- Email: [email protected]
Telehealth by an Interprofessional Team in Patients With CKD: A Randomized Controlled Trial
Original Investigation Telehealth by an Interprofessional Team in Patients With CKD: A Randomized Controlled Trial Areef Ishani, MD, MS,1,2 Juleen Christopher, PhD,1 Deirdre Palmer, MD,1 Sara Otterness, NP,1 Barbara Clothier, MS,3 Sean Nugent, BA,3 David Nelson, PhD,3 and Mark E. Rosenberg, MD,2 on behalf of the Center for Innovative Kidney Care* Background: Telehealth and interprofessional case management are newer strategies of care within chronic disease management. We investigated whether an interprofessional team using telehealth was a feasible care delivery strategy and whether this strategy could affect health outcomes in patients with chronic kidney disease (CKD). Study Design: Randomized clinical trial. Setting & Participants: Minneapolis Veterans Affairs Health Care System (VAHCS), St. Cloud VAHCS, and affiliated clinics March 2012 to November 2013 in patients with CKD (estimated glomerular filtration rate , 60 mL/min/1.73 m2). Interventions: Patients were randomly assigned to receive an intervention (n 5 451) consisting of care by an interprofessional team (nephrologist, nurse practitioner, nurses, clinical pharmacy specialist, psychologist, social worker, and dietician) using a telehealth device (touch screen computer with peripherals) or to usual care (n 5 150). Outcomes: The primary end point was a composite of death, hospitalization, emergency department visits, or admission to skilled nursing facilities, compared to usual care. Results: Baseline characteristics of the overall study group: mean age, 75.1 6 8.1 (SD) years; men, 98.5%; white, 97.3%; and mean estimated glomerular filtration rate, 37 6 9 mL/min/1.73 m2. Telehealth and interprofessional care were successfully implemented with meaningful engagement with the care system. One year after randomization, 208 (46.2%) patients in the intervention group versus 70 (46.7%) in the usual-care group had the primary composite outcome (HR, 0.98; 95% CI, 0.75-1.29; P 5 0.9). There was no difference between groups for any component of the primary outcome: all-cause mortality (HR, 1.46; 95% CI, 0.42-5.11), hospitalization (HR, 1.15; 95% CI, 0.80-1.63), emergency department visits (HR, 0.92; 95% CI, 0.68-1.24), or nursing home admission (HR, 3.07; 95% CI, 0.71-13.24). Limitations: Older population, mostly men, potentially underpowered/wide CIs. Conclusions: Telehealth by an interprofessional team is a feasible care delivery strategy in patients with CKD. There was no statistically significant evidence of superiority of this intervention on health outcomes compared to usual care. Am J Kidney Dis. 68(1):41-49. Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. This is a US Government Work. There are no restrictions on its use. INDEX WORDS: Telemedicine; case management; chronic kidney disease (CKD); interprofessional relations; chronic disease management; video monitoring; virtual visit; remote monitoring; patient education; hypertension; hospitalization; mortality; randomized controlled trial.
Editorial, p. 5
S
trategies are needed to improve the care of patients with chronic disease. This care should reduce adverse health outcomes, provide a timely and convenient care experience for patients regardless of their
From the 1Section of Renal Diseases and Hypertension, Minneapolis Veterans Affairs Health Care System; 2Division of Renal Diseases and Hypertension, University of Minnesota Medical School; and 3Center for Chronic Disease Outcomes Research, Minneapolis Veterans Affairs Health Care System, Minneapolis, MN. * A list of Center for Innovative Kidney Care members appears in the Acknowledgements. Received September 29, 2015. Accepted in revised form January 14, 2016. Originally published online March 1, 2016. Trial registration: www.ClinicalTrials.gov; study number: NCT01446029. Am J Kidney Dis. 2016;68(1):41-49
location, and be of high value with the potential to reduce overall health system costs.1 Telehealth is an example of such a strategy. It has been used, with or without case management, in various forms to manage patients with chronic illnesses such as chronic obstructive pulmonary disease, heart failure, and diabetes mellitus.2-6 Despite its growing use, telehealth outcomes have been variable Address correspondence to Areef Ishani, MD, MS, Department of Medicine, Nephrology Section, Minneapolis VA Medical Center, One Veterans Dr (111J), Minneapolis, MN 55417. E-mail: [email protected] Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. This is a US Government Work. There are no restrictions on its use. 0272-6386 http://dx.doi.org/10.1053/j.ajkd.2016.01.018
41
Ishani et al
and its expense is often considerable, emphasizing the need to carefully assess its effectiveness.7,8 Previous studies have suggested that interprofessional case management or home monitoring could improve intermediate outcomes such as blood pressure in patients with chronic kidney disease (CKD).9-11 The feasibility and effectiveness of such interventions on patientcentered outcomes is unclear. We conducted a randomized controlled trial to determine whether an intervention consisting of telehealth with interprofessional team case management could be effectively implemented and whether it could improve the combined end point of death, hospitalization, emergency department visits, or admission to a nursing home in patients with moderate to severe CKD compared to usual care.
METHODS Patients We randomly assigned 451 participants to receive the intervention and 150 participants to receive usual care from the Minneapolis Veterans Affairs Health Care System (VAHCS), the St. Cloud, MN, VAHCS, and their affiliated community-based clinics. A registry of patients with CKD (estimated glomerular filtration rate [eGFR] , 60 mL/min/1.73 m2 based on the most recent creatinine value within past year) was created. From March 2012 through October 2012, patients older than 18 years whose most recent eGFR (calculated using the IDMS-traceable 4-variable MDRD [Modification of Diet in Renal Disease] Study equation) within the last year was ,60 mL/ min/1.73 m2 were identified from the registry and invited to participate in the study. Patients from rural areas were oversampled.12 Patients who were unable to give consent, had life expectancy less than 1 year, lived in a skilled nursing facility, or had a primary care provider unwilling to allow participation were excluded. Potential participants were invited to be in the study by mail and then by telephone. Interested individuals were scheduled for an inhome visit, during which the written informed consent process was completed, baseline measurements were collected, and participants were randomly assigned to receive the intervention or usual care. Participants randomly assigned to the intervention received a second in-home visit during which the video monitoring device with peripherals and broadband (most commonly cellular connection) were installed. Participants in the usual-care group were invited to attend a CKD education class and asked to follow up with their primary care providers regarding management of their kidney disease. Participants assigned to usual care may or may not have been seen by a nephrologist during the course of the trial, at the discretion of their primary care provider. The exