees with readily accessible specialty advice and education, which may serve to channel learning health system goals of advancing timeliness and quality of care. This may be particularly important for NPPs, trainees, and physicians with less than 10 years of practice, who are more likely to report improved management skills as a result of e-consults.

Despite the benefits, e-consults add tasks and responsibilities to the PCP. In our study, e-consults increased by 300% and in-person consults remained relatively stagnant. Many PCPs recognize the benefits of e-consults; however, they also have reported an added burden of increased administrative work and broadened clinical responsibility.

 If low-value tasks are not shifted and dedicated time for these non–in-person activities is not allocated,

 PCPs will likely face increasing task burden and higher potential for burnout.

Adoption and sustained use of e-consults across the national VHA implicates the value of this provision for integrating care and improving access.

 The shared EHR, limited third-party billing, lack of specialists in all clinic locations, and credentialing across care locations and states facilitate implementation within the VHA. Other regions and health systems have identified modalities to overcome barriers.

 With expansion of data sharing across health systems through direct EHR collaboration and health information exchanges,

 e-consults have the potential to enhance access in a timely and cost-effective manner.

 Despite broad uptake, evidence on the effectiveness of e-consults to improve quality of care and condition-specific outcomes has been modest.

 Further work looking at e-consult use in individual specialties and conditions is needed, including specific reasons for e-consults, potential risks, and opportunities for specialty-specific growth to improve access, quality, and costs. This research could inform guidelines for implementation and expansion of appropriate and high-quality e-consult use.

There were several limitations to this work. First, each VHA facility offering specialty care established e-consult services at its discretion and with different consultation processes. Second, there was no control group and the data were retrospective; thus, we cannot determine to what extent e-consults substituted for in-person visits, affected specialist workload, were used appropriately, and resulted in patient satisfaction with this form of care. Third, the VHA has separate teledermatology, teleradiology, and teleretinal programs that allow for store-and-forward image transmission, but these e-consults were not included in this analysis. Fourth, both drive time and dollars saved are idealized maximum estimates based on the assumption that an e-consult served as a complete substitution for an in-person consult. Lastly, the VHA’s experience with e-consults may not be generalizable to the broader diverse American health care landscape because of differences in insurance coverage, reimbursement for e-consult care, and other economic considerations that differentiate VHA from non-VHA care.

e-Consults have emerged as an important tool for improving access to specialty care, as is evidenced by their broad and sustained implementation and use across the VHA since 2012. Variation in e-consult growth across clinical sites and specialties suggests that there exists further potential for expansion and use of e-consults within and outside the VHA.

The authors would like to thank Emily Ashmore for helping to create Figure 1 of the manuscript.

 Division of Endocrinology, Rocky Mountain Regional Veterans Affairs Medical Center (DRS), Aurora, CO; Division of Endocrinology, Metabolism and Diabetes, University of Colorado, Anschutz Medical Campus (DRS), Aurora, CO; Center for Access and Delivery Research and Evaluation, Iowa City VA Healthcare System (PJK, BH, MO), Iowa City, IA; Department of Internal Medicine, University of Iowa Carver College of Medicine (PJK, MO), Iowa City, IA; Primary Care Analytics Team, VA Puget Sound Health Care System (CW), Seattle, WA; Department of Medicine, James J. Peters VA Medical Center (MRA), Bronx, NY; Department of Medicine, Icahn School of Medicine at Mount Sinai (MRA), New York, NY.

 This material is based upon work supported (or supported in part) by the Department of Veterans Affairs (VA), Veterans Health Administration, VA Office of Rural Health and the Office of Research and Development, Health Services Research and Development Service through the Center for Access and Delivery Research and Evaluation (CIN 13-412).

 The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (DRS, PJK, MO, MRA); acquisition of data (PJK, BH, CW); analysis and interpretation of data (DRS, PJK, BH, CW, MRA); drafting of the manuscript (DRS, CW, MO, MRA); critical revision of the manuscript for important intellectual content (DRS, PJK, MO, MRA); statistical analysis (BH, CW); supervision (PJK); and software (CW).

 David R. Saxon, MD, MS, Division of Endocrinology, Metabolism and Diabetes, University of Colorado, Anschutz Medical Campus, 12801 E 17th Ave, Mail Stop 8106, Aurora, CO 80045. Email: david.saxon@cuanschutz.edu.

