Type 2 Diabetes: Intervention Design and Implementation

Summary

Home telehealth has been used extensively in the management of type 2 diabetes.  Most interventions are designed to enable remote monitoring through the transmission of blood glucose values from patient to provider.  Videoconferencing may be used for appointments with physicians or dietitians.  In some cases, there is an online educational component as well.

Details

Basic Model

The standard type 2 diabetes intervention offered monitoring-based management of patients’ blood glucose. Blood pressure was also often monitored. In some cases there was monitoring of additional biometric values, such as weight (Stone et al., 2010; Yoo et al, 2009; Cho et al., 2006; Bond et al., 2007, 2010), exercise and/or diet (Tjam et al., 2006; Robertson et al., 2007; Bond et al., 2007, 2010), social activity (Robertson et al., 2007; Cho et al., 2006), medication use and administration (the biodang.com studies; Bond et al., 2007, 2010), and diabetes knowledge (Dang et al., 2007). In the VA-CCHT interventions only diabetes-related symptoms and health status were monitored; self-measured biometric values were not transmitted to providers.

Transmission of biometric values from the patient to provider often occurred via a telehealth device, such as a glucose or blood pressure meter with transmission capabilities, Health Hero Buddy, or another kind of “dialogue box”. But in most cases, patients logged onto a website to upload or manually enter the self-measured values. Some of the monitoring interventions that used websites also featured a patient diabetes education component in addition to the transmission capacity (Shea et al., 2006, 2009; Trief et al., 2006, 2009; Chang et al., 2007; Izquierdo et al., 2007; Bond et al., 2007, 2010; McMahon et al., 2005; Levine et al., 2009).

In seven interventions transmission occurred via a mobile phone (Cho et al., 2009; Kim et al., 2005; Rodríguez-Idígoras et al., 2009; Istepanian et al., 2009a, b; Trudel et al., 2007; Yoo et al, 2009), or landline phone (Luzio et al., 2007). Patients used meters to self-measure values; the meters transmitted the value via Bluetooth to the mobile phone or the patients manually input the values to send to the provider.

Provider contact with patient was mostly carried out via email or some kind of website-based messaging, or with a landline telephone. Patients received feedback from the provider via SMS in most of the interventions carried out as part of the biodang.com study and Yoo et al. (2009). In the Rodríguez-Idígoras et al. (2009) intervention providers placed a call to the patients’ mobile phone to give monitoring feedback.

In the Bond et al. (2007, 2010) interventions, patient-provider contact occurred on an instant message/chat platform, as well as during online discussion groups. Moreover, Bond et al. (2007, 2010) are unique because the intervention website also allowed for peer-to-peer instant messaging and email among patients. Peer support was also facilitated in Robertson et al. (2007) and Tjam et al. (2006).

In Tildesley et al. (2010), the provider did not provide feedback using an ehealth technology; while providers received the values via the intervention, feedback was reserved for face-to-face consultations.

Consultation-based monitoring typifies interventions that use videoconferencing. This technology was only used for monitoring with a small section of patients in the VA-CCHT interventions and as another mode of patient-provider communication in the IDEATel interventions in addition to biometric meters with transmission capabilities and a website-based messaging service.

Human Resource Requirements

Nurses delivered the telehealth intervention in a large majority of the diabetes studies. Nurses were responsible for monitoring and, in some instances, contacting other providers when necessary. Physicians performed patient monitoring duties in five interventions (Bujnowska-Fedak et al., 2006; Cho et al., 2009; Trudel et al., 2007; Tildesley et al, 2010; Yoo et al., 2009).

Multidisciplinary teams usually involved at least two of the following: physicians, nurses, diabetes educators, endocrinologists, and dieticians. Multidisciplinary providers were directly involved in delivering the telehealth intervention in the IDEATel study and in nine other studies (Cho et al., 2006;  Yoon & Kim, 2008; Rodríguez-Idígoras et al., 2009; Dang et al., 2007; Istepanian et al., 2009a, b; Luzio et al., 2007; McMahon et al., 2005; Stone et al., 2010).

Intervention Time Requirements

Provider time requirements for intervention delivery were not reported systematically. In most cases, the provider time spent on analysis of monitoring values and delivering feedback was mentioned only as time cost approximates. Based on what was reported, the amount of time required by the provider in delivering feedback to/contacting patients is the briefest when the response is made via an online message or mobile text message (SMS); more time from the provider is required for phone calls, and videoconferencing sessions are the longest in duration, but brief feedback messages were generally delivered more frequently.

