Community-Based Telemonitoring for Hypertension Management: Practical Challenges and Potential Solutions

Research Setting

Four senior centers in Southwestern Ohio were selected to serve as field centers for this project. Two of these centers served as intervention centers (telehealth kiosks installed) and the other two served as control centers. Criteria for selection of these centers included location in a rural, underserved area, provision of meals at least once per week, availability of a sufficient number of hypertensive seniors to meet recruitment targets, and support of the study by senior center staff and willingness to allow study staff to visit on a regular basis to engage in study activities. The intervention facilities met additional criteria, including adequate space to accommodate the telehealth kiosk, availability of high speed Internet connectivity service (at digital subscriber line or cable modem speeds), willingness to permit installation of kiosks onsite and to have a staff member trained in operating the equipment and assisting seniors if necessary. In addition, we looked for demonstration of strong, stable leadership; commitment to the objectives of the research; and availability staff whose roles permitted involvement in the project.

Participants and Enrollment

The client base of the four senior centers described above is ∼1,200 persons, from which a target enrollment of 144 seniors with diagnosed hypertension (36 persons at each center) was established. Eligible participants were 55 years or older, lived in a rural underserved area, visited the senior center one or more times a week, had physician-diagnosed hypertension that was stable with oral therapy, and agreed to interact with study nurses. Staff consented participants, and obtained permission from enrolled participants to contact primary care physicians for the purpose of obtaining recommended intervention protocols and defining referral criteria for changes in BP treatment.

Procedures

Telehealth kiosks were installed at two of the four senior centers. Participants in the intervention arm were asked to monitor their BP at least once per week for 10 months, but they could engage in self-monitoring any time they used the senior center.

BP data from intervention participants were streamed to a central server managed by a telehealth vendor. University-based nurse-researchers monitored BP information and made referrals according to criteria defined by participants' primary care physicians at the time of enrollment. Follow-up and other health utilization measures were collected from both intervention and control facilities.

Vendor personnel installed a telehealth kiosk at each of the two intervention centers and tested the system before screening and enrollment. Ongoing technical support was provided throughout the follow-up period. Participants accessed the kiosks using ID cards that were kept at the two intervention centers. These cards were freely available to study participants during normal business hours at the centers. Participants were trained on use of the kiosk (which included a card reader, BP cuff, and touch screen computer connected to the Internet), and senior center staff was available to assist participants with kiosk use throughout the follow-up period. Participants with high BP readings were instructed via an on-screen pop up to retest their BP 20 min later and to answer a questionnaire about health behaviors to help the recipients of the reports understand the causes of the elevated BP measurement. All information was made available to the nurse-researchers who were engaged in remote monitoring and who then implemented the referral protocol recommended by the participants' physicians. The rest of this report deals with practical considerations associated with use of telemonitoring in the senior center setting.

Internet Connectivity

Internet connectivity at the intervention sites was not optimal despite preimplementation screening of the centers for adequacy of service. One site in particular had trouble obtaining service and maintaining continuity from its local ISP. There was a period of ∼10 days during which Internet connectivity was unstable or unavailable. Difficulties at this center led not only to problems with the functioning of the telehealth kiosk, but they also created frustration for both participants and senior center staff. Most notably, service interruptions resulted in the inability to transfer data from the kiosk to the vendor's server. Although service interruptions were generally brief and no emergency situations arose during these times, nurses were, nonetheless, unable to monitor BP in real-time during these episodes of interrupted service. In this sense, Internet service interruptions not only undermined compliance efforts among participants, but they also influenced the ability of nurses to respond to health data in real-time.

ISP Support

Technical support from the local ISP was at best inconsistent, and at worst, unavailable. The ISP had difficulty opening the port required to allow the vendor's technicians to access the telehealth computer remotely over the Internet using Virtual Network Computing. This difficulty limited the vendor's ability to remotely troubleshoot and provide needed technical support to one site—an essential component of the project. Repeated attempts were made to establish remote access through the static IP address assigned to the kiosk, but these efforts were unsuccessful. Senior center staff lacked the knowledge and resources to resolve technical issues independently. These circumstances required one of the study nurses to travel to the senior center and address technical issues with the vendor over the phone. Although effective, this solution was suboptimal: The nurse lacked expertise with the equipment, had other study obligations, and was based over an hour from the senior center. Moreover, the site's technical issues were easily diagnosed by the vendor and could have been resolved in minutes with remote access had the ISP been responsive to the needs of the senior center and the research staff.

