|
Utilization of Secure Messaging to Primary Care DepartmentsJose Yakushi, MD1; Mose Wintner, PhD2; Naomi Yau2; Lina Borgo, MPH2; Edwin Solorzano, MD1 Perm J 2020;24:19.177 [Full Citation] https://doi.org/10.7812/TPP/19.177E-pub: 04/29/2020ABSTRACTIntroduction: Secure messaging is a platform for email communication between patients and their physicians. Although patient-generated emails are associated with increased use of clinical services, greater member retention, and improved quality of care, secure messaging has a marked impact on primary care physicians’ workload. Objective: To understand how the email topic and volume vary by demographics and clinical factors among members of a managed care organization. Methods: We analyzed all secure messages sent to primary care departments by adult members of Kaiser Permanente Southern California (KPSC) in 2017. Results: Members with a higher volume of office visits and telephone appointment visits generated a higher volume of emails to primary care physician. Members with a Centers of Medicare and Medicaid Services Hierarchical Condition Category diagnosis history sent 3 times as many emails as those without such a diagnosis history. Women accounted for nearly two-thirds of emails despite making up only half of the KPSC member population. Less than one-fourth (21.4%) of members sent 2.3 million total emails to their physician. Medical advice was the most common reason for sending secure messages (24.7%) in a sample studied (n = 2397). Discussion: These findings confirm the need for additional research to more accurately quantify the additional burden from secure message utilization on primary care physicians. Knowing the factors associated with secure messaging usage and message content could assist in building more efficient staffing models and creating more efficient routing that matches the message content with a physician’s scope of practice. INTRODUCTIONSecure messaging is a protected electronic communication service between patients and physicians. Known for its ability to foster physician-patient relationships, secure email has become a ubiquitous communication tool in health care settings.1 Studies have shown that patient satisfaction has generally increased since adopting secure messaging in family medicine, with satisfaction scores higher on questions related to “ease of communicating with the physician” and “understanding of diagnosis/treatment plan”.2 Some organizations initially adopted secure messaging as a way to address low-acuity medical issues, assuming this might reduce unnecessary telephone or in-person office visits.3,4 Others have targeted young and healthy patients in an early attempt to increase use of secure messaging, assuming the millennial population is more likely to use this virtual service than the older population.2,5 The rapid adoption of secure messaging has furthered the focus on and improved the quality of patient-centered care6,7, a pillar of Kaiser Permanente’s (KP) care delivery system.8 KP’s secure messaging system was implemented in 2005, first beginning with the KP Hawaii Region. Secure messaging is facilitated through KP’s member portal (https://kp.org), which is integrated with KP’s electronic health record, HealthConnect. HealthConnect represents KP’s implementation of the Epic suite of software (Epic Systems Corp, Verona, WI). To access the member portal, patients are required to register for a kp.org account and be authenticated users.8 In the first quarter of 2017, more than 67% KP Southern California (KPSC) members were registered on the member portal, an 8.3% growth from the prior year. In the fourth quarter of that same year, 23.1% of all portal-registered members in KPSC sent 1 or more secure email messages to their physicians; an average of 1272 emails were sent per 1000 registered members in a year. When the 3.6% membership increase from the previous year is factored in, 2017 saw an 11% increase in secure messages from 2016, with a total of 5,807,833 secure messages sent.8 Secure messaging is now a common form of physician-patient interaction through which care is provided and common concerns are addressed by the care team. This study analyzed all adult-generated secure messages in the KPSC Region. The purpose of this study was to better understand the usage of secure messaging between KPSC members and primary care departments. This was accomplished by 1) comparing the use of secure messaging against the utilization of office visits and telephone appointment visits (TAVs) in primary care, 2) examining the demographic characteristics of patients who use KP’s secure messaging platform, and 3) investigating the content of secure messages that KP members sent to their primary care physicians (PCPs). METHODSStudy Design and Population After obtaining KP institutional review board approval, we conducted a retrospective study analyzing patient demographic data (eg, age, sex), clinical factors (eg, Health Plan type, time between a sent message and nearest completed office visit), and primary care utilization data (eg, encounter volume). All data examined in this study pertain to the utilization of primary care services in Southern California, by adult KPSC members from January 2017 to December 2017. Primary care included both family medicine and internal medicine specialties. Data from the Pediatrics and Obstetrics/Gynecology Departments were not included in