Clinical effects of an optimised care program with telehealth in heart failure patients in a community hospital in the Netherlands

Congestive heart failure (CHF) is a complex clinical syndrome, which represents an increasingly common health problem for patients, healthcare providers and reimbursing agencies. The continuing ageing of the population is expected to increase the prevalence of heart failure in the Netherlands from 120,000 (~ 1 % of the population) in 2008 to approximately 200,000 in the coming decade [1]. In 2011, the number of hospital admissions with heart failure as the primary discharge diagnosis was almost 30,000 [1]. Of all patients who were admitted once for heart failure, 14 % of them were re-admitted to the hospital within 6 months after their first index admission [2].

Guidelines of the European Society of Cardiology (ESC) recommend a multidisciplinary approach that coordinates care along the continuum of CHF, often implemented as in-person follow-up visits [3]. Recently, alternative approaches such as telehealth and remote monitoring via cardiac implanted electronic devices [4, 5] have been proposed to assess physiological parameters related to CHF exacerbation more frequently, thereby enabling remote disease management [6]. Telehealth uses various information and communication technologies to assist in the management of an existing long-term medical condition of a patient at home by delivering clinical care and non-clinical services, including education on the disease and the capture and transmission of questionnaire data and vital signs. However, the optimal approach to non-invasive remote monitoring is uncertain, and an ESC guideline recommendation is not yet supported because the randomised controlled trials (RCTs) performed to date have given inconsistent results [3].

Telehealth for heart failure has had mixed and heterogeneous effects across studies, necessitating further evaluation within specific healthcare systems and patient populations. A systematic review on telehealth efficacy in RCTs concluded that telehealth reduced all-cause mortality by 34 % and CHF-related hospitalisations by 21 % and improvements in secondary endpoints such as quality of life and disease knowledge were also observed [7]. However, some studies included in the systematic review did not corroborate the beneficial effects of telehealth. The TEN-HMS study, the world’s first multi-centre RCT for telemonitoring in the UK, Germany and the Netherlands, showed that the use of telemonitoring reduced all-cause mortality by 29 % in a 240-day follow-up, though this was not significant when comparing with usual care [8]. In the HartMotief study, which was a multi-centre trial in eight Dutch hospitals, telemonitoring also failed to reduce mortality in a 288-day mean follow-up [9]. Furthermore, some later studies were not able to demonstrate telehealth efficacy. In the TIM-HF study, telehealth was implemented by means of a call centre and compared with an already excellent conventional care. In this study, telehealth failed to reduce all-cause mortality or a combined endpoint of cardiovascular mortality or hospitalisation in a mean follow-up of 21 months [10]. The Tele-HF study was unable to show reduction in mortality and admission for a structured telephone support system in a 6-month follow-up [11]. Poor patient compliance may have had an impact on the findings, as the use of the system was laborious for end users. The Whole System Demonstrator (WSD), a large RCT of telehealth involving 3230 patients with CHF, chronic obstructive pulmonary disease or diabetes mellitus in the UK, showed that telemonitoring lowered mortality and emergency admissions by 45 and 20 %, respectively, in comparison with usual care in a 1-year follow-up [12]. Cost-effectiveness analysis of WSD results did not show telehealth to be a cost-effective alternative to usual care [13], but unfortunately the decrease in mortality in the telehealth group was not taken into account in this analysis. Even though most available data suggest that telehealth is a promising strategy for improving disease management of CHF patients, more data are needed to determine the most advantageous approaches.

The first objective of this study was to assess the impact of telehealth-based disease management on unplanned admissions, disease severity and quality of life in CHF patients. The unique aspect of this study was its setting: a single community hospital in the north of the Netherlands, where cardiologists and nurse practitioners collaborate within the hospital outpatient heart failure unit. Furthermore, primary care, local home care agencies and nursing homes play a key role in the care protocol of CHF patients through their general practitioners, home care nurses and physicians. The second objective was to determine, which patient characteristics were predictive of unplanned admissions in a telehealth program. As unplanned readmissions are costly, a predictive score to quantify the risk of readmission would help clinicians to identify patients who might benefit from more intensive post-discharge care, even beyond telemonitoring.

In line with previous findings [19], the number of unplanned admissions for heart failure decreased in patients enrolled in a disease management program after telemonitoring was added to the program. Furthermore, blood pressure decreased independently of medication and NT-proBNP level decreased significantly, a trend that was observed earlier in internet-based home monitoring [20]. Finally, disease-specific knowledge was maintained at a high level and quality of life increased during the telehealth intervention.

Interventions using telehealth technology have the potential to reduce avoidable readmissions, and the evidence base for its beneficial effect is growing [12, 21]. However, various previous RCTs on the effect of telehealth showed disappointing results [10, 11, 22]. For instance, in the TEHAF study, the combined endpoint of heart failure admission and all-cause mortality was similar for both control and intervention groups. This could be in part attributed to already excellent care management in the control groups of these studies [23]. While multicentre RCTs are the gold standard of clinical evidence, a rigorous RCT is also very restrictive and telehealth results may be more superior when there is freedom in the implementation [24]. Therefore, reporting the results of this local telehealth project may contribute to the further development and critical evaluation of these systems in daily practice.

By providing education to promote change in patient behaviour for proper self-management, telehealth may have improved clinical indicators and quality of life [25]. Blood pressure readings improved over time in line with previous findings (see Paré et al. [26] for a review). We did not investigate if any medication adjustments triggered by telemonitoring affected blood pressure. Continuous surveillance of vital sign data in addition to usual care allowed for an immediate and adequate response to deviations from the accepted limits. As described previously [27], we found access to patient medical records for the staff who monitored the data to be a crucial factor in care coordination.

Predicting admissions in heart failure patients is difficult [28]; however, we showed that unplanned admissions in the next 12 months could be accurately predicted from baseline measurements for CHF patients who were in the telehealth program using multivariable logistic regression. Strong predictors included NYHA class IV and elevated NT-proBNP levels, which are indicative of increased disease severity. Emergency admissions may not be avoidable by a telehealth program alone in these patients.

Main limitation of the current study is the pre-post design without a control group. Therefore, one needs to be cautious when interpreting these results as being brought about by the telehealth program. The number of admissions on remeasurement is likely to be lower due to the statistical phenomenon of regression to the mean. An estimation of this effect [29] led us to believe that only 17 % of the observed reduction in unplanned admissions could have been explained by this effect. Furthermore, a learning effect may have occurred by sending the MLHF and DHFKS questionnaires every 8 and 12 weeks. In addition, based on the answers of the DHFKS, the nurse practitioner could decide to send the patient a relevant educational video. This may have influenced the DHFKS score as an outcome measure.

In the Netherlands, virtually all hospitals work with a CHF clinic staffed with either nurses or nurse practitioners, as it is a performance indicator monitored by the Dutch Healthcare Inspectorate [30]. Provided patient records are electronically available, the use of telemonitoring and implementation of its data in day-to-day care can be done without a substantial increase in workload for the nursing staff. However, contrary to nurse practitioners, nurses are not allowed to prescribe medication, which could impair a quick response.

In future, we are looking to further expand the current optimised care program with telehealth to other hospitals in the region. However, the inadequacy of reimbursing systems in the Netherlands to pay for this type of coordinated care constitutes a potential threat to its wider implementation, thereby possibly withholding improved chronic care from patients who could benefit and missing the opportunity to reduce overall costs of the care for CHF patients.