Health-related quality of life on tele-monitoring for users with pacemakers 6 months after implant: the NORDLAND study, a randomized trial

Controlled, randomized, non-masked clinical trial with patient assignment to conventional follow-up in the hospital or tele-monitoring follow-up by electronic data transmission with 6 months monitoring from the date of implantation.

The randomization process was performed as follows: A person unrelated to the study prepared a total of 50 sealed envelopes, 25 which included at note reading “tele-monitoring” and 25 with a note reading “hospital monitoring”. The envelopes were thoroughly mixed and numbered from 1 to 50. When a patient had accepted the invitation to participate in the study and signed the informed consent he was given a consecutive study number and allocated to follow-up in accordance with the specification included in the corresponding envelope. Thus, the investigators had no knowledge of or influence on the randomization result prior to inclusion.

The study was carried out in Nordland Hospital, Bodø, Norway, a hospital with a pacemaker center that serves a population of 170,000 inhabitants and performs around 80–90 pacemakers implants per year.

All patients who were scheduled for implant with a pacemaker between August of 2014 and October of 2015 and met all the inclusion criteria (age 18 years or older, ability to give informed consent and to operate the home monitor, life expectancy > 1 year) and none of the exclusion criteria (scheduled for implantable cardioverter-defibrillator (ICD) or cardiac resynchronization therapy (CRT), participation in other trials) were invited to participate. A total of 76 patients were screened and 50 patients were included and randomized to either telemonitoring (TM, n = 25) or hospital monitoring (HM, n = 25). No changes to methods were performed after trial commencement.

According to their diagnosis, patients received either a single (VVIR) or a dual chamber pacemaker (DDDR). Patients assigned to tele-monitoring received either a Biotronik Estella SR-T/ DR-T or a Biotronik Evia SR-T/ DR-T. Patients assigned to hospital monitoring received either one of the aforementioned Biotronik pacemakers, a St Jude Medical Endurity SR/DR or a Sorin Reply 200 SR/DR.

Home monitoring was performed through the Biotronik Home Monitoring® system, an internet-based TM service for users with Biotronik implantable heart devices. Biotronik devices equipped with Home Monitoring (T-devices) have additional storage capacity and contain a small RF-antenna for wireless communication and data transmission from the implant to a wireless patient device, the CardioMessenger. Home Monitoring has no negative effects on devices’ battery life [24].

Every night, the CardioMessenger automatically collects and transmits important, encrypted health information to the Biotronik service center using the global network of T-Mobile and its partners (GPRS). The transmitted patient data is collected, automatically analyzed and filtered at the Biotronik Home Monitoring Service Center, according to patients‘needs as defined by their physician. Health and system-related issues are ranked and marked in order of importance. All event and trend reports can be accessed and reviewed on a protected online platform. Furthermore, according to pre-set definitions, the physician can receive automatic warnings (e.g. by e-mail or text message) concerning safety issues such as premature battery depletion, lead fracture etc.

HRQoL of the users with pacemakers was assessed through the Norwegian version of the utility scores of the EuroQol-5D (EQ-5D) [25] from − 1 [the worst state of health] to 1 [the best state of health], and the self-rated HRQoL was analyzed through of the EQ-5D Visual Analogue Scale (EQ-5D-VAS), from 0 [worst imaginable health state or death] to 100 [perfect health]).

To measure the effects of symptoms, functional limitations and psychological reactions commonly associated with heart failure or its treatment on an individual’s quality of life, the Norwegian validated version of the MLHFQ. It consists of 21 questions regarding patients’ perception of the effects of HF on their daily lives; the score for each question ranges from 0 to 5, producing a total score of 0 to 105, with a higher score indicating a poorer HRQoL [26‐29].

Differences between TM and HM groups, in regards to: 1) Cardiovascular events (AE), including exitus, percutaneous coronary intervention, angina and lead dislodgement 2) monitoring visits and transmissions from home vs. hospital and 3) changes of medication and pacemakers reprogramming.

Patients were given 3 interviews (on the day of the pacemakers implant, and at 1 and 6 months after the implant) in which the corresponding questionnaires were administered. These interviews were carried out in a personalized way during hospitalization or by telephone interviews after discharge.