Hardware

Before enrollment, it was discovered that one site's computer was unable to communicate with the wireless BP measurement device via Bluetooth dongle. Once this problem was identified, a simple software update immediately resolved the problem. During the study, one of the computers froze repeatedly, which led the research team to suspect the cause to be a virus or malware. The vendor thoroughly tested the problematic computer to determine the cause of the freezing. Because the problem persisted after reformatting and reinstalling all software and hardware drivers, the vendor determined the cause was most likely a defect in hardware. At that point, the hardware was replaced by the vendor. Replacement of the hardware inadvertently, but fortuitously, addressed complaints participants had made about sensitivity of the touch screens. A number of users experienced frustration with the touch screens because of a perception of insufficient screen sensitivity. When the hardware was replaced, it contained a newer, more sensitive screen. This enhanced satisfaction among many users and eliminated complaints related to screen sensitivity. In addition, study nurses placed a mouse at the kiosk as an alternative to the touch screen.

The BP cuff presented separate hardware issues. The BP device inflates the cuff to a predetermined pressure, but when inflation is not high enough to accurately measure systolic BP, the cuff deflates and re-inflates to higher level to obtain a better reading. This is a routine device function, but resulted in confusion among users. Some participants removed their arms partially or completely from the cuff between readings, causing the subsequent measurement to fail and produce an error message. In this case, participants must recognize the error and retake the reading. In most cases, an accurate reading was obtained but confusion over this function should be noted. Inaccurate or missing readings undermine the ability of the nurses to provide targeted remote monitoring, and ultimately to assist patients in reaching their disease management goals.

Software

Staff identified several software-related issues that resulted in participant confusion and in some instances disrupted data collection. For example, although the software displayed a “Send” button on the kiosk screen after the device captured the BP measurement, this function was designed for off-line applications that required the user to press “Send” to deliver data to the vendor's server. In this setting, typically, BP readings uploaded automatically and were transferred to the server without user intervention. An exception to this process occurred after an Internet outage, in which case BP data had to be entered manually into the computer and users needed to press “Send” to transfer the data once Internet connectivity was restored. Using the “Send” button at any other time resulted in transfer and storage of duplicate readings in the BP reports, which were easy to identify and correct in the BP database. However, this issue resulted in periodic confusion and frustration among users regarding correct use of the “Send” button.

When a high BP reading was detected, a pop-up message appeared on the screen to alert participants and prompt them to complete an on-screen health questionnaire. After replacement of the hardware at one site (described above), this pop up no longer appeared. Vendor's technical staff enabled this function remotely.

Another software issue involved the kiosks' going into a default sleep mode, which temporarily prevented patients from logging in. To address this issue, the sleep mode was disabled. Although this problem was easy to resolve, it is important to recognize as a cause for confusion and possible loss of data if the problem source had not been discovered quickly. For example, there was a 2-week period during which study staff could not initialize new ID cards at the senior centers. One of the study nurses resolved this issue with telephone-based technical support from the telehealth vendor. However, had this problem occurred during the enrollment period, it would have prevented new participants from enrolling and being monitored. It should be emphasized that no ID cards needed to be reissued during this period, so there was no disruption in monitoring for participants who had started the program. These situations further underscore the importance of basic computer literacy and adequate training among on-site staff and reliable remote technical support from telehealth vendors.

Technology Literacy in Congregate Settings

As expected, participant anxiety and technology literacy, especially in conjunction with Internet connectivity, ISP as well as software and hardware problems, were major barriers to compliance and collection of accurate and consistent BP readings. However, despite a perception on the part of study investigators that participants would have limited experience with automatic BP cuffs, 99% of participants had used automatic cuffs (and many had used them hundreds of times) before the study. Nevertheless, some participants did experience anxiety as they learned to use the kiosks, and at times when connectivity, software, or hardware problems occurred, this anxiety increased and resulted in elevated BP readings. Interestingly, we observed that elevated BP readings appeared to be associated with certain activities that are common at senior centers such as card playing, which was very competitive among those seniors, as we learned from nurse-researchers. Some participants who monitored their BP immediately after a card game had high initial measurements that later decreased at the second reading. It should be noted that participants were instructed to do their BP measurements after a period of rest, but they did not consistently follow these instructions.

