Internet-based device-assisted remote monitoring of cardiovascular implantable electronic devices: an evidence-based analysis
G Pron,L Ieraci,K Kaulback,Medical Advisory Secretariat, Health Quality Ontario
Abstract:Objective: The objective of this Medical Advisory Secretariat (MAS) report was to conduct a systematic review of the available published evidence on the safety, effectiveness, and cost-effectiveness of Internet-based device-assisted remote monitoring systems (RMSs) for therapeutic cardiac implantable electronic devices (CIEDs) such as pacemakers (PMs), implantable cardioverter-defibrillators (ICDs), and cardiac resynchronization therapy (CRT) devices. The MAS evidence-based review was performed to support public financing decisions. Clinical Need: Condition and Target Population: Sudden cardiac death (SCD) is a major cause of fatalities in developed countries. In the United States almost half a million people die of SCD annually, resulting in more deaths than stroke, lung cancer, breast cancer, and AIDS combined. In Canada each year more than 40,000 people die from a cardiovascular related cause; approximately half of these deaths are attributable to SCD. Most cases of SCD occur in the general population typically in those without a known history of heart disease. Most SCDs are caused by cardiac arrhythmia, an abnormal heart rhythm caused by malfunctions of the heart’s electrical system. Up to half of patients with significant heart failure (HF) also have advanced conduction abnormalities. Cardiac arrhythmias are managed by a variety of drugs, ablative procedures, and therapeutic CIEDs. The range of CIEDs includes pacemakers (PMs), implantable cardioverter-defibrillators (ICDs), and cardiac resynchronization therapy (CRT) devices. Bradycardia is the main indication for PMs and individuals at high risk for SCD are often treated by ICDs. Heart failure (HF) is also a significant health problem and is the most frequent cause of hospitalization in those over 65 years of age. Patients with moderate to severe HF may also have cardiac arrhythmias, although the cause may be related more to heart pump or haemodynamic failure. The presence of HF, however, increases the risk of SCD five-fold, regardless of aetiology. Patients with HF who remain highly symptomatic despite optimal drug therapy are sometimes also treated with CRT devices. With an increasing prevalence of age-related conditions such as chronic HF and the expanding indications for ICD therapy, the rate of ICD placement has been dramatically increasing. The appropriate indications for ICD placement, as well as the rate of ICD placement, are increasingly an issue. In the United States, after the introduction of expanded coverage of ICDs, a national ICD registry was created in 2005 to track these devices. A recent survey based on this national ICD registry reported that 22.5% (25,145) of patients had received a non-evidence based ICD and that these patients experienced significantly higher in-hospital mortality and post-procedural complications. In addition to the increased ICD device placement and the upfront device costs, there is the need for lifelong follow-up or surveillance, placing a significant burden on patients and device clinics. In 2007, over 1.6 million CIEDs were implanted in Europe and the United States, which translates to over 5.5 million patient encounters per year if the recommended follow-up practices are considered. A safe and effective RMS could potentially improve the efficiency of long-term follow-up of patients and their CIEDs. Technology: In addition to being therapeutic devices, CIEDs have extensive diagnostic abilities. All CIEDs can be interrogated and reprogrammed during an in-clinic visit using an inductive programming wand. Remote monitoring would allow patients to transmit information recorded in their devices from the comfort of their own homes. Currently most ICD devices also have the potential to be remotely monitored. Remote monitoring (RM) can be used to check system integrity, to alert on arrhythmic episodes, and to potentially replace in-clinic follow-ups and manage disease remotely. They do not currently have the capability of being reprogrammed remotely, although this feature is being tested in pilot settings. Every RMS is specifically designed by a manufacturer for their cardiac implant devices. For Internet-based device-assisted RMSs, this customization includes details such as web application, multiplatform sensors, custom algorithms, programming information, and types and methods of alerting patients and/or physicians. The addition of peripherals for monitoring weight and pressure or communicating with patients through the onsite communicators also varies by manufacturer. Internet-based device-assisted RMSs for CIEDs are intended to function as a surveillance system rather than an emergency system. Health care providers therefore need to learn each application, and