This study can contribute to shed some light on the impact of home telemonitoring of chronic elderly patients on healthcare resource use. Furthermore, this research provides evidence of the feasibility of the use of information and communication technology (ICT) applications by elderly patients with limited computer literacy. The results of this study show a positive effect of home telemonitoring of in-home patients with a high degree of comorbidity on healthcare resource usage.
The primary outcome measure was the number of hospitalisations at 12 months post-randomisation. In this respect, it was particularly striking that 42.9% of patients (9 people) in the IG completed the follow-up without any hospital admissions (the mean number of admissions in the previous year having been 3.4), and such figure was significantly different from the rate of 13.6% among the patients in the CG. This trend of a decrease in the number of hospital admissions is consistent with the results of other analysis we performed to assess hospitalisations, although in some cases the trend did not reach statistical significance: comparison of the number of admissions per patient; ORs for hospitalisations; NNTs and the comparison of the values in the year prior to inclusion relative to those during the 12 months of follow-up.
The observation that telemonitoring can reduce hospital admission is in agreement with the results of other benchmark studies. In the Whole System Demonstrator study, patients in the IG had fewer admissions, with an OR of 0.82 (95% CI 0.70 to 0.97) [8
]. Similarly, in the meta-analysis by Polisena and colleagues concerning HF [5
] three RCTs [16
] yielded an overall RR of 0.77 (95% CI 0.65 to 0.90), and in a meta-analysis on COPD by the same authors [6
] two studies [19
] demonstrated a decrease in the number of admissions of 32% to 46% among patients included in the IG. Further, another meta-analysis on HF [21
] found a protective effect with a RR of 0.93 for all-cause admissions and of 0.71 for specific admissions due to HF, the latter being statistically significant. The 2010 Cochrane review [4
] showed that telemonitoring decreased the number of total hospital admissions and admissions due to HF by 44% and 21%, respectively, thus confirming the findings of earlier research [21
The majority of the hospital admission that occurred during the completion of the study were cause-specific (81%), that is, for health problems related to one of the two health conditions considered in the present study (HF and CLD). In this respect, 84.8% of the admissions related to cardiovascular problems were due to HF, while 94.3% of respiratory-related admissions were due to respiratory exacerbations. This fact underlines the importance of a good clinical management of the aforementioned conditions and, especially, the significance of the early detection of the episodes of worsening or exacerbation, as the conditions with which patients have been diagnosed are responsible for most of the admissions. In this sense, it is essential that when applying telemonitoring interventions, a good disease management is combined with the deployment of the technology.
Additionally, a trend towards shorter hospital stays among telemonitored patients was also observed. In this regard, the mean days of hospital stay for those patients who completed 12 months of follow-up was of 9 days in the IG vs. 10.7 days in the CG, despite the differences observed were not statistically significant. Moreover, considering the total 121 hospital admissions that occurred during the study, the length of hospital stay due to all causes was 9.6 days vs. 12.2 days and number of days in hospital due to specific causes was 9.8 vs. 12.5, in the intervention and control groups, respectively, although the differences were not statistically significant. Several other authors have reported decreases in the length of hospital stay in patients with HF and COPD [4
], but results are not consistent across all studies [23
] and some RCTs have shown the opposite effect [22
Regarding the impact of home telecare on mortality, fewer deaths were observed among patients in the IG than in the CG (3 vs. 8 patients) and the mortality-rate was lower than would be expected with respect to estimates based on published data and our statistics from previous years for patients with similar characteristics. Nevertheless, the sample size of the present study is not large enough to draw solid conclusions. To date, several studies have found a lower mortality-rate when telemonitoring was used on HF patients [4
], but there is no consensus for the aforementioned condition [23
], and even less consistent results have been published for patients with COPD [6
During the course of the study, home telemonitoring of the trial patients led to changes in how care was delivered by the participating primary care health centres. In this respect, there were significantly more telephone contacts between telemonitored patients and their GPs and/or nurses (with a mean of 22.6 telephone calls per patient in the IG, vs. 8.5 in the CG) and fewer home visits, mainly due to significantly fewer home nursing visits (with a mean of 15.3 nursing visits per patient in the IG vs. 25.4 in the CG). A very slight increase in home visits by GPs was however observed in the IG, which we believe may be influenced by a closer delivery of care and increased initial actions undertaken by doctors, until the participating GPs became familiar with the telemonitoring process. Thus, assuming that the intervention is effective decreasing the number of hospitalisations, such redistribution of healthcare activities and resources may entail savings, considering the costs associated with telephone contacts compared to home visits. On the other hand, these changes could have a significant impact on the way primary care professionals work, leaving more time for nurses to spend on other relevant task that could improve the management of this type of patients. Nonetheless, the impact of such operational modifications has not been directly addressed in this study and requires further investigation. Few studies have assessed the effect of home telemonitoring on telephone calls and home visits and although some studies have been consistent with our findings [20