Cuff Placement and Compliance

Research participants did not always use the appropriate BP cuff size or follow instructions about proper cuff placement consistently, and they often required supplemental instructions by senior center staff. Incorrect cuff placement observed by nurse researchers included placing the cuff on the elbow, and using the cuff over bulky clothes, which prevented a reading from being obtained. If participants made a correction, such as removing the bulky item, the cuff sometimes re-inflated to the point of discomfort. Although proper cuff placement was demonstrated to study participants during training on the use of the kiosk and picture-illustrated instructions on proper cuff placement were posted on both kiosks, study participants needed additional reminders to follow these instructions and periodic supervision to ensure that proper placement was consistent.

At both sites, ∼73% of participants used the BP kiosk once a week throughout the study. Site 1 averaged 50% compliance with second BP readings; Site 2 averaged 74% compliance. Both sites averaged 13% compliance with the health questionnaire. The kiosk was designed to prompt participants to take a second BP reading if the first reading was high, and to complete the questionnaire. During the study, staff observed that compliance with completing the questionnaire and re-measuring BP was low. In response, study staff changed the pop-up message to instruct users to complete the questionnaire after the first high-BP reading and then retake the BP. However, between 25% and 50% of participants did not complete second readings and, as suggested above, 87% did not complete the questionnaires. Failure to complete a second reading could result from a number of issues, including forgetfulness, frustration over cuff inflation, poor touch screen sensitivity, inconsistent pop-up reminders, or any of the other hardware and software issues discussed that had the potential to influence participants' experiences interacting with the technology. Similarly, failure to complete the questionnaire could have resulted from inconsistent pop-up reminders, difficulty navigating the user-interface to access the questionnaire, or broader dissatisfaction on the part of users.

Physician Engagement

When a hypertensive participant wished to enroll in the study, his or her primary care physician provided BP thresholds to guide the study nurses' interventions. Ideally, physicians would have defined specific BPs at which participants should modify diet or exercise, consider a medication change, consult their physician via telephone, schedule an appointment with their physician, or seek immediate medical attention. Further, these thresholds would be based on the physicians' knowledge of their patients' histories and be aimed at minimizing the need for in-person visits when possible. In practice, however, most physicians asked to be alerted if their patients' readings were over 140/90 and provided no additional instructions to study nurses. Because of these generic referral instructions, study nurses were underutilized; if BP readings were high, they called patients and recommended a physician appointment. Eventually, some participants tired of this advice, particularly the subset of participants with consistently high readings. In some cases, these participants refused to answer nurses' calls. Others reported that their physicians had modified their recommendations and set a higher BP threshold but had failed to notify study nurses.

Internet Connectivity

Inconsistent Internet connectivity was a problem for both data collection and patient compliance. Internet service outages at one site prevented the transfer of BP readings to the vendor's server. Some participants wrote down their measurements and entered them manually when connectivity was restored. However, some did not write down readings and others did not bother to take their BP measurement during Internet outages, resulting in lost data. Future implementation projects should take measures to ensure reliable connectivity to prevent data loss and noncompliance. Sites could also consider mounting a logbook at the kiosk in which seniors enter their BP readings and ID numbers during an Internet outage. A logbook would serve both as a physical record of BP, preventing data loss, and as a visual reminder for seniors to complete and record readings even in the case of Internet outage.

ISP Support

Like inconsistent connectivity, insufficient local ISP support was a hindrance to optimal use of the kiosks. Failure of the local ISP to establish a remote connection prevented the vendor from providing technical support to the intervention sites. Reliable Internet service and remote access are critical to the delivery of technical support, which is in turn critical to successful implementation of telehealth technology and remote patient monitoring applications. These issues are particularly important in the underserved and geographical isolated areas that were targeted in this study, in which Internet providers are often small, transient, and lacking in adequate customer support resources. If reliable ISP support is not available or a remote connection cannot be established, both vendors and study staff must ensure that an onsite staff person is trained in basic software and hardware functions and is willing to troubleshoot problems with telephone support from the vendor.

Hardware

Once diagnosed, hardware issues were resolved either through replacement of the equipment or software updates.

As noted above, re-inflation of BP cuffs reflects normal functioning of the device. Efforts should be directed toward educating and reminding the users about proper cuff size, cuff placements, and the need to remain still while the measurements are being taken. These factors can cause the BP device to fail to capture a correct reading and subsequently cause the cuff to re-inflate.