as more than one application may be used at one site, multiple applications may need to be reviewed for alarms. All RMSs deliver system integrity alerting; however, some systems seem to be better geared to fast arrhythmic alerting, whereas other systems appear to be more intended for remote follow-up or supplemental remote disease management. The different RMSs may therefore have different impacts on workflow organization because of their varying frequency of interrogation and methods of alerts. The integration of these proprietary RM web-based registry systems with hospital-based electronic health record systems has so far not been commonly implemented. Currently there are 2 general types of RMSs: those that transmit device diagnostic information automatically and without patient assistance to secure Internet-based registry systems, and those that require patient assistance to transmit information. Both systems employ the use of preprogrammed alerts that are either transmitted automatically or at regular scheduled intervals to patients and/or physicians. The current web applications, programming, and registry systems differ greatly between the manufacturers of transmitting cardiac devices. In Canada there are currently 4 manufacturers—Medtronic Inc., Biotronik, Boston Scientific Corp., and St Jude Medical Inc.—which have regulatory approval for remote transmitting CIEDs. Remote monitoring systems are proprietary to the manufacturer of the implant device. An RMS for one device will not work with another device, and the RMS may not work with all versions of the manufacturer’s devices. All Internet-based device-assisted RMSs have common components. The implanted device is equipped with a micro-antenna that communicates with a small external device (at bedside or wearable) commonly known as the transmitter. Transmitters are able to interrogate programmed parameters and diagnostic data stored in the patients’ implant device. The information transfer to the communicator can occur at preset time intervals with the participation of the patient (waving a wand over the device) or it can be sent automatically (wirelessly) without their participation. The encrypted data are then uploaded to an Internet-based database on a secure central server. The data processing facilities at the central database, depending on the clinical urgency, can trigger an alert for the physician(s) that can be sent via email, fax, text message, or phone. The details are also posted on the secure website for viewing by the physician (or their delegate) at their convenience. Research Questions: The research directions and specific research questions for this evidence review were as follows: To identify the Internet-based device-assisted RMSs available for follow-up of patients with therapeutic CIEDs such as PMs, ICDs, and CRT devices.To identify the potential risks, operational issues, or organizational issues related to Internet-based device-assisted RM for CIEDs.To evaluate the safety, acceptability, and effectiveness of Internet-based device-assisted RMSs for CIEDs such as PMs, ICDs, and CRT devices.To evaluate the safety, effectiveness, and cost-effectiveness of Internet-based device-assisted RMSs for CIEDs compared to usual outpatient in-office monitoring strategies.To evaluate the resource implications or budget impact of RMSs for CIEDs in Ontario, Canada. Literature Search: The review included a systematic review of published scientific literature and consultations with experts and manufacturers of all 4 approved RMSs for CIEDs in Canada. Information on CIED cardiac implant clinics was also obtained from Provincial Programs, a division within the Ministry of Health and Long-Term Care with a mandate for cardiac implant specialty care. Various administrative databases and registries were used to outline the current clinical follow-up burden of CIEDs in Ontario. The provincial population-based ICD database developed and maintained by the Institute for Clinical Evaluative Sciences (ICES) was used to review the current follow-up practices with Ontario patients implanted with ICD devices. Search Strategy: A literature search was performed on September 21, 2010 using OVID MEDLINE, MEDLINE In-Process and Other Non-Indexed Citations, EMBASE, the Cumulative Index to Nursing & Allied Health Literature (CINAHL), the Cochrane Library, and the International Agency for Health Technology Assessment (INAHTA) for studies published from 1950 to September 2010. Search alerts were generated and reviewed for additional relevant literature until December 31, 2010. Abstracts were reviewed by a single reviewer and, for those studies meeting the eligibility criteria full-text articles were obtained. Reference lists were also examined for any additional relevant studies not identified through the search. Inclusion Criteria: published between 1950 and September 2010;English language full-reports and human studies;original