1. Barnett ML, Yee HF Jr, Mehrotra A, Giboney P. Los Angeles safety-net program eConsult system was rapidly adopted and decreased wait times to see specialists. 

. 2017;36(3):492-499. doi:10.1377/hlthaff.2016.1283

2. Kim-Hwang JE, Chen AH, Bell DS, Guzman D, Yee HF Jr, Kushel MB. Evaluating electronic referrals for specialty care at a public hospital. 

. 2010;25(10):1123-1128. doi:10.1007/s11606-010-1402-1

3. North F, Uthke LD, Tulledge-Scheitel SM. Integration of e-consultations into the outpatient care process at a tertiary medical centre. 

. 2014;20(4):221-229. doi:10.1177/1357633X14533886

4. Gleason N, Ackerman S, Shipman SA. eConsult—transforming primary care or exacerbating clinician burnout? 

. 2018;178(6):790-791. doi:10.1001/jamainternmed.2018.0762

5. Lee MS, Ray KN, Mehrotra A, Giboney P, Yee HF Jr, Barnett ML. Primary care practitioners’ perceptions of electronic consult systems: a qualitative analysis. 

. 2018;178(6):782-789. doi:10.1001/jamainternmed.2018.0738

6. Davis A, Gilchrist V, Grumbach K, James P, Kallenberg R, Shipman SA. Advancing the primary/specialty care interface through eConsults and enhanced referrals. 

. 2015;13(4):387-388. doi:10.1370/afm.1829

7. Liddy C, Drosinis P, Keely E. Electronic consultation systems: worldwide prevalence and their impact on patient care—a systematic review. 

. 2016;33(3):274-285. doi:10.1093/fampra/cmw024

8. Kirsh S, Carey E, Aron DC, et al. Impact of a national specialty e-consultation implementation project on access. 

Rurality Definitions and Methods Work Group Report

10. Rodriguez KL, Burkitt KH, Bayliss NK, et al. Veteran, primary care provider, and specialist satisfaction with electronic consultation. 

. 2015;3(1):e5. doi:10.2196/medinform.3725

11. Kwok J, Olayiwola JN, Knox M, Murphy EJ, Tuot DS. Electronic consultation system demonstrates educational benefit for primary care providers. 

. 2018;24(7):465-472. doi:10.1177/1357633X17711822

12. Vimalananda VG, Gupte G, Seraj SM, et al. Electronic consultations (e-consults) to improve access to specialty care: a systematic review and narrative synthesis. 

. 2015;21(6):323-330. doi:10.1177/1357633X15582108

13. Knox M, Murphy EJ, Leslie T, Wick R, Tuot DS. e-Consult implementation success: lessons from 5 county-based delivery systems. 

. 2020;26(1):e21-e27. doi:10.37765/ajmc.2020.42149

14. Liddy C, Moroz I, Afkham A, Keely E. Sustainability of a primary care–driven eConsult service. 

. 2018;16(2):120-126. doi:10.1370/afm.2177

15. Tuot DS, Liddy C, Vimalananda VG, et al. Evaluating diverse electronic consultation programs with a common framework. 

. 2018;18(1):814. doi:10.1186/s12913-018-3626-4

16. Afable MK, Gupte G, Simon SR, et al. Innovative use of electronic consultations in preoperative anesthesiology evaluation at VA medical centers in New England. 

. 2018;37(2):275-282. doi:10.1377/hlthaff.2017.1087

17. Cecchini M, Rose MG, Wong EY, Neparidze N. The implementation of electronic hematology consults at a VA hospital. 

. 2016;127(12):1610-1611. doi:10.1182/blood-2015-09-672113

18. Patel PS, Jiang B, Marcelli M, Mediwala SN, Vasudevan MM. Electronic consultation: an effective alternative to in-person clinical care for patients with diabetes mellitus. 

. 2019;13(1):152-153. doi:10.1177/1932296818807472

19. Strymish J, Gupte G, Afable MK, et al. Electronic consultations (e-consults): advancing infectious disease care in a large Veterans Affairs healthcare system. 