], others were not [4
We believe that when assessing the impact and overall implications of home telemonitoring it is essential to consider the global effects of the technology itself in addition to the usage of healthcare resources. Thus, other complex factors such as, the organizational changes, standards and perceptions of health and safety, effect on the patients’ QoL, economical implications, patients’ and healthcare professionals’ satisfaction as well as the effect of telemonitoring on family caregivers should also be taken into account. Some of these aspects have been assessed by our team and will be published shortly. A cost-effectiveness analysis of the TELBIL study has also been undertaken [25
We would like to highlight three key characteristics that differentiate this study from others and make it particularly relevant. Firstly, the telemonitoring was managed by primary care professionals (GPs and nurses) who regularly see the patients in the health centres or at home. The fact that primary care professionals are in charge of the telemonitoring intervention is particularly important, since these are the healthcare professionals that routinely carry out the follow-up of in-home patients and, thus, telemonitoring could have a greater positive impact than when applied to hospital-based interventions. In this regard, the integration of the new telemonitoring intervention to the routine practice at the health centres could improve the care provided. To the best of our knowledge, this is one of the first RCTs in which primary care professionals are in charge of the telemonitoring procedure. The implication of primary healthcare professionals has been further explored in the present study. On the one hand, the implication of around 70 primary care professionals has enabled us to undertake a qualitative analysis with focus groups to further assess the satisfaction and specific contribution of the participating GPs and nurses [26
]. On the other hand, the factors related to the healthcare professionals’ acceptance of the new telemonitoring technology have been evaluated through and extension of the Technology Acceptance Model (TAM), showing that the perception of facilitators in the organisational context is the most important variable to consider for increasing healthcare professionals’ intention to use the telemonitoring technology [27
Secondly, the patients included in the present study had challenging characteristics: they had a higher mean age (81 years) than that targeted by most other published studies [4
] and were patients with advanced diseases and high levels of comorbidity (51.7% were under home oxygen therapy, 46.5% had both of the diseases targeted in this study, 86.2% scored high comorbidity (>2) on the Charlson index and 79.3% showed signs of clinical deterioration) and, in line with these features, the patients included in the trial were heavy users of healthcare resources. In the view of the above mentioned peculiarities, this trial demonstrates the feasibility of implementing telemonitoring interventions as an alternative mode of health care provision for medically unstable patients with high degree of physical and functional deterioration. Furthermore, there were only 5 individuals not wishing to continue using the telemonitoring system and even these had all successfully managed to handle the devices. The observed low dropout-rate contrasts with other research into the feasibility and perceptions of this technology, which suggests that the older the patient and/or caregiver the more obstacles to the adoption of telemonitoring [28
Thirdly, we have focussed on two common chronic disorders (HF and CLD), which have been managed using the same telemonitoring system and in a very similar way, while practically all previous RCTs on telemonitoring technology have targeted a single chronic condition [29
]. This point is important, since the levels of comorbidity among this kind of patients make it difficult to consider a single condition in isolation from other existing health problems. Hence, the telemonitoring management approach should be adapted to reflect the clinical and functional status of the patients, rather than focussed on a specific disease.
The results of our study should be interpreted in light of some limitations. First, the number of patients included in the study was limited by the available telemonitoring devices. Twenty eight patients received telemonitoring and 21 patients in the IG and 22 patients in the CG completed the follow-up. We have, therefore, observed a deviation in the estimated power mainly due to an increase in the expected losses and a lower reduction in the number of hospital admissions than previously envisaged. We have recalculated the statistical power, taking into account the results obtained, with a confidence level of 95% in a bilateral contrast and 21 patients in the IG and 22 in the CG. Thus, the present study has a statistical power of 58% to detect significant differences between the percentages of patients with ≥1 hospital admissions in the IG (57.1%) and in the CG (86.4%).
Another limitation of this study is that due to the interactive nature of the intervention, it was not possible to blind the health care professionals providing the intervention or the participants involved in the study. However, despite this limitation, most of the data presented in this manuscript are objective and have been obtained from medical registers. The veracity of the data obtained has been double-checked and the statisticians in charge of the data analysis have been blinded to group assignment.