reports including clinical evaluations of Internet-based device-assisted RMSs for CIEDs in clinical settings;reports including standardized measurements on outcome events such as technical success, safety, effectiveness, cost, measures of health care utilization, morbidity, mortality, quality of life or patient satisfaction;randomized controlled trials (RCTs), systematic reviews and meta-analyses, cohort and controlled clinical studies. Exclusion Criteria: non-systematic reviews, letters, comments and editorials;reports not involving standardized outcome events;clinical reports not involving Internet-based device assisted RM systems for CIEDs in clinical settings;reports involving studies testing or validating algorithms without RM;studies with small samples (<10 subjects). Outcomes of Interest: The outcomes of interest included: technical outcomes, emergency department visits, complications, major adverse events, symptoms, hospital admissions, clinic visits (scheduled and/or unscheduled), survival, morbidity (disease progression, stroke, etc.), patient satisfaction, and quality of life. Summary of Findings: The MAS evidence review was performed to review available evidence on Internet-based device-assisted RMSs for CIEDs published until September 2010. The search identified 6 systematic reviews, 7 randomized controlled trials, and 19 reports for 16 cohort studies—3 of these being registry-based and 4 being multi-centered. The evidence is summarized in the 3 sections that follow. 1. Effectiveness of Remote Monitoring Systems of CIEDs for Cardiac Arrhythmia and Device Functioning In total, 15 reports on 13 cohort studies involving investigations with 4 different RMSs for CIEDs in cardiology implant clinic groups were identified in the review. The 4 RMSs were: Care Link Network® (Medtronic Inc,, Minneapolis, MN, USA); Home Monitoring® (Biotronic, Berlin, Germany); House Call 11® (St Jude Medical Inc., St Pauls, MN, USA); and a manufacturer-independent RMS. Eight of these reports were with the Home Monitoring® RMS (12,949 patients), 3 were with the Care Link® RMS (167 patients), 1 was with the House Call 11® RMS (124 patients), and 1 was with a manufacturer-independent RMS (44 patients). All of the studies, except for 2 in the United States, (1 with Home Monitoring® and 1 with House Call 11®), were performed in European countries. The RMSs in the studies were evaluated with different cardiac implant device populations: ICDs only (6 studies), ICD and CRT devices (3 studies), PM and ICD and CRT devices (4 studies), and PMs only (2 studies). The patient populations were predominately male (range, 52%–87%) in all studies, with mean ages ranging from 58 to 76 years. One study population was unique in that RMSs were evaluated for ICDs implanted solely for primary prevention in young patients (mean age, 44 years) with Brugada syndrome, which carries an inherited increased genetic risk for sudden heart attack in young adults. Most of the cohort studies reported on the feasibility of RMSs in clinical settings with limited follow-up. In the short follow-up periods of the studies, the majority of the events were related to detection of medical events rather than system configuration or device abnormalities. The results of the studies are summarized below: The interrogation of devices on the web platform, both for continuous and scheduled transmissions, was significantly quicker with remote follow-up, both for nurses and physicians.In a case-control study focusing on a Brugada population–based registry with patients followed-up remotely, there were significantly fewer outpatient visits and greater detection of inappropriate shocks. One death occurred in the control group not followed remotely and post-mortem analysis indicated early signs of lead failure prior to the event.Two studies examined the role of RMSs in following ICD leads under regulatory advisory in a European clinical setting and noted:– Fewer inappropriate shocks were administered in the RM group.– Urgent in-office interrogations and surgical revisions were performed within 12 days of remote alerts.– No signs of lead fracture were detected at in-office follow-up; all were detected at remote follow-up.Only 1 study reported evaluating quality of life in patients followed up remotely at 3 and 6 months; no values were reported.Patient satisfaction was evaluated in 5 cohort studies, all in short term follow-up: 1 for the Home Monitoring® RMS, 3 for the Care Link® RMS, and 1 for the House Call 11® RMS.– Patients reported receiving a sense of security from the transmitter, a good relationship with nurses and physicians, positive implications for their health, and satisfaction with RM and organization of services.– Although patients reported that the system was easy to implement and required less than 10 minutes to transmit information, a variable proportion of patients (range, 9% 39%) reported that they needed the assistance of a caregiver for their transmission.