. 2017;64(8):1123-1125. doi:10.1093/cid/cix058

20. Elnahal SM, Clancy CM, Shulkin DJ. A framework for disseminating clinical best practices in the VA health system. 

. 2017;317(3):255-256. doi:10.1001/jama.2016.18764

21. Pimentel CB, Gately M, Barczi SR, et al. GRECC Connect: geriatrics telehealth to empower health care providers and improve management of older veterans in rural communities. 

22. Baig MM, Antonescu-Turcu A, Ratarasarn K. Impact of sleep telemedicine protocol in management of sleep apnea: a 5-year VA experience. 

. 2016;22(5):458-462. doi:10.1089/tmj.2015.0047

23. Liddy C, Smyth C, Poulin PA, Joschko J, Rebelo M, Keely E. Improving access to chronic pain services through eConsultation: a cross-sectional study of the Champlain BASE eConsult service. 

. 2016;17(6):1049-1057. doi:10.1093/pm/pnw038

24. Oseran AS, Wasfy JH. Early experiences with cardiology electronic consults: a systematic review. 

. 2019;215:139-146. doi:10.1016/j.ahj.2019.06.013

25. Kim EJ, Orlander JD, Afable M, et al. Cardiology electronic consultation (e-consult) use by primary care providers at VA medical centres in New England. 

. 2019;25(6):370-377. doi:10.1177/1357633X18774468

26. Venkatesh RD, Campbell EJ, Thiim M, et al. e-Consults in gastroenterology: an opportunity for innovative care. 

. 2019;25(8):499-505. doi:10.1177/1357633X18781189

27. Ahmed S, Kelly YP, Behera TR, et al. Utility, appropriateness, and content of electronic consultations across medical subspecialties: a cohort study. 

. 2020;172(10):641-647. doi:10.7326/M19-3852

28. Olayiwola JN, Anderson D, Jepeal N, et al. Electronic consultations to improve the primary care–specialty care interface for cardiology in the medically underserved: a cluster-randomized controlled trial. 

. 2016;14(2):133-140. doi:10.1370/afm.1869

Articles

This study evaluates the growth in electronic consultation use over the first 7 years after its implementation across the entire Veterans Health Administration system.

Growth of Electronic Consultations in the Veterans Health Administration

 To examine the effect of a patient activation intervention with financial incentives to promote switching to a thiazide in patients with controlled hypertension using calcium channel blockers (CCBs).

The Veterans Affairs Project to Implement Diuretics, a randomized clinical trial, was conducted at 13 Veterans Affairs primary care clinics.

 Patients (n = 236) with hypertension previously controlled using CCBs were randomized to a control group (n = 90) or 1 of 3 intervention groups designed to activate patients to talk with their primary care providers about switching to thiazides: Group A (n = 53) received an activation letter, group B (n = 42) received a letter plus a financial incentive to discuss switching from a CCB to a thiazide, and group C (n = 51) received a letter, a financial incentive, and a telephone call encouraging patients to speak with their primary care providers. The primary outcome was thiazide prescribing at the index visit.

At the index visit, the rate of switching to a thiazide was 1.1% in the control group and 9.4% (group A), 26.2% (group B), and 31.4% (group C) for the intervention groups (

 < .0001). In adjusted analysis, patients randomized to group C were significantly more likely to switch from a CCB to thiazide at the index visit (odds ratio, 4.14; 95% CI, 1.45-11.84; 

This low-cost, low-intensity patient activation intervention resulted in increased rates of switching to a thiazide in those whose hypertension was controlled using another medication, suggesting that such interventions may be used to overcome medication optimization challenges, including clinical inertia.

Medication optimization remains difficult to achieve in practice. Patient-centered interventions in this context, which emphasize increased communication and shared decision-making between the patient and provider, have shown promise.

In this patient activation intervention for hypertension, intervention patients with their hypertension already controlled using another medication were more likely to switch to a thiazide diuretic compared with usual care control patients. Specifically, adjusted odds of switching were greatest for patients randomized to receive a custom letter, modest financial incentive, and brief nurse-led reminder call.

These findings suggest that low-cost, low-intensity patient activation interventions may be useful in addressing medication optimization challenges and clinical inertia.

Deprescribing refers to the process of removing, reducing, or switching medications to reduce risks and improve patient outcomes.

 A conceptually overlapping process is that of medication optimization, which seeks to maximize medication benefit through appropriate prescribing; it shares similar strategies with deprescribing and may involve removing, reducing, or switching medications.