To ensure that the sample size was representative, due to the limited capacity of patient enrolment in the hospitaI, the following formula was used: \( \boldsymbol{n}=\frac{\boldsymbol{N.}{\boldsymbol{Z}}^{\mathbf{2}}.\boldsymbol{p}\left(\mathbf{1}-\boldsymbol{p}\right)}{\left(\boldsymbol{N}-\mathbf{1}\right){\boldsymbol{e}}^{\mathbf{2}}+{\boldsymbol{Z}}^{\mathbf{2}}.\boldsymbol{p}\left(\mathbf{1}-\boldsymbol{p}\right)} \) Thus, from a target population of n = 70, assuming a standard deviation of Z = 1.96, with a margin of error e = 0.10, a proportion of p = 0.5, and a 10% of possible losses, we got a size sample necessary of n = 45. We did not have any prior information on the value that was expected to be found. For this reason, 50% has been used as the statistical power (p = 0.5) meaning that, if there was any effect, it would be detected at the 50% of the total events.

Between August of 2014 and October of 2015, 50 patients (mean age: 75 ± 12 years, 48% women) were enrolled in the NORDLAND study, of which 25 were included in the TM group and 25 in the HM group (Fig. 1). Dizziness was the most common symptom, and the patients were most commonly referred from the hospital’s own cardiology ward or from other local hospitals. The most common pacing indication was sick sinus syndrome (48%) followed by atrioventricular block (40%). In 88% of the cases, DDDR pacemakers were used. Hypertension was the most frequent comorbidity (64%). None of the baseline characteristics, including HRQoL were different between the two groups (Table 1).

In basal analysis patients had a mean score of 0.78 (95%CI:0.72;0.85) for EQ5D utilities and 64.83 (95% CI:56.74;70.14) for EQ5D VAS, which was increased to 0.81 (95%CI: -0.75;0.87) and 67.80 (95%CI: 62.49;73.10) respectively at 1 month after PM implantation (Table 2). Changes between baseline and 1 month EQ5D utilities and VAS were not statistically significant (p = 0.81 and p = 0.20, respectively). TM and HM groups showed no significant differences in EQ5D utilities and VAS scores between enrollment and at month (Table 2), nor at any EQ5D utilities dimensions. Regarding MLHFQ, the patients had a mean score at baseline of 24.58 (95%CI:19.30;29.86). This improved significantly to 16.55 (95%CI:11.40–21.70) at the 1 month assessment (p <  0.05). There were no significant differences in MLHFQ scores between TM and HM groups (20.20 vs. 28.96; p = 0.07) at baseline or at 1 month of follow-up (18.64 vs. 14.38, p = 0.59) (Table 3).

After 6 months of follow-up, 6% of patients had experienced at least one AE (TM: 6%, HM: 4.2%; p = 0.40) (Table 4). After PM implantation, 38.2% of patients were hospitalized at least once (TM: 42%, HM: 33.4%; p = 0.55). The main causes of hospitalization were due to coronary problems (TM, 16% vs. HM, 20.8%) and PM dysfunction (TM, 12% vs. HM, 0%), however no significant differences between the groups were found. One patient (HM group) died for non-cardiovascular causes. As for workload, a mean of 1.20 ± 0.50 in-hospital visits per patient were registered during the follow-up period (TM:1.24 vs. HM:1.17; p = 0.26).

At 6 months, the EQ5D utilities score (Table 5) was increased to 0.79 (95%CI: 0.71;0.86; p = 0.54), 0.01 points more than baseline and 0.02 points less than the scores reached in month 1 (95% CI: -0.16;0.10; p = 0.64). On the EQ-5D VAS, the mean score reached 66.74 points, 1.91 points higher than in basal analysis (relative increase of 2.86%; p = 0.42) and 1.06 points less than the obtained in month 1 after implant (relative decrease of 1.59%; p = 0.72).

The MLHFQ scores (Table 5) revealed a significant improvement for both groups (− 9.65 points; 95% CI: 9.04–20.02), with significant differences compared to baseline scores (p <  0.001) and 2.02 points less than the obtained at month 1 after implant (reduction of 12.21%; P = 0.39). At the end of the study period, HM group experienced the better response with respect to month 0 (TM: − 4.4 vs. HM: − 15.75). At 6 months, there were no significant differences between the two groups (TM: 15.80 versus HM: 13.21; p = 0.97).

Among the main results of the NORDLAND study, we found that, a) there was not significant differences in HRQoL assessed through EQ-5D utilities and VAS questionnaires administered at 1 and 6 months after implant between users included in the two groups of follow-up, b) in MLHF questionnaire were found significant differences in both groups between the enrolment and the end of follow-up period, c) safety was similar between both groups and no significant differences were found, d) that number of visits to hospital for both groups was similar and no significant differences were found.