this study. At KP, members seeking care for primary care services have several options. For the purposes of this study, we focused on 3 main primary care services: Office visits, TAVs, and secure messaging. To quantify the utilization of these modalities, we counted instances of the following primary care services for each individual who held a KPSC membership in 2017 (4.7 million KP members): 1. Office visits in 2017 (n = 4,898,596) to primary care departments by adult KPSC members 2. Telephone appointment visits (n = 1,185,590) with primary care departments by adult KPSC members in 2017 3. Secure messages (n = 2,316,309) from adult KPSC members to primary care departments in 2017 that were not replies to a previous message (ie, initiations of message threads by members, not by physicians). Furthermore, we performed random sampling of secure messages (n = 2397 of total 2,316,309) for email topic analysis. In this study, patients with a chronic condition were defined as having a Centers for Medicare and Medicaid Services (CMS) Hierarchical Condition Category (HCC) diagnosis history (used in the CMS risk-adjustment model10) listed in their medical record. The CMS HCC was originally designed to estimate future health care costs for patients and is now often used to communicate patient complexity. Of the 2.3 million email messages sent in 2017, we analyzed the content and corresponding metadata of 2397 messages. Our researchers tagged each email message with up to 3 primary message topic tags, using a detailed description of each message topic. The detailed description and table were modified from the Veterans Affairs study by Shimada et al.11 Table 1 depicts the various email topics. Statistical Analysis We performed a negative binomial regression to assess the number of messages sent by a patient. Independent variables included primary care utilization counts; geographic area of the patient’s medical center; and patient’s age, sex, number of chronic illnesses, race/ethnicity, and Health Plan product type. The reference patient we compared all variables to was white, male, age 52 years (median age in our dataset), had no chronic illnesses, and had a KP Health Plan in his home Region. This regression accounts for each variable independent of one another. For example, although being a Medicare member and age may be related, this regression represents them as 2 separate variables, independent of one another. Transformed versions of office visit counts and TAV counts were used in the regression, namely log(variable + 1).
RESULTSUtilization of Primary Care Services Main Drivers for Secure Message Utilization As shown in Figure 1, the number of completed primary care office visits by a member was most strongly associated with a member sending a greater number of secure messages to the primary care department. The number of TAVs was the second strongest factor. Other factors associated with a member sending more secure messages included being a Medicare member, being a woman, and having a prior CMS HCC diagnosis. Because this regression looks at each variable independently, this study’s findings show that having an office visit was more indicative of sending a secure message than being a Medicare member or having a chronic disease. Conversely, being a KP-Medicaid member and self-identifying as black and/or Hispanic were factors associated with fewer secure messages sent. Average counts of office visit utilizations were similar between racial groups, but propensities for secure messaging varied widely (Table 2). Though there was only a small difference between average numbers of office visits between black, Hispanic, and white members, members who self-identified as black and/or Hispanic sent fewer messages than other races with similar average number of office visits. As seen in Table 3, members who had at least 1 office visit had, on average, more than 2 office visit in 2017. Members who had at least 1 TAV had, on average, 1.63 TAVs. Members who initiated a secure message thread initiated nearly 3 messages on average. Table 3 also shows the average counts for members who did not have any office visits, TAVs, or secure messages sent. Correlation between Secure Messaging and Office Visits/TAVs Figure 2 illustrates how age and sex were associated with utilization of different primary care services. The spike in office visit utilization correlates with a spike in utilization of TAVs and secure messages. The observed spike is likely related to the free Medicare annual wellness office visit that members are encouraged by KP to schedule once they become a Medicare member. Additionally, KP PCPs often use TAVs to follow-up with Medicare members after a Medicare wellness visit and often encourage members to use secure messaging to follow-up after their Medicare wellness visit. In Figure 3, the largest volume of messages is highlighted, detailing that 7.69% of messages sent by adult members to primary care departments were sent within 1 day of a PCP office visit. Another 14.85% of messages were sent within 3 days of a PCP office visit, and 37.5% of messages were sent within 7 days of a PCP office visit. Utilization by Chronic Condition Diagnosis Of KPSC members, 42.5% had a chronic condition, yet this population accounted for two-thirds of primary care service utilization. Those KP members with a chronic condition used TAVs and secure messaging more than in-person