The economic effects of home telehealth can be discussed using direct or indirect measures of costs and cost savings. Direct measures include figures such as cost per patient, cost per telehealth unit, and annual costs. Examples of indirect measures are probability of patient hospitalization, number of emergency room visits, and other types of health service use. Though a formal economic analysis would typically assign these services a fixed dollar value, other study designs often report changes in service use without attempting to translate these changes into costs or cost savings.
Though there are a number of cost analyses of home telehealth for heart failure management, most limit themselves to calculations of average cost per patient. It is generally unclear whether these estimates include equipment, training, or human resources. In addition, start-up costs are only occasionally separated from maintenance costs. Without this knowledge, it is exceedingly difficult to determine an intervention’s financial sustainability and potential for long-term cost savings. Crude extrapolation from the information available appears to show that cost savings are possible – generally as a result of decreased service use – but that an intervention is equally likely to require a greater investment than usual care. However, findings from four recent studies that found or projected substantial cost savings are encouraging. As these studies were retrieved in a final scan of 2011-2012 literature and were not subjected to the same level of analysis as the other studies included in this review, we will refer the reader to the studies for more information. See Almond et al. (2011), Dickerson et al. (2011), Mark et al. (2012), and Sohn et al. (2012) for more details.
Routine comparisons of mean and median costs per patient, another near-universal omission, might also be illuminating; in one study, a large difference between the two figures implies that a few patients accounted for a disproportionate amount of resource use. A fascinating series of sub-analyses from a recent study found that an intervention was more expensive than usual care among patients with of New York Heart Association (NYHA) classification III IV, but significantly less expensive than usual care when only NYHA II patients were included in the analysis. As this study was retrieved in a final scan of 2011-2012 literature and was not subject to the same level of analysis as the other studies included in this review, we will not attempt to expand on these conclusions. See Sohn et al. (2012) for more details.
Home telehealth has been seen to reduce the number and length of hospital stays in numerous well-designed and strongly executed studies, but the generalizability of this finding appears low. Significant benefit is seen in slightly over one-third of published cases.
As a general rule, home telehealth does not appear to result in increased or decreased use of primary care. Though a recent randomized controlled multi-centre trial found that contacts with primary caregivers were significantly reduced in patients using home telehealth, this study was retrieved in a final scan of 2011-2012 literature and was not subject to the same level of analysis as the other studies included in this review. The reader should consult the full text of the study (Josiane et al., 2012) for more details.
Most home telehealth interventions result either in no change in emergency services use or in a change favourable to the intervention group.
There is no indication at present that home telehealth changes use of specialist care. However, few studies have looked at this outcome that conclusions would be premature.
Emergency Services Use
Summary: Emergency services use was reported in 11 studies. There is strong evidence that most telehealth interventions result in either no change in emergency services use or a change favourable to the intervention group.
Study Details: Six studies found no significant differences between intervention and control groups in use of emergency or urgent care (Wakefield et al., 2008a; Woodend et al., 2008; Myers et al., 2006; Soran et al., 2008; Bowles et al., 2009; Kulshreshtha et al., 2010).
A significant difference favouring the intervention group was reported in 4 studies. Dansky et al. (2008) found a significantly lower percentage of patients with emergency department visits in Intervention Group 1 when compared with the control group at 60 days (30% vs. 18%, p<.05). Intervention Group 1 used a 2-way monitoring and videoconferencing system. Intervention Group 2 was provided with a 1-way monitoring system. No difference was observed between the Intervention Group 2 and the control group, nor did the difference between Intervention Group 1 and the control group persist to the 120-day endpoint.
Hudson et al. (2005), using a 1-group study design, found a decrease in emergency room visits per 1000 member years during a remote monitoring program. Length of enrollment in the program varied, but averaged over 12 months. The decrease in emergency services use was not significant at the program’s close, but did reach significance when measured from pre-intervention to the 6-month post-program follow up (1013.8 to 773.6, p=.01). It should be noted that clinical instability and/or recent use of emergency services was a criterion for study eligibility.
The findings of Hudson et al. (2005) are supported by the results reported in controlled studies by Kashem et al. (2008), who found a significantly lower rate of emergency room visits in the intervention group (p=.05), and Dar et al. (2009), who found significantly fewer emergency heart failure admissions in the intervention group (36% vs. 81%, p=.01).
Significant differences favouring the control group were found only in Cleland et al. (2005), who conducted a 3-arm study in which 426 enrollees were analyzed using intention-to-treat. At 450 days, the 168 patients assigned randomly to home telemonitoring had a higher number of emergency room visits per 1000 days than the usual care group (1.8 vs. 0.5). The number of visits in the third arm, which received usual care plus monthly contacts with a heart failure specialist nurse, was also elevated (1.6). Significance was not given. Interestingly, this difference did not translate into poorer clinical outcomes; both the telemonitoring group and the monthly contact group had significantly fewer days lost to death or hospitalization that the usual care group.
Summary: Number, time, and/or length of (re-) hospitalizations were reported in 19 studies. There is some evidence that telehealth can reduce number of (re-) hospitalizations: 7 of 18 studies found significant benefit. The strength of the evidence for reduced length of hospitalizations is similar, with 3 of 10 studies reporting shorter stays or fewer days lost to hospitalization among intervention patients. The remaining studies in all 3 groups find no significant differences.
Study Details: See Table C.6.3.3.: System Outcomes – (Re-) Hospitalizations, below, for details.
Recent Developments: A scan of material from 2011-2012, a time period not covered by our initial searches, found 1 additional article that addressed (re-)hospitalizations (Boyne et al., 2012). In this randomized controlled multi-centre trial, patient using home telemonitoring had significantly reduced hospital admissions. See Boyne et al. (2012) for more details.