– The majority of patients would recommend RM to other ICD patients.– Patients with hearing or other physical or mental conditions hindering the use of the system were excluded from studies, but the frequency of this was not reported.Physician satisfaction was evaluated in 3 studies, all with the Care Link® RMS:– Physicians reported an ease of use and high satisfaction with a generally short-term use of the RMS.– Physicians reported being able to address the problems in unscheduled patient transmissions or physician initiated transmissions remotely, and were able to handle the majority of the troubleshooting calls remotely.– Both nurses and physicians reported a high level of satisfaction with the web registry system. 2. Effectiveness of Remote Monitoring Systems in Heart Failure Patients for Cardiac Arrhythmia and Heart Failure Episodes Remote follow-up of HF patients implanted with ICD or CRT devices, generally managed in specialized HF clinics, was evaluated in 3 cohort studies: 1 involved the Home Monitoring® RMS and 2 involved the Care Link® RMS. In these RMSs, in addition to the standard diagnostic features, the cardiac devices continuously assess other variables such as patient activity, mean heart rate, and heart rate variability. Intra-thoracic impedance, a proxy measure for lung fluid overload, was also measured in the Care Link® studies. The overall diagnostic performance of these measures cannot be evaluated, as the information was not reported for patients who did not experience intra-thoracic impedance threshold crossings or did not undergo interventions. The trial results involved descriptive information on transmissions and alerts in patients experiencing high morbidity and hospitalization in the short study periods. 3. Comparative Effectiveness of Remote Monitoring Systems for CIEDs Seven RCTs were identified evaluating RMSs for CIEDs: 2 were for PMs (1276 patients) and 5 were for ICD/CRT devices (3733 patients). Studies performed in the clinical setting in the United States involved both the Care Link® RMS and the Home Monitoring® RMS, whereas all studies performed in European countries involved only the Home Monitoring® RMS. 3A. Randomized Controlled Trials of Remote Monitoring Systems for Pacemakers: Two trials, both multicenter RCTs, were conducted in different countries with different RMSs and study objectives. The PREFER trial was a large trial (897 patients) performed in the United States examining the ability of Care Link®, an Internet-based remote PM interrogation system, to detect clinically actionable events (CAEs) sooner than the current in-office follow-up supplemented with transtelephonic monitoring transmissions, a limited form of remote device interrogation. The trial results are summarized below: In the 375-day mean follow-up, 382 patients were identified with at least 1 CAE—111 patients in the control arm and 271 in the remote arm.The event rate detected per patient for every type of CAE, except for loss of atrial capture, was higher in the remote arm than the control arm.The median time to first detection of CAEs (4.9 vs. 6.3 months) was significantly shorter in the RMS group compared to the control group (P < 0.0001).Additionally, only 2% (3/190) of the CAEs in the control arm were detected during a transtelephonic monitoring transmission (the rest were detected at in-office follow-ups), whereas 66% (446/676) of the CAEs were detected during remote interrogation. The second study, the OEDIPE trial, was a smaller trial (379 patients) performed in France evaluating the ability of the Home Monitoring® RMS to shorten PM post-operative hospitalization while preserving the safety of conventional management of longer hospital stays. Implementation and operationalization of the RMS was reported to be successful in 91% (346/379) of the patients and represented 8144 transmissions.In the RM group 6.5% of patients failed to send messages (10 due to improper use of the transmitter, 2 with unmanageable stress). Of the 172 patients transmitting, 108 patients sent a total of 167 warnings during the trial, with a greater proportion of warnings being attributed to medical rather than technical causes.Forty percent had no warning message transmission and among these, 6 patients experienced a major adverse event and 1 patient experienced a non-major adverse event. Of the 6 patients having a major adverse event, 5 contacted their physician.The mean medical reaction time was faster in the RM group (6.5 ± 7.6 days vs. 11.4 ± 11.6 days).The mean duration of hospitalization was significantly shorter (P < 0.001) for the RM group than the control group (3.2 ± 3.2 days vs. 4.8 ± 3.7 days).Quality of life estimates by the SF-36 questionnaire were similar for the 2 groups at 1-month follow-up. 3B. Randomized Controlled Trials Evaluating Remote Monitoring Systems for ICD or CRT Devices: The 5 studies evaluating the impact of