office visits. In Table 4, patients with a chronic condition diagnosis history sent more than 3 times as many secure messages to their PCP and had 2 times as many office visits compared with patients without any chronic condition diagnosis. Utilization by Sex Approximately half of KPSC members (51.7%) in 2017 were women (Table 4), but female members accounted for almost two-thirds of primary care utilization. On average, 5 of 8 messages sent to a PCP were sent by women. In a comparison of the 3 different primary care services, the biggest difference in utilization between sexes was seen in TAVs, with female members accounting for 65% of TAV encounters. Utilization of All Service Modalities Less than one-fourth of the member population was responsible for the sum of 2.3 million secure messages to PCPs in 2017 (Table 5). According to Figure 4, younger men used primary care services the least. For example, less than 5% of 18-year-old men used secure messaging. Overall, women across all ages used primary care office visits and TAVs more than men did. However, in secure messaging, after age 70 years, men consistently sent more messages to their primary care department than did women. Message Content Nearly 1 in 4 secure messages to the primary care office fell into the medical advice category. The top categories included medication issues and questions, informing the physician (ie, messages not requiring an action or reply), medication refills, and scheduling (Figure 5). Nearly 25% of all messages examined were about medications.
DISCUSSIONThis study demonstrates the interdependency of patient demographics and clinical factors on secure message utilization. Studies have shown that secure messages and TAVs were implemented to create more efficient primary care services.3,4,12 Our study findings reveal that members with a higher utilization of office visits and TAVs sent more secure messages than those who did not use those clinical services. Through our variable-independent regression analysis, we found that having an office visit and having a TAV are more indicative of sending secure messages than being a Medicare member or having a chronic condition, by a factor of 6 and 13, respectively, although being a Medicare member and having a chronic condition are still significant in attributing to larger secure message volumes. At KPSC, our PCPs often prompt patients at the end of their office visits to email them if they have any further questions. Whereas our study results suggest an association between the volume of secure messaging and office visits, other studies have shown that the relationship between secure messaging, office visits, and TAVs are complex. Some studies have shown that messaging can replace the number of office visits or TAVs; other studies have suggested that secure messaging either increases or does not alter the utilization of office visits or telephone calls.4,12-20 Further studies must be done to confirm these correlations. Studies have found that patients with certain demographic backgrounds used care services overall more than others did.21,22 Our study findings confirm previous research by demonstrating that certain patient variables, such as being a woman, being white, having a chronic condition, and being older, contribute to higher secure messaging utilization. Moreover, results of this study reveal the magnitude at which certain clinical and patient demographic factors affect secure message utilization. A member having a CMS HCC diagnosis history may send 3 times as many emails to primary care than a member without an HCC diagnosis history. Women also account for almost two-thirds of secure messaging to primary care departments excluding Obstetrics/Gynecology. Because of this exclusion, our results likely underestimate the message gap between sexes. In a comparison of the 3 different primary care services, the biggest difference in utilization between sexes was seen in virtual services, with female members using 65% of TAV encounters and 62.4% of secure messaging encounters. Many studies have researched the heavy adoption of secure messaging in outpatient medicine and obstetrics/gynecology.4,12-22 Data shows that the adoption of the secure messaging platform is not yet widespread. Nearly 25% of the KPSC member population used the secure messaging platform in 2017. This group sent approximately 2.3 million messages in 2017. This finding can also be interpreted as more than 75% of KPSC members have yet to adopt secure messaging. Past KP trends data suggest, however, that the use of secure messaging increases each year.8,9 The primary care team needs to be well positioned to account for the increased general adoption of secure messaging. Our study adds to existing research by looking at message topics to primary care departments. Consistent with other studies, medical advice is ranked as one of the top message topics that members are inquiring about.22,24,25 The top 6 message topics following medical advice are very similar to the message topic study conducted at Veterans Affairs medical centers,11 with a slight difference in topic ranks. The variation could be caused by the small sample size and difference in the service population. Knowing what types of messages patients are sending will help when building, prioritizing, or updating routing protocols. Practical Applications Today, drivers of secure message utilization are not