Software

Duplicate readings that resulted from users pressing the “Send” button unnecessarily were easy for study staff to identify and ignore on a small scale, but may overwhelm personnel in a large-scale implementation project. To address these issues, the software platform could be modified either by adding an instructional pop-up or displaying the send button only when user action is required.

Vendor's technicians were able to make the necessary modifications to reinstate the pop-up prompt and disable the kiosks' sleep mode. Critically, the nurses and site staff were able to identify these problems and report them to the vendor. The period in which new ID cards could not be issued was longer and had the potential to prevent new users from being monitored or interrupt the monitoring of active users who lost their cards. It is important to emphasize the need for computer literacy among study nurses and/or nutrition center staff and, in particular, remote technical support so that problems can be identified and resolved without disruption to health monitoring, data collection, or user experience.

Technology Literacy in Congregate Settings

Elevated BPs induced by environmental factors were observed and must be considered. Some elevated readings were related to anxiety over cuff use. Although many patients reported familiarity with the cuffs, some, nonetheless, experienced anxiety. Participants may benefit from additional instruction in kiosk and cuff use to reduce their anxiety, as well as regular reminders from nutrition center staff in future studies.

Participants were reminded to conduct BP readings after a period of rest. However, the concept of what constitutes a period of rest needed to be revaluated in this setting since most individuals interpret this to mean absence of physical activity. Contrary to this definition, we observed that certain sedentary activities such as card games, which were popular among participants, could influence BP readings. Participants also reported anxiety over traveling in inclement weather that was reflected in their BP readings if they failed to observe 20 min of rest after arriving at the center. These observations point to the importance of understanding the unique characteristics of this and other congregate service settings in which telemonitoring can be implemented. Staff who are familiar with participants' activities may need to issue additional instructions regarding BP readings, for example, advising them to sit quietly with a book or TV show before measuring BP.

Cuff Placement and Compliance

To address cuff-related confusion nutrition center staff and nurse researchers wrote the cuff size appropriate for each participant on their ID card and posted multiple reminders on the walls. In poststudy interviews, staff recommended that a procedural cheat sheet be posted at kiosks or dispensed to participants with their ID cards each day to eliminate the cuff-size and placement confusion as well as the questionnaire compliance. The cheat sheet would remind participants about proper cuff use (position and orientation on the arm, remaining still), returning after 20 min for a second reading if the first is high, and answering the health questionnaire after a high reading. The cheat sheet would replace multiple, additional instructions hung near the kiosks and which participants appeared to ignore. Another possible solution is to place the kiosk near a staff member station so assistance can be provided when necessary. On the other hand, this may over-burden the staff and undermine the notion of independent self-monitoring. Instead, periodic group retraining sessions may be an efficient way to ensure participants use the equipment correctly. Questionnaire compliance could be further enhanced by modifying the software to display the questionnaire automatically after the first high BP reading, rather than require users to navigate back to the home screen to find the questionnaire. Study staff suggested this modification based on their observations of compliance, but the vendor could not implement this change during the course of the study. However, it is a possible solution for future implementation studies to consider.

Physician Engagement

Physician discomfort with telehealth and telemonitoring is not surprising given the novelty of the field. In this project, physicians asked to be contacted if their patients' BPs were over 140/90, thereby making minimal use of nurses and the technology. Ideally, physicians would use their familiarity with the patient's history to define personalized BP thresholds at which specific care plans should be enacted, such as retaking the measurement; being diligent about an exercise program and retesting in a week; modifying diet and then retesting; and making a physician appointment. Our current healthcare system favors the latter approach, which often prevents seniors from interacting effectively with healthcare professionals because of challenges such as long distances, difficulties with transportation, and the disconnect between scheduled office visits and clinically meaningful changes in disease status. For example, it is often the case that a medication titration is needed at times that do not coincide with scheduled visits, and that scheduled visits can be associated with unnecessary costs for routine monitoring that could be done more efficiently in the community. These barriers and inefficiencies are exacerbated among rural elders with limited economic resources, and include seniors who utilize subsidized programs and services of community-based senior centers. Advances in telehealth are intended to increase care options for chronic care management but will require extensive educational campaigns to secure both patient and physician support.