 Both concepts are important given the pervasiveness of polypharmacy; half of older adults take 5 or more medications, at least 1 of which may be potentially inappropriate.

 Achieving optimal medication usage thus may improve numerous outcomes in this population, including quality, cost, adherence, and reduction of adverse drug events,

 but it remains difficult to achieve in practice.

 Several factors that may inhibit optimal medication usage have been identified, including clinical inertia, provider assumptions about patient treatment preferences, patient beliefs about the appropriateness of the medication in question, and patient fears related to cessation (eg, withdrawal effects).

Such factors underscore the need to employ a patient-centered process to achieve medication optimization, which emphasizes increased communication and shared decision-making between the patient and provider in clinical encounters.

 Studies in the deprescribing literature employing patient-centered approaches have reported favorable results,

 suggests that interventions designed to actively involve the patient are more effective than those that do not (eg, those focused solely on prescribers).

Direct-to-consumer (DTC) marketing has been used by the pharmaceutical industry for some time to motivate discussions between patients and physicians to affect prescribing.

 Fundamentally, DTC marketing seeks to increase patient activation, a term that refers to the patient-centered process of imparting knowledge, skills, and confidence, which may facilitate patients becoming more active and informed participants in health care delivery.

 Patient activation interventions have shown benefit in terms of better health outcomes and care experiences,

 and interventions that facilitate patient initiation of discussion about medication optimization may be especially beneficial as they provide opportunities for shared communication and decision-making.

The present study examines data from the Veterans Affairs Project to Implement Diuretics, a randomized clinical trial designed to test the efficacy of a patient activation intervention with financial incentives to promote thiazide prescribing (ClinicalTrials.gov identifier: NCT 00265538). When the trial was initiated in 2006,

 the prevailing hypertension (HTN) guidelines from the Joint National Committee (JNC) 7 for the Management of Hypertension in Adults specifically stated that “thiazide-type diuretics should be used in drug treatment for most patients with uncomplicated [HTN].”

 In addition to being recommended for first-line therapy, thiazides were considerably cheaper than many other antihypertensives, with a favorable efficacy and safety profile. The 2014 guidelines (formerly JNC 8) recommend selection among 4 classes, including thiazides.

 The arm of the trial examined in the current study aimed to promote switching to thiazides in a group of patients with controlled blood pressure (BP) using calcium channel blockers (CCBs) due to the recommendations at the time and considerable cost savings, aligning with the concept of medication optimization in which medications that may not pose significant harm—but are suboptimal—are targeted for switching or reduction. Although much of the deprescribing literature focuses on medication discontinuation, switching is also an important principle of deprescribing/medication optimization and deserves further attention.

 Thus, our primary interest in this study was to examine whether a low-cost, low-intensity intervention would affect medication switching for those already at their BP goal taking a nonpreferred agent. We hypothesized that patients assigned to the most “active” arm of the intervention would exhibit the greatest degree of medication switching.

Participants were sampled from all patients who received primary care at 13 Veterans Affairs (VA) outpatient clinics affiliated with 2 large VA medical centers (N = 61,019) (

). To be eligible for participation, patients must have had a prior diagnosis of HTN, not be taking a thiazide diuretic, have 2 or more VA clinic visits in the past 12 months, not be in a long-term care facility, and have no known thiazide or sulfa allergy or prior adverse drug event. From this, 2 subpopulations were derived: (1) patients not at an appropriate BP goal at the 2 most recent clinic visits, referred to as 

, and (2) patients at an appropriate BP goal, but taking a CCB, referred to as 

. Attainment of goal BP at baseline was determined using the mean of all clinic-recorded BPs from the 2 clinic visits immediately preceding the initial study clinic appointment (hereafter referred to as the index visit). Complete inclusion and exclusion criteria, as well as primary trial results focusing on those not at an appropriate BP goal, have been reported elsewhere.

Participants were recruited following a staged protocol that randomized 168 primary care providers (71 affiliated with the Iowa City VA Medical Center and 97 affiliated with the Minneapolis VA Medical Center) to intervention and control groups and then selected eligible patients within these groups. Upon primary care provider randomization, patients with a primary care appointment within 30 days were identified and screened using the electronic health record. Eligible patients (n = 2853) were mailed a letter describing the study objectives, enrollment process, and data collection. One week later, patients were contacted by telephone to obtain informed consent and administered the Pfeiffer Short Portable Mental Status Questionnaire (SPMSQ) to exclude patients with cognitive impairment (defined as > 2 errors on the SPMSQ).