commonly considered in certain aspects of the secure message management system in primary care settings. Knowing the factors associated with secure messaging utilization and the message content could assist in the following: 1) building more efficient staffing models; 2) creating more efficient routing that matches the secure message content with maximum scope of practice allowed for nurses, pharmacists, and advanced practice physicians; and 3) using message volume as a weighted factor that influences a PCP’s panel ceiling, or cap. Staffing Models Our analysis suggests that certain populations tend to send a larger number of secure messages and therefore produce additional virtual work for physicians. The factors in these populations should be considered by health systems when developing staffing and resource allocation protocols to address this imbalance. For example, if a physician is projected to have a heavier secure message workload based on his/her panel, nurses experienced with handling secure messaging can be assigned to assist. Routing Secure Messages Understanding what patients are messaging about can aid in designing a more efficient secure messaging model to route and prioritize messages, stratifying them for the appropriate recipient. For example, if messages about medications have a high volume, there may be value in routing messages about medication to pharmacists directly, bypassing the primary care team. Optimized Panel Calculation To account for workload owing to messaging volumes, our research findings suggest that some patient variables should be considered when health systems calculate PCP panel sizes. PCPs whose panels contain white, Medicare-eligible women with a chronic condition likely have a heavier workload because of the increased volume of secure messages they receive from these patient populations. Limitations This study has several limitations. First, our study included secure messages through https://kp.org only to the Family Medicine and Internal Medicine Departments, excluding the Pediatrics and Obstetrics/Gynecology Departments. Our analysis did not include any messages to specialty departments. Second, we examined data from only a single KP Region, KPSC. Our study does not include messaging, office visit, and TAV data trends from other KP Regions. Studies have found that the relationship between secure messaging, office visits, and TAVs is complex. Some study findings have shown that messaging can replace the number of office visits or TAVs; other studies have suggested that secure messaging either increases or does not alter the utilization of office visits or telephone calls.4,12-20 Therefore, our study, being an observational study, cannot claim whether messaging drives other forms of care or vice versa. To make such a claim about causality, we would need a randomized controlled trial. The results in this article speak only to associations. Last, it is important to note that the total number of secure messages sampled in this study reflects only the 21.4% of KPSC members who used secure messaging and does not reflect the population as a whole. As more and more members adopt the service, further studies are needed to determine if the observations found in this study stay consistent. CONCLUSIONThis study examined several factors that were often associated with utilization of secure messaging, ranging from having an office visit to having a chronic condition. Factors most strongly associated with increased utilization of secure messaging were the numbers of primary care office visits and TAVs. Demographic factors associated with secure message utilization included sex, age, HCC diagnosis history, and Health Plan type. Being white, female, and older and having a Medicare Health Plan type and a previous HCC diagnosis were associated with higher utilization of secure messages. The findings of this study reveal the magnitude of these factors on secure message utilization and challenge many existing assumptions around secure messaging. Disclosure StatementThe author(s) have no conflicts of interest to disclose. AcknowledgmentsThe authors thank the physicians and operational directors in the Kaiser Permanente Southern California Region, the Health Innovation team under the Southern California Permanente Medical Group and Peter Jung from Kaiser Permanente Business Systems and Reporting for their assistance. Kathleen Louden, ELS, of Louden Health Communications performed a primary copy edit. Author Affiliations1 Department of Family Medicine, Downey Medical Center, CA Corresponding AuthorJose Yakushi, MD (jose.a.yakushi@kp.org) How to Cite this ArticleYakushi J, Wintner M, Yau N, Borgo L, Solorzano E. Utilization of secure messaging to primary care departments. Perm J 2020;24:19.177. DOI: https://doi.org/10.7812/TPP/19.177 References1. Garrido T, Meng D, Wang JJ, Palen TE, Kanter MH. Secure e-mailing between physicians and patients: Transformational change in ambulatory care. J Ambul Care Manage 2014 Jul-Sep;37(3):211-8. DOI: https://doi.org/10.1097/JAC.0000000000000043 PMID:24887522
|
ETOC
Click here to join the eTOC list or text ETOC to 22828. You will receive an email notice with the Table of Contents of The Permanente Journal.
CIRCULATION
2 million page views of TPJ articles in PubMed from a broad international readership.
Indexing
Indexed in MEDLINE, PubMed Central, EMBASE, EBSCO Academic Search Complete, and CrossRef.