Several studies [10, 16‐18, 30] have assessed HRQoL in users with pacemakers. In these clinical trials, the SF-36 [16, 17, 30] and EQ5D [10, 18] questionnaires were used to assess HRQoL. Results were improved but no significant differences between both groups in either the effectiveness or the safety of TM and HM were found.

Coinciding with previous trials [16, 17, 30] there was not found significant improvement in health-related quality of life for both groups between the enrolment and the last month of the study. In other more recent studies [17, 18], users with pacemaker implants included in both groups reported significant improvements in HRQoL assessed through the EQ-5D [10] and SF-36 [18]. On the other hand, in NORDLAND study the MLHF questionnaire score was improved in the studied groups (21.78% vs. 54.42%) and at 6 months there were statistically significant differences between the groups. These results were similar to those obtained in other studies in which HRQoL questionnaires for specific heart-diseases such as MLHFQ [30], AQUAREL [18], Duke Activity Status Index (DASI) [10] were improved. In all the studies, similar results have been found regardless their research design, either randomized or non-randomized.

Previous studies [7, 10, 17, 30] showed a reduction in the percentage of in-office follow-up visits (31 to 58%). These results contrast with the obtained in the NORDLAND study where the visits to the hospital were 2% higher in TM group with respect to the other group (TM:1.24 vs. HM:1.17; p = 0.26). Additionally, in NORDLAND study, there were 98 transmissions scheduled for home patient (TM group) plus 10 extra ones (unscheduled transmissions).

The results on early detection of cardiovascular events and safety were similar to those obtained in previous studies [10, 16, 17, 31‐35]. There were no found significant differences among the two groups with regards to cardiovascular events. The cumulative number of cardiovascular events was slightly higher in the TM (6%) than in the HM group (4.2%) although no statistically significant differences were found. This is relevant since it has been demonstrated that either a high burden of atrial fibrillation can be an important predictor of stroke and even mortality [5, 11, 13]. On the other hand, results found in previous studies showed that users included in TM group were more satisfied, the number of in-office visits to hospital was reduced, feeling safer, and being closely monitored by competent health care professionals with negligible impact of tasks to be fulfilled to comply with monitoring [5, 18, 36, 37] and although the number of studies on HRQoL and costs performed up to present on telemonitoring of pacemakers is quite reduced, preliminary studies suggest costs-saving [15]. A very reduced number of studies on effectiveness in pacemakers telemonitoring have used the HRQoL (EQ5D and SF-36) [10, 16, 17] questionnaires.

To our knowledge, the NORDLAND study is the first randomized clinical trial (RCT) evaluating HRQoL in the tele-monitoring of users with pacemakers. The RCT is the most scientifically rigorous method of hypothesis testing available [38], and is regarded as the gold standard trial for evaluating the effectiveness of interventions [39]. Therefore, the NORDLAND study provides an important input for future systematic reviews and clinical decision-making with regards to the use of tele-monitoring of users with pacemakers. In addition, to our knowledge, the NORDLAND study is first using the Minnesota Living with Heart Failure Questionnaire, which offers an innovative approach in the field of internet-based health technologies assessments.

Despite the relevant results obtained, the study has some limitations. First, it is a trial where patients and health staff knew the type of monitoring assigned, possibly this aspect may have influenced in their behavior. However, such influence has been decreased as much as possible since the NORDLAND study is a randomized trial. This ensures major evidence level, less chance of bias due to random selection of the groups and it might be repeatable and comparable with other studies. Second, the final size of patients included in this trial (with a limited number of pacemakers implants per year) made it difficult to detect differences between the two groups. The lack of a larger final sample size and the limited differences between groups may have reduced the real statistical power. However, the proximity between the results found in both groups in the analyses performed do not indicate significant differences. In addition, our study is a single centre study which may also be an advantage as different clinical practice among different hospitals sometimes makes it difficult to conduct multicenter studies.

Moreover, the comparison and generalization of results to other settings may be supported as participant characteristics were quite similar to those reported in previous studies [2‐7, 10, 16, 17] in terms of age, gender, symptoms and indications.

Finally, it must be considered the phenomenon known as the Hawthorne effect [40, 41]. The fact that people tend to change their behavior when they are the targets of interest and attention, regardless of the specific nature of an intervention could be a limitation of our results. In such situations, patients become eager to please their physicians and make them feel successful. Additionally, patients wish to participate so that “good” results can be achieved in the study [40, 41].