Telephone Self-management Interventions for BP Control
Telephone Self-management Interventions for BP Control
Telehealth disease management programs are increasingly implemented in clinical practice but with little empirical evidence about their effectiveness on patient or cost outcomes. In our evaluation of the HINTS trial, patients with home BP telemonitoring and receiving either a nurse-administered behavioral intervention or a nurse/physician medication management intervention showed significant gains in BP control at 12 months. At 18 months, BP control for intervention patients overall was no different than usual care. In subgroup analysis, combined intervention resulted in significant decreases in short-term and longer term BP outcomes among patients with poor baseline BP control, which suggests a benefit of intervention for individuals who may have been more appropriate for intervention.
Our cost analysis results inform a nascent evidence base on chronic disease management modeled in a telephone- and home-based framework. Combined intervention resulted in observed net savings in outpatient care and overall medical care as well as the lowest mean cost difference and total cost, but these differences were not statistically significant relative to the other intervention arms. Nurse contact may have accounted for some difference in intervention costs, but this is unlikely because nurse time costs for the combined intervention were the second least costly of the 3 interventions. The similarities in total VA costs across the HINTS interventions are not surprising, given that VA medical costs were substantially larger than the HINTS intervention costs and that there were no significant differences in health care use. Furthermore, the large variation in medical costs prevented sufficient precision to identify differences with certainty. Altogether, these findings suggest that, unless the HINTS interventions improve BP control and reduce the actual number of clinic events, patient care costs will not change.
Although implementation costs of telephone-based monitoring and disease management were not statistically significantly higher than usual care total costs (ie, intervention and direct medical costs), intervention costs were nontrivial. The HINTS intervention was costly and time-consuming to deliver, especially for interventions involving medication management that required nurse and physician time. From a program cost perspective, implementation on a larger scale likely would entail lower health system costs than seen in this trial. First, nurse costs in this trial were higher than they would be if the HINTS interventions were implemented in a medical center or health system because nurses would be spending time on other clinic tasks rather than the HINTS trial protocols. Second, although intervention start-up and supply costs constituted roughly half of total intervention costs per patient in the HINTS trial, lower cost scenarios that mirror discounts and economies of scale (eg, volume discounts on home BP monitors or reduced training costs for personnel) would lower intervention costs when scaled for wider implementation. These cost reductions would need to be weighed against the health benefits of gains in BP control that are achieved over time.
A limitation of this study merits acknowledgment. We only had access to VA data files when determining postintervention health care use and medical costs, and thus, we were unable to account for health care services provided outside VA (eg, VA-contract care, Medicare). However, VA was likely to be the principle source of health care for the HINTS patients because of the trial's eligibility requirement of active use of VA primary care and study participation that required visits to VA clinics.
Our cost analysis of the HINTS interventions sheds light on several important considerations for wide-scale implementation of telephone-based self-management programs. Although the overall costs of implementing the HINTS interventions were not significantly different from that of usual care, implementation costs are still nontrivial. Health systems are likely to find that intervention costs may be further reduced for wider scale implementation. Second, for telephone-based intervention management programs to be worth their investment, they must create significant improvement in clinical outcomes. The findings from our subgroup analysis suggests that, to optimize investment in home and telephone-based interventions such as the HINTS, it is critically important for health systems to consider applying interventions to specific patient subpopulations that are most likely to benefit from intensive home-based monitoring and self-management efforts. Third, implementation should also consider whether intervention generates other patient-centered outcomes or efficiencies in other aspects of medical care. With overall improvements in 12-month clinical outcomes that were not sustained at 18 months, the long-term return on investment in clinical outcomes remains unclear. It is possible that short-term gains in 12-month BP control produced downstream effects in medical care savings (ie, because of slowing progression of chronic disease) that were not observed in this study. To inform implementation efforts and the refinement of existing programs, future research should quantify the indirect benefits and costs of interventions, assess longer term BP control and health care use and cost outcomes, and examine the extent to which subpopulations may benefit from telephone-based chronic disease management interventions such as those tested in the HINTS.