 Of the 2853 patients who were contacted by mail, 2019 were attempted by phone; 1151 declined to participate, including 86 with no telephone. Enrollment ended when the target number of participants were enrolled. Consenting patients (n = 236) were randomized to an intervention (n = 146) or 1 of 2 usual care control groups (n = 90), one group of approximately 50 who were seen by providers with no intervention patients and another group of approximately 50 seen by providers also with intervention patients. We conducted several analyses that demonstrated that these patient groups were no different from one another and thus collapsed them for this study. Sample sizes were selected to provide sufficient (≥ 80%) power to detect an 18% absolute difference in thiazide use between the control group and a single treatment group. All study procedures were approved by The University of Iowa’s institutional review board (#200506705).

Patients randomized to the intervention were assigned to 1 of 3 intervention groups. Group A (n = 53) received a customized letter 1 to 3 weeks before their scheduled index visit. Information was tailored to the patient’s treatment regimen, BP, and cardiovascular risk factors. The letter requested that the patient initiate a discussion with their primary care provider about switching to a thiazide and provided specific options for doing so. It included a postcard for patients to have providers sign and return, documenting whether thiazides were discussed during the visit.

Group B (n = 42) received the same patient education letter in addition to a financial incentive ($20 by check) to initiate a discussion with their primary care provider (see 

] for an example letter representing group B). The incentive was awarded upon receipt of the signed postcard. For patients with a prescription co-pay (55/93; 59.1% across groups B and C), the letter also indicated that thiazide co-pays for 6 months ($48) would be reimbursed upon receipt of the postcard and confirmation that a thiazide was prescribed.

Group C (n = 51) received the same letter and financial incentives as group B in addition to a brief telephone call within 1 week before their clinic visit. Calls were conducted by a trained health educator who discussed the mailed information, the patient’s role in initiating a discussion with the primary care provider about BP medications, and potential strategies for initiating these conversations. All patients received up to 5 call attempts prior to their scheduled index visit; the mode number of calls was 2. Calls lasted, on average, between 5 and 10 minutes.

The primary outcome was thiazide prescription at the index visit; switching to a thiazide from a CCB was determined by reviewing each patient’s progress note and medication list after the index visit, which included discontinued orders to establish that CCBs were stopped. Intervention and control groups were compared at baseline in terms of sociodemographic characteristics and other clinical variables of importance (eg, baseline BP, smoking, mental status, diabetes) using standard univariate tests including 

 tests. Variables showing a (conservative) significant difference between groups at 

 < .10 were included as covariates in adjusted analyses.

Because of the clustered nature of these data, statistical analyses must account for correlation in order for the observed results to be valid given that traditional regression models assume independence of observations. Therefore, analyses utilized hierarchical logistic regression using the generalized linear mixed model procedure in SPSS version 24.0 (IBM). Models incorporated random physician effects to account for within-physician clustering; clinic sites were included as fixed effects. Analyses tested whether there was a significant difference between the log odds of thiazide prescription for the control group vs the 3 intervention groups at the index visit.

During recruitment, 2853 eligible patients were identified. Of these, 236 who were at an appropriate BP goal and taking a CCB for HTN treatment were consented and enrolled. These patients were randomized to either intervention (n = 146: group A = 53; group B = 42; group C = 51) or control (n = 90) groups. Mean age for the intervention groups was 66.7 years and mean systolic/diastolic BP at baseline was 121.5/74.5 mm Hg (see 

, which also includes a breakdown by intervention group assignment). Mean age for the control group was 66.9 years and mean BP at baseline was 122.2/73.2 mm Hg. Only race varied significantly between groups and was added as a covariate in the adjusted model. No differential attrition by group was observed, which includes appointment no-shows and elective study dropouts. Only 2 of 51 (3.9%) group C participants were not reached by phone, making them functionally more like group B participants; these participants were analyzed according to the group they were randomized to.