RMSs with ICD/CRT devices were conducted in the United States and in European countries and involved 2 RMSs—Care Link® and Home Monitoring ®. The objectives of the trials varied and 3 of the trials were smaller pilot investigations. The first of the smaller studies (151 patients) evaluated patient satisfaction, achievement of patient outcomes, and the cost-effectiveness of the Care Link® RMS compared to quarterly in-office device interrogations with 1-year follow-up. Individual outcomes such as hospitalizations, emergency department visits, and unscheduled clinic visits were not significantly different between the study groups.Except for a significantly higher detection of atrial fibrillation in the RM group, data on ICD detection and therapy were similar in the study groups.Health-related quality of life evaluated by the EuroQoL at 6-month or 12-month follow-up was not different between study groups.Patients were more satisfied with their ICD care in the clinic follow-up group than in the remote follow-up group at 6-month follow-up, but were equally satisfied at 12- month follow-up. The second small pilot trial (20 patients) examined the impact of RM follow-up with the House Call 11® system on work schedules and cost savings in patients randomized to 2 study arms varying in the degree of remote follow-up. The total time including device interrogation, transmission time, data analysis, and physician time required was significantly shorter for the RM follow-up group.The in-clinic waiting time was eliminated for patients in the RM follow-up group.The physician talk time was significantly reduced in the RM follow-up group (P < 0.05).The time for the actual device interrogation did not differ in the study groups. The third small trial (115 patients) examined the impact of RM with the Home Monitoring® system compared to scheduled trimonthly in-clinic visits on the number of unplanned visits, total costs, health-related quality of life (SF-36), and overall mortality. There was a 63.2% reduction in in-office visits in the RM group.Hospitalizations or overall mortality (values not stated) were not significantly different between the study groups.Patient-induced visits were higher in the RM group than the in-clinic follow-up group. The TRUST Trial: The TRUST trial was a large multicenter RCT conducted at 102 centers in the United States involving the Home Monitoring® RMS for ICD devices for 1450 patients. The primary objectives of the trial were to determine if remote follow-up could be safely substituted for in-office clinic follow-up (3 in-office visits replaced) and still enable earlier physician detection of clinically actionable events. Adherence to the protocol follow-up schedule was significantly higher in the RM group than the in-office follow-up group (93.5% vs. 88.7%, P < 0.001).Actionability of trimonthly scheduled checks was low (6.6%) in both study groups. Overall, actionable causes were reprogramming (76.2%), medication changes (24.8%), and lead/system revisions (4%), and these were not different between the 2 study groups.The overall mean number of in-clinic and hospital visits was significantly lower in the RM group than the in-office follow-up group (2.1 per patient-year vs. 3.8 per patient-year, P < 0.001), representing a 45% visit reduction at 12 months.The median time from onset of first arrhythmia to physician evaluation was significantly shorter (P < 0.001) in the RM group than in the in-office follow-up group for all arrhythmias (1 day vs. 35.5 days).The median time to detect clinically asymptomatic arrhythmia events—atrial fibrillation (AF), ventricular fibrillation (VF), ventricular tachycardia (VT), and supra-ventricular tachycardia (SVT)—was also significantly shorter (P < 0.001) in the RM group compared to the in-office follow-up group (1 day vs. 41.5 days) and was significantly quicker for each of the clinical arrhythmia events—AF (5.5 days vs. 40 days), VT (1 day vs. 28 days), VF (1 day vs. 36 days), and SVT (2 days vs. 39 days).System-related problems occurred infrequently in both groups—in 1.5% of patients (14/908) in the RM group and in 0.7% of patients (3/432) in the in-office follow-up group.The overall adverse event rate over 12 months was not significantly different between the 2 groups and individual adverse events were also not significantly different between the RM group and the in-office follow-up group: death (3.4% vs. 4.9%), stroke (0.3% vs. 1.2%), and surgical intervention (6.6% vs. 4.9%), respectively.The 12-month cumulative survival was 96.4% (95% confidence interval [CI], 95.5%–97.6%) in the RM group and 94.2% (95% confidence interval [CI], 91.8%–96.6%) in the in-office follow-up group, and was not significantly different between the 2 groups (P = 0.174). The CONNECT Trial: The CONNECT trial, another major multicenter RCT, involved the Care Link® RMS for ICD/CRT devices in a15-month follow-up study of 1,997 patients at 133 sites in