Primary Care Use
Summary: Primary care use is reported in 6 studies. As a general rule, home telehealth does not appear to result in increased or decreased use of primary care.
Study Details: Four studies found no significant differences between groups in scheduled or unscheduled clinic visits (Kashem et al., 2008) or number of primary care visits (Dar et al., 2009; Woodend et al., 2009). Bowles et al. (2009), reporting descriptive data only, write that providers made an average of 9 in-person visits to each of 2 control groups and 11 to the telehealth intervention group.
Significant differences between study arms were found in Cleland et al. (2005), which used a 3-arm design. While no significant differences were found between the usual care group and the telehealth intervention group, the telephone intervention group had a significantly greater number of patient-provider contacts than the telehealth group at 450 days. Whether this is due to greater need or readier accessibility of providers is not clear.
In Myers et al. (2006), intervention participants had significantly fewer home medical visits per month than the comparison group (58 vs. 8.2, p<.0001). Note that historical data was used for the comparison group.
Recent Developments: A scan of material from 2011-2012, a time period not covered by our initial searches, found 1 additional article that addressed primary care use (Boyne et al., 2012). In this randomized controlled multi-centre trial, patient using home telemonitoring had significantly fewer contacts with caregivers. As this study was not subject to the same level of analysis as the other studies included in this review, we will not attempt further analysis. See Boyne et al. (2012) for more details.
Specialist Care Use
Summary: One study reports on specialist care use. There is no indication at present that home telehealth changes use of specialist care. However, few studies have looked at this outcome that conclusions would be premature.
Study Details: Woodend et al. (2008) found no significant difference between control and intervention groups in number of visits made to a cardiologist at 1 month, 3 months, or 12 months.
Telehealth Costs and Cost Comparisons
Summary: Telehealth costs and/or cost comparisons were found in 7 studies. Reporting in this area is poor; the means by which costs are calculated are rarely made explicit, with only 1 study giving a detailed breakdown of costs (Scalvini et al., 2005a). It is generally unclear whether cost estimates include equipment, training, or human resources. In addition, start-up costs are only occasionally separated from maintenance costs. Without this knowledge, it is exceedingly difficult to determine whether or not an intervention is financially sustainable. Routine comparisons of mean and median costs per patient, another near-universal omission, would also be illuminating; in the 1 study reporting this (Soran et al., 2010), the large difference between the 2 costs implies that a few patients accounted for a disproportionate amount of resource use.
Study Details: Lower overall costs in the control group were found in 3 studies. The size of the difference in cost was insignificant in Dar et al. (2009), and unreported in Balk et al. (2008). In Soran et al. (2010), the mean cost per patient was $17,838 (with a median of $8,901) in the intervention group vs. $15,967 (with a median of $9,772) in the control group.
In 4 studies, cost savings were found in the intervention group. The most detailed breakdown of costs and cost savings were provided in a study by Scalvini et al. (2005a), who gave separate figures for hardware and software costs, maintenance, provider time, and training. Total costs associated with the telehealth intervention, which was delivered to 230 patients over 1 year, were €75,426. Cost comparisons were given for hospitalization charges and telecare services. At 6 months, these charges were €180,744 in the control group vs. €116,952 in the intervention group and €0 in the control group vs. €14,429 in the intervention group, respectively. At 12 months, charges were €140,874 in the control group vs. €95,688 in the intervention group and €0 in the control group vs. €11,806 in the intervention group, again respectively.
Myers et al. (2006) estimates intervention group cost savings at $189.92 per patient over the 2-month monitoring period, while Bondmass (2007) notes that hospitalization and intervention costs are significantly higher in the control group (p=.033 and .001, respectively; effect sizes not given).
Hudson et al. (2005) conducted a 1-arm study with multiple pre- and post-measurements and found a significant reduction in costs per patient per month when comparing the intervention and post-intervention phases with the pre-intervention phase ($629.79, $508.54, and $1,002.81, respectively; p=.04 and .02, respectively).
Recent Developments: A scan of material from 2011-2012, a time period not covered by our initial searches, found 4 additional studies that addressed telehealth costs and cost comparisons (Almond et al., 2011; Dickerson et al., 2011; Sohn et al., 2012; Stampehl et al., 2012). Dramatic results were seen in a telehealth program described in Dickerson et al. (2011), which reported savings of $10,000 per patient over 6 months. Stampehl et al. (2012), though not reporting endpoint data, state that ‘preliminary data suggest substantially reduced costs’. Almond et al. (2011) come to slightly different conclusions in their examination of results from a home telehealth program in the UK. The authors report that initial costs were considerable, but that reduced use of medical services leads them to expect ‘substantial savings over the medium term.’ An interesting sub-analysis was conducted by Sohn et al. (2012), who carried out a controlled trial with patients using home telehealth units in conjunction with telephone follow-up. Overall costs were significantly lower in the intervention group when all patients were included in the analysis. This remained true in a sub-group analysis of patients with an NYHA II classification. However, costs were slightly higher in the intervention group when only NYHA III and IV patients were analyzed. As these studies were not subject to the same level of analysis as the other studies included in this review, we will not attempt further analysis. See Almond et al. (2011), Dickerson et al. (2011), Sohn et al. (2012), and Stampehl et al. (2012) for more details.
 Kulshreshtha et al. (2010) noted higher all-cause emergency room visits in the intervention group than in the control, although the difference was not significant. Interestingly, this trend was accompanied by a trend – again, not significant – toward lower rates of all-cause and heart failure-related readmission rates in the intervention group.