Rate of thiazide prescribing, representing a switch from a CCB to thiazide, was the primary outcome of interest. At the index visit, the unadjusted thiazide prescription rates were 1.1% (1/90) for the control group and 9.4% (5/53) for group A, 26.2% (11/42) for group B, and 31.4% (16/51) for group C (

 < .0001). Adjusted odds ratios (ORs) for the 3 intervention groups relative to control are presented in 

. At the index visit, the odds of switching to a thiazide were not significantly greater compared with the control group for group A (OR, 2.37; 95% CI, 0.82-6.82; 

 = .11) and only marginally so for group B (OR, 2.77; 95% CI, 0.96-8.01; 

 = .06). Group C (OR, 4.14; 95% CI, 1.45-11.84; 

 < .01) had significantly greater odds of switching to a thiazide than the control group, the magnitude of which has been suggested to be equivalent to a medium effect size.

This study suggests that a relatively low-cost, low-intensity intervention may affect medication optimization behavior in the context of HTN, even for patients currently controlled on another medication. At the index visit, approximately one-third of patients in group C switched from a CCB to thiazide, which was significantly different from the control group in the adjusted analysis. Perhaps most noteworthy is the fact that these patients had HTN that was controlled at the time of intervention and, presumably, most had no pressing reason to switch other than those primarily tied to financial benefit (ie, nominal monetary incentive, prescription co-pay assistance).

Patients were first activated to engage their primary care providers through a customized letter, which only modestly improved switching compared with control (group A; see unadjusted rates above and adjusted odds ratios in Table 2). Adding financial incentives resulted in a marginally significant increase in the odds of switching for group B and, further, a brief reminder phone call resulted in significantly increased odds for group C. The use of financial incentives to drive behavior change has shown success across multiple health-related contexts, including medication adherence, smoking cessation, and weight loss.

 It has, however, notably failed in others.30 In this study, a letter plus modest financial incentives were not enough to achieve statistically significant switching (

 = .06), but adding a brief reminder call previsit was enough to reach significant change (

 < .01). Although a formal budget impact analysis was not part of the study, the financial incentive was considerably less than the ongoing cost of a CCB and the letters were automatically generated from data in the electronic health record and thus low cost. The phone calls took approximately 5 to 10 minutes but improved the potency of the intervention and served to remind patients about their appointments as well as to bring their medications with them, a frequent part of usual care.

Given that group C participants received a tailored letter encouraging discussion with the provider, modest financial incentives, and a brief phone call, it is likely that successful interventions will require a multifaceted approach to facilitate medication optimization, including active engagement of the patient.

 This is further underscored by calls to implement interventions designed to address patient, provider, and/or systems-level factors.

 In the present study, the tailored letter not only provided information specific to the patient’s condition but also served as a means of facilitating discussion about medication options. Although, on its own, it was not enough to drive significant change, it does represent a key first step in shared decision-making and presumably resulted in some dialogue about the benefits vs harms of switching. Because a majority of (although not all) patients prefer to be actively involved in the decision-making process,

 which includes being offered choices and the freedom to express opinions, it would seem that simple tools such as a patient-directed letter may hold value as one instrument toward broader medication optimization when combined with additional elements like financial incentives and other means to further engage the patient. Even for those preferring to leave final decisions to the provider, such letters could serve as a prompt for discussion, which itself may combat clinical inertia and misperceptions about patient preferences on the part of the provider.

 Indeed, a companion qualitative study of participating intervention providers found that, in addition to very favorable overall perceptions of the intervention and belief in its suitability for wider use, providers specifically noted that the intervention helped align patient objectives with their own.

Not all multimodal interventions are equally feasible in practice, however. Many interventions have demonstrated success in improving deprescribing outcomes, but there is often a disconnect between research and practice in that many interventions may be unrealistic for widespread adoption in busy clinic settings.

 Compared with interventions that require the involvement of multiple providers, considerable effort on the part of a lone provider, and/or complex referral or care coordination processes, the present approach may be more logistically practical and demand fewer resources. Future research should be designed with an eye toward what is reasonably attainable in the clinical setting to address this important issue.

This study has several limitations. First, it is reasonable to argue that because some patients did not have a compelling reason to switch other than financial incentives and then-contemporary practice guidelines, there may have been no need to switch at all. However, this study suggests that financial incentives, when coupled with other modest interventional components, can facilitate discussion of the issue, which is an important first step in the deprescribing/medication optimization process. Previous qualitative work from this trial found that patient participants liked being involved because it allowed them the opportunity to initiate discussions with the provider on a topic of importance.