Provider response via web/mobile message

Cho et al. (2006; 2006b) reported that it took five minutes per patient for the physician to upload the patient-provided data, interpret it, and send a reply. This was done every two weeks; therefore a physician spent three to four hours biweekly responding to a group of 40 patients. Relative to face-to-face interviews, the time commitment required by the intervention was relatively brief.

The care manager in Ralston et al. (2009) reported that they spent an average of four hours per week updating care plans and communicating via secure email with 42 intervention group patients. The care manager was responsible for responding Monday to Friday, when necessary, to patients’ messages, and reviewing blood glucose levels at least once per week.

Providers in Levine et al. (2009) reported that they typically spent between 30 seconds to five minutes responding to patient messages; they were responsible for checking the monitoring website daily and responding to patients when necessary.

Provider response via telephone call

Telephone calls made to patients by care coordinators in Chumbler et al. (2009) generally lasted 15 to 30 minutes. There was no information provided regarding the frequency at which phone calls to patients were made.

Research assistants in Forjuoh et al. (2007; 2008) made 117 phone calls to 18 participants during the 6-month course of the study; each call lasted a mean of five minutes.

Teleconsultation

Reporting on an intervention that was part of the IDEATel study, an individual intervention group patient in Trief et al. (2006; 2009) received 28.3 ± 15.2 home televisits in the first year of the program (approximately biweekly). Each televisit lasted around 30 minutes on average.

Technical Requirements

Some type 2 diabetes telehealth monitoring interventions required an ehealth technology, such as a glucose meter or blood pressure cuff with transmission capabilities, or “dialogue boxes” like the Health Hero Buddy, with screens that featured prompted answers from patients regarding health status and disease knowledge. But most interventions used websites accessed via internet-enabled personal computers, mobile phones, landlines, or videoconferencing technology.

The interventions generally used only one technology for health service delivery, or a mixture of blood pressure cuffs and glucose meters for self-measurement plus a technology for transmission and/or communication. The exception to this was the IDEATel studies: patients participating in these interventions received a home telemedicine unit, which included an internet-enabled personal computer with videoconferencing capacities, a blood pressure cuff and glucose meter that transmitted values to the clinical database via the personal computer, and personal access to clinical data and a specialized educational webpage.

Algorithms were used in approximate one third of all included studies. The algorithms performed different functions. Some interventions used intelligent systems that generated alerts to providers when patients’ self-reported values were out of range (Rodríguez-Idígoras et al., 2009; Trudel et al., 2007; Luzio et al., 2007); prioritized monitoring reports to direct the providers attention to particular issues (Dang et al., 2007) and aided in provider decision-making (Shea et al., 2006, 2009; Bujnowska-Fedak et al., 2006; Kim et al., 2006; McMahon et al., 2005); delivered diabetes management dialogues with patients using dialogue boxes (Chumbler et al., 2005a,b); or provided feedback and educational messages to patients upon receipt of transmitted self-measured values (Yoo et al, 2009; Robertson et al., 2007).

Departures from Basic Model

The studies that did not fit the standard diabetes model delivered interventions that were categorized as health education, lifestyle support tools, and/or peer support. Most common was a web-based self-management program that offered a combination of health and diabetes education and lifestyle support tools, such as action plans, audio relaxation exercises, guided one-week modules, and health and exercise diaries (Forjuoh et al., 2007, 2008; Song et al., 2009; Lorig et al., 2010; Glasgow et al., 2010). The interventions in Lorig et al. (2010) and Song et al. (2010) also facilitated peer support with interactive bulletin boards and online discussion boards.

Sun et al. (2010) used an intervention that combined in-person education with telehealth-facilitated follow-up: patients received a five-day in-person diabetes education course at a local hospital, and then long-term follow-up advice from a community physician via telephone, SMS, or email.

Grant et al. (2008) was a unique intervention that gave patients access to an electronic personal health record in preparation for appointments with their physician. A question-and-answer algorithm helped patients to identify personal health concerns and produced a diabetes care plan; the intervention aimed to engage patients in personal diabetes management, involve them in care-related decision-making, and improve communication with providers.

Kim and Kang (2006), Timmerberg et al. (2008) and Izquierdo et al. (2010) were most similar to the standard model because they offered monitoring-based and consultation-based management, respectively. But Kim & Kang (2006) had nurses monitor physical activity rather than biometric data. A website supported patients’ personalized exercise plans and allowed patients to log information about exercise sessions. Izquierdo et al. (2010) was an intervention within the larger IDEATel study that delivered patient consultations with dieticians via videoconferencing with no transmission of self-measured values. Timmerberg et al. (2008), like Izquierdo et al. (2010), delivered group and individual patient videoconference sessions with a dietician.