Discussion
Telehealth disease management programs are increasingly implemented in clinical practice but with little empirical evidence about their effectiveness on patient or cost outcomes. In our evaluation of the HINTS trial, patients with home BP telemonitoring and receiving either a nurse-administered behavioral intervention or a nurse/physician medication management intervention showed significant gains in BP control at 12 months. At 18 months, BP control for intervention patients overall was no different than usual care. In subgroup analysis, combined intervention resulted in significant decreases in short-term and longer term BP outcomes among patients with poor baseline BP control, which suggests a benefit of intervention for individuals who may have been more appropriate for intervention.
Our cost analysis results inform a nascent evidence base on chronic disease management modeled in a telephone- and home-based framework. Combined intervention resulted in observed net savings in outpatient care and overall medical care as well as the lowest mean cost difference and total cost, but these differences were not statistically significant relative to the other intervention arms. Nurse contact may have accounted for some difference in intervention costs, but this is unlikely because nurse time costs for the combined intervention were the second least costly of the 3 interventions. The similarities in total VA costs across the HINTS interventions are not surprising, given that VA medical costs were substantially larger than the HINTS intervention costs and that there were no significant differences in health care use. Furthermore, the large variation in medical costs prevented sufficient precision to identify differences with certainty. Altogether, these findings suggest that, unless the HINTS interventions improve BP control and reduce the actual number of clinic events, patient care costs will not change.
Although implementation costs of telephone-based monitoring and disease management were not statistically significantly higher than usual care total costs (ie, intervention and direct medical costs), intervention costs were nontrivial. The HINTS intervention was costly and time-consuming to deliver, especially for interventions involving medication management that required nurse and physician time. From a program cost perspective, implementation on a larger scale likely would entail lower health system costs than seen in this trial. First, nurse costs in this trial were higher than they would be if the HINTS interventions were implemented in a medical center or health system because nurses would be spending time on other clinic tasks rather than the HINTS trial protocols. Second, although intervention start-up and supply costs constituted roughly half of total intervention costs per patient in the HINTS trial, lower cost scenarios that mirror discounts and economies of scale (eg, volume discounts on home BP monitors or reduced training costs for personnel) would lower intervention costs when scaled for wider implementation. These cost reductions would need to be weighed against the health benefits of gains in BP control that are achieved over time.
A limitation of this study merits acknowledgment. We only had access to VA data files when determining postintervention health care use and medical costs, and thus, we were unable to account for health care services provided outside VA (eg, VA-contract care, Medicare). However, VA was likely to be the principle source of health care for the HINTS patients because of the trial's eligibility requirement of active use of VA primary care and study participation that required visits to VA clinics.
Our cost analysis of the HINTS interventions sheds light on several important considerations for wide-scale implementation of telephone-based self-management programs. Although the overall costs of implementing the HINTS interventions were not significantly different from that of usual care, implementation costs are still nontrivial. Health systems are likely to find that intervention costs may be further reduced for wider scale implementation. Second, for telephone-based intervention management programs to be worth their investment, they must create significant improvement in clinical outcomes. The findings from our subgroup analysis suggests that, to optimize investment in home and telephone-based interventions such as the HINTS, it is critically important for health systems to consider applying interventions to specific patient subpopulations that are most likely to benefit from intensive home-based monitoring and self-management efforts. Third, implementation should also consider whether intervention generates other patient-centered outcomes or efficiencies in other aspects of medical care. With overall improvements in 12-month clinical outcomes that were not sustained at 18 months, the long-term return on investment in clinical outcomes remains unclear. It is possible that short-term gains in 12-month BP control produced downstream effects in medical care savings (ie, because of slowing progression of chronic disease) that were not observed in this study. To inform implementation efforts and the refinement of existing programs, future research should quantify the indirect benefits and costs of interventions, assess longer term BP control and health care use and cost outcomes, and examine the extent to which subpopulations may benefit from telephone-based chronic disease management interventions such as those tested in the HINTS.
Source...