 Such empowerment may carry over to other aspects of the clinical encounter. Second, we do not have information regarding the exact reasons why patients and providers decided to switch from a CCB to a thiazide; it is assumed that the financial incentives and practice guidelines largely influenced this change, but it is possible that at least some patients were candidates for switching irrespective of our intervention, perhaps due to low tolerance of the medication at the time of visit. This concern is alleviated by the presence of a control group, however. Third, we do not have specific race/ethnicity information for a number of patient participants in this study, which is primarily due to issues with reporting and classification in the VA medical record at the time of this trial.

 Fourth, we did not monitor downstream efforts to address the appropriateness of the switch (eg, follow-up laboratory tests, BP monitoring) and thus cannot speak to patients who did not respond well to the new medication. Finally, HTN guidelines evolved over time, making some recommendations obsolete. However, in addition to the clear connections to deprescribing/medication optimization, this study also may have implications specifically for therapeutic substitution (ie, when a lower-cost alternative within the same class is available), particularly when developing methods to inform and engage patients about less expensive options that are otherwise equivalent for the same indication.

These results suggest that a low-intensity intervention activating patients toward increased communication (in this case, discussion of current HTN medications in clinic) may be relevant to deprescribing and broader medication optimization efforts, helping to overcome clinical inertia. Future work should evaluate the practical barriers of implementing even simple interventions into complex and busy clinical practices and develop ways to overcome those barriers. 


 Center for Access Delivery Research & Evaluation (CADRE), VA Iowa City Healthcare System (MBH, MWVW, PJK), Iowa City, IA; Department of Behavioral Sciences & Social Medicine, College of Medicine, Florida State University (MBH), Tallahassee, FL; Division of Geriatrics, University of California San Francisco (MAS), San Francisco, CA; San Francisco VA Health Care System, San Francisco (MAS), CA; Division of General Internal Medicine, The University of Iowa Carver College of Medicine (BC, MWVW, PJK), Iowa City, IA; Department of Pharmacy Practice and Science, The University of Iowa College of Pharmacy (BC), Iowa City, IA; Department of Psychological & Brain Sciences, The University of Iowa (MWVW), Iowa City, IA.

 The research reported here was supported by the Department of Veterans Affairs, Veterans Health Administration, Health Services Research and Development Service Merit Review Grant (IMV 04-066-1). The views expressed in this article are those of the authors and do not necessarily represent the view of the Department of Veterans Affairs.

 Dr Steinman has received grants from the National Institutes of Health (NIH) related to appropriate medication use and has received honoraria for talks at universities and participation in programs sponsored by other nonprofit entities around issues related to appropriate medication use. Dr Carter is an investigator on 3 NIH grants and receives honoraria as a member of 2 Data and Safety Monitoring Boards for NIH. 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.

 Concept and design (MBH, BC, MWVW, PJK); acquisition of data (BC, PJK); analysis and interpretation of data (MBH, MAS, MWVW, PJK); drafting of the manuscript (MBH, PJK); critical revision of the manuscript for important intellectual content (MBH, MAS, BC, MWVW, PJK); statistical analysis (MBH, PJK); provision of patients or study materials (PJK); obtaining funding (MBH, PJK); administrative, technical, or logistic support (PJK); supervision (BC, PJK); and content expertise (MAS).

 M. Bryant Howren, PhD, MPH, Florida State University College of Medicine, 1115 W Call St, Tallahassee, FL 32306. Email: matthew.howren@med.fsu.edu.

1. Scott IA, Hilmer SN, Reeve E, et al. Reducing inappropriate polypharmacy: the process of deprescribing. 

In patients with hypertension, a patient activation intervention increased rates of switching to a thiazide, suggesting that such interventions may address medication optimization challenges.

Peter J. Kaboli, MD, MS

Articles

Growth of Electronic Consultations in the Veterans Health Administration

Effect of a Patient Activation Intervention on Hypertension Medication Optimization: Results From a Randomized Clinical Trial

Assessing the Accuracy of Computerized Medication Histories

Associations With Reduced Length of Stay and Costs on an Academic Hospitalist Service