Effects of tailored telemonitoring on functional status and health-related quality of life in patients with heart failure

The TEHAF study, a randomised controlled trial performed between 2007 and 2010, compared tailored telemonitoring using the Health Buddy to usual care in HF patients [10]. Further study details have been published elsewhere [7]. In brief, HF patients with New York Heart Association (NYHA) functional class 2–4 that were treated by a cardiologist and were in the care of a heart failure nurse (HFN) were included. HF was defined as ‘at least one episode of fluid retention requiring diuretics, either with an echocardiographic left ventricular ejection fraction ≤40% or a preserved ejection fraction with diastolic dysfunction’ [7]. Patients were included in the study either during a visit to the outpatient clinic of one of the participating centres (Sittard, Heerlen, Maastricht) or when they were visited by a HFN, in the home situation. Exclusion criteria were: unable to give informed consent, visual impairment, hearing impairment in combination with living alone, no command of the Dutch language, planned hospital admission within 3 months and/or chronic obstructive pulmonary disease, Parkinson’s disease, extracorporeal dialysis, (pre)dementia or another disease with an expected shortened lifespan. A total of 382 patients were included. The trial was completed by 301 patients (79%). All patients who did not finish the study were excluded from all further analyses. The follow-up period was 12 months. Before randomisation, written informed consent was obtained. Randomisation was computer-generated with stratification per centre. Four times during the study, data on patients’ self-care behaviour (measured with the European Heart Failure Self-Care Behaviour Scale) [11], self-efficacy (measured with the Barnason Efficacy Expectation Scale) [12] and compliance (measured with the Heart Failure Compliance Scale) [13] were collected (Fig. 1).

The primary objective of our reanalysis of the TEHAF study was to assess the effect of tailored telemonitoring on functional status, defined as change in metabolic equivalents (METS) scores, comparing an intervention group and a control group.

The secondary objective was to assess the effect of tailored telemonitoring on HRQoL in HF patients, defined as change in Borg rating of perceived exertion scale (Borg), visual analogue scale (VAS) and EuroQol five dimensions questionnaire (EQ-5D) scores during follow-up, comparing an intervention group and a control group. Additional secondary endpoints were mortality rate, number and length of HF-related hospital admissions, and number of outpatient clinic visits due to HF during 1 year of follow-up.

At baseline and after 12 months, patient characteristics, Borg, METS, VAS and EQ-5D scores were obtained as described below. Additionally, clinical data about mortality, number and length of hospital admissions and number of outpatient clinic visits due to HF were collected after 12 months.

The Health Buddy, an electronic device consisting of a display screen and four buttons, was assigned to the intervention group. On a daily basis, patients received disease-specific education and self-care support, as well as questions about their symptoms, knowledge and behaviour. First, all collected data were sent to a protected server. Then, the HF specialist could see the data via an i‑Care desktop. Based on the answers given, patients were classified into low-, medium- or high-risk profiles for clinical deterioration. If necessary, patients were contacted for further analysis and support. Every 3 months, and in-between if necessary, questions were readjusted on an individual basis. For example, patients with frequent hospital admissions were allocated to a program that focused more on symptom monitoring, whereas patients with a low disease-specific education level were allocated to a program that contained a highly intensive education program. European Society of Cardiology guidelines were recommended to treat patients in both groups [10].

METS describe the amount of energy used to perform a physical task at a certain level of intensity. It indicates patients’ global cardiopulmonary fitness. In this study, METS were obtained by using a dedicated questionnaire consisting of 39 items (see Electronic Supplementary Material). Patients were asked about their capability to and limitations in performing certain physical tasks. The more demanding the task, the higher the METS obtained [14]. Of all performable tasks, the one with the highest METS was regarded as the reference value for that person’s cardiopulmonary fitness.

EQ-5D is a scoring system to assess HRQoL. It consists of a questionnaire and a VAS. Using the questionnaire, patients’ current health status is depicted with regard to five different domains: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Based on the answers given, utility scores can be calculated to describe HRQoL. Utility scores can vary from −0.59 (‘health status worse than death’) to 1 (‘perfect health’) [15].

The VAS consists of a vertical line displaying values from 0 to 100, where 100 stands for ‘best imaginable health state’ and 0 represents ‘worst imaginable health state’ [16].

Borg is a diagnostic test used to rate the level of exertion of an individual. It consists of a questionnaire that can be used to evaluate the effort when performing a certain task [17, 18]. In the present study, the key question asked was: ‘How exhausting for you is taking a shower or bath?’ Potential answers and thus scores ranged from 6 (‘no effort’) to 20 (‘maximal effort’).

This study was approved by the local ethics committees of Maastricht (METC 15-4-195.1/ivb) and Heerlen/Sittard [METC 15-N-188(15-4-195)] University. The investigation conforms to the principles outlined in the Declaration of Helsinki [19].

Descriptive statistics were used to describe the baseline characteristics of the population, stratified by treatment group. To analyse differences in baseline variables, Student’s t-test (if data were normally distributed) or the Mann-Whitney test (if data were not normally distributed) was used for continuous variables and χ² for discrete variables. Single stochastic regression imputation [R code: mice(data, method = ‘pmm’, m = 1, maxit = 1, seed = 1)] was used to impute missing values for (total number of missing values in brackets): ejection fraction (4), duration of HF (19), pacemaker (13), NYHA class (12), marital status (2), educational level (16), left bundle branch block (6) and cause of HF (3) [20]. Results shown are with imputed numbers only. Excluding patients in whom some values were missing did not change the results (data not shown).