the United States. The primary objective of the trial was to determine whether automatically transmitted physician alerts decreased the time from the occurrence of clinically relevant events to medical decisions. The trial results are summarized below: Of the 575 clinical alerts sent in the study, 246 did not trigger an automatic physician alert. Transmission failures were related to technical issues such as the alert not being programmed or not being reset, and/or a variety of patient factors such as not being at home and the monitor not being plugged in or set up.The overall mean time from the clinically relevant event to the clinical decision was significantly shorter (P < 0.001) by 17.4 days in the remote follow-up group (4.6 days for 172 patients) than the in-office follow-up group (22 days for 145 patients).– The median time to a clinical decision was shorter in the remote follow-up group than in the in-office follow-up group for an AT/AF burden greater than or equal to 12 hours (3 days vs. 24 days) and a fast VF rate greater than or equal to 120 beats per minute (4 days vs. 23 days).Although infrequent, similar low numbers of events involving low battery and VF detection/therapy turned off were noted in both groups. More alerts, however, were noted for out-of-range lead impedance in the RM group (18 vs. 6 patients), and the time to detect these critical events was significantly shorter in the RM group (same day vs. 17 days).Total in-office clinic visits were reduced by 38% from 6.27 visits per patient-year in the in-office follow-up group to 3.29 visits per patient-year in the remote follow-up group.Health care utilization visits (N = 6,227) that included cardiovascular-related hospitalization, emergency department visits, and unscheduled clinic visits were not significantly higher in the remote follow-up group.The overall mean length of hospitalization was significantly shorter (P = 0.002) for those in the remote follow-up group (3.3 days vs. 4.0 days) and was shorter both for patients with ICD (3.0 days vs. 3.6 days) and CRT (3.8 days vs. 4.7 days) implants.The mortality rate between the study arms was not significantly different between the follow-up groups for the ICDs (P = 0.31) or the CRT devices with defribillator (P = 0.46). Conclusions: There is limited clinical trial information on the effectiveness of RMSs for PMs. However, for RMSs for ICD devices, multiple cohort studies and 2 large multicenter RCTs demonstrated feasibility and significant reductions in in-office clinic follow-ups with RMSs in the first year post implantation. The detection rates of clinically significant events (and asymptomatic events) were higher, and the time to a clinical decision for these events was significantly shorter, in the remote follow-up groups than in the in-office follow-up groups. The earlier detection of clinical events in the remote follow-up groups, however, was not associated with lower morbidity or mortality rates in the 1-year follow-up. The substitution of almost all the first year in-office clinic follow-ups with RM was also not associated with an increased health care utilization such as emergency department visits or hospitalizations. The follow-up in the trials was generally short-term, up to 1 year, and was a more limited assessment of potential longer term device/lead integrity complications or issues. None of the studies compared the different RMSs, particularly the different RMSs involving patient-scheduled transmissions or automatic transmissions. Patients’ acceptance of and satisfaction with RM were reported to be high, but the impact of RM on patients’ health-related quality of life, particularly the psychological aspects, was not evaluated thoroughly. Patients who are not technologically competent, having hearing or other physical/mental impairments, were identified as potentially disadvantaged with remote surveillance. Cohort studies consistently identified subgroups of patients who preferred in-office follow-up. The evaluation of costs and workflow impact to the health care system were evaluated in European or American clinical settings, and only in a limited way. Internet-based device-assisted RMSs involve a new approach to monitoring patients, their disease progression, and their CIEDs. Remote monitoring also has the potential to improve the current postmarket surveillance systems of evolving CIEDs and their ongoing hardware and software modifications. At this point, however, there is insufficient information to evaluate the overall impact to the health care system, although the time saving and convenience to patients and physicians associated with a substitution of in-office follow-up by RM is more certain. The broader issues surrounding infrastructure, impacts on existing clinical care systems, and regulatory concerns need to be considered for the implementation of Internet-based RMSs in jurisdictions involving different clinical practices.