Linear regression was used to examine the impact of the following baseline covariates on HRQoL (i.e. EQ-5D utility score, VAS score, BORG score and METS score) measured 1 year later: baseline HRQoL, treatment group, ejection fraction, duration of HF, NYHA functional classification, pacemaker, marital status, age, sex, educational level.

Interaction terms were tested and forward selection was used to select covariates/interaction terms [using the stepAIC() function in R] using a p-value of 0.05 as selection criteria [k = qchisq (0.05, 1, lower.tail = F)].

Additionally, regression techniques were used to examine the impact of baseline HRQoL (i.e. EQ-5D utility score, VAS score) and of functional parameters (BORG score and METS score) on overall survival, number and length of hospital admissions and number of outpatient clinic visits due to HF (all after 1 year). Cox regression was used to analyse overall survival.

Functional status measured in METS remained unchanged in the intervention group, whereas it decreased in the control group over time. The changes over time were in favour of the intervention group (regression coefficient 0.318; p-value = 0.01). For VAS (p = 0.95), EQ-5D (p = 0.83) and Borg (p = 0.22), no differences over time between intervention and control group could be found (Tab. 2).

High VAS, EQ-5D utility and METS scores at baseline (T0) increased survival probability after 12 months (T12); the same was true for a low Borg score (Tab. 3). The hazard ratio for each point increase in METS for survival was 0.73 (95% CI 0.58–0.93), whereas the hazard ratio for Borg was 1.21 (95% CI 1.11–1.31). The hazard ratios for VAS and EQ-5D were 0.98 and 0.20, respectively (95% CI 0.96–0.99 and 0.07–0.54 respectively).

Functional status is an important outcome in HF patients given that exercise capacity is one of the key determinants affecting daily-life activities in these patients. In addition, it is a strong predictor of mortality in patients with cardiovascular disease [21]. Therefore, improving physical activity is an important aim to achieve in treating patients with HF.

This study showed that METS remained stable over time in the study group, whereas a decline of METS was observed in the control group. At the same time, patients in the intervention group had 50% fewer routine face-to-face contacts with a HF specialist. It can be hypothesised that the education program of the Health Buddy stimulates participants to improve their exercise capacity. Another possible explanation could be the fact that telemonitoring increased adherence to pharmacological and non-pharmacological treatment in this study population [7]. Also, self-care increased in the telemonitoring group. Both increased therapy adherence and self-care may have positively influenced cardiopulmonary fitness. Nevertheless, the TEHAF study was not designed for this outcome, making these findings very likely due to chance.

Giordano et al. showed that the application of a home-based telesurveillance program improved functional status (measured by NYHA class and left ventricular ejection fraction) and physical performance (measured by 6‑min walk) in patients with HF [22]. Instead of focusing on patient education, their program was mainly used for remote monitoring by collecting vital parameters and regular telephone consultations with a HFN. Remote monitoring was also integrated in the Health Buddy system and may be the most important component of telemonitoring in improving functional status in HF patients.

Skobel et al. showed that a remote-controlled exercise program was more effective in increasing functional status than conventional training methods [23]. Therefore, more comprehensive telemonitoring programs may help to further improve the functional capacity of HF patients, but this needs to be properly studied in prospective trials.

Improving the perceived burden of physical activity is important to HF patients as many feel tired, fatigued and unable to perform physical tasks. Furthermore, there is some discrepancy between objective measurement (e.g. cardiopulmonary exercise tests) and symptom perception (e.g. NYHA class) [24]. Therefore, it might be most important to improve the subjectively perceived burden of a particular domain to positively influence the patient’s well-being.

Using telemonitoring, no improvement of HRQoL in HF patients could be found in our study. This is consistent with the results of Balk et al., who used a TV-based telemonitoring program to improve HRQoL in HF patients [25]. Their intervention consisted of providing education and monitoring symptoms and behaviour, which made their results comparable to those of our study. A recently published randomised controlled trial (RCT) compared use of an e‑health adjusted care pathway and the heartfailurematters.org website to usual care in HF patients. No effect on HRQoL after 12 months of follow-up could be found [26].

Functional status and HRQoL at baseline predicted survival after 1 year of follow-up. It has been shown previously that METS [27], EQ-5D and VAS [28] can predict survival in HF patients. Therefore, these scores may be valuable in assessing patients’ prognosis during post-clinical follow-up.

In our METS questionnaire, the highest obtainable score was 8.0. Therefore, exercise capacity of very fit patients may be underestimated. However, only one patient achieved a METS of 8.0. The original TEHAF study was underpowered to detect differences in functional status or HRQoL; thus, our findings are only hypothesis-generating. Furthermore, the control group was well treated, making the detection of clinical effects of the intervention difficult. We did not use HF-specific questionnaires and no questionnaires were used to provide detailed information on all relevant domains of HRQoL. Also, our analysis was done according to the per protocol principle, which could have led to a selective exclusion of patients. Therefore, attrition bias is possible. Still, the numbers of patients that did not finish the study were comparable between the two groups.

In this study, we used the EQ-5D to measure HRQoL. Other instruments such as the Minnesota Living With Heart Failure questionnaire were developed more specifically for HF patients. Still, the EQ-5D is an appropriate tool to measure HRQoL. Also, a more general questionnaire like the EQ-5D has a broader applicability because comparison among different medical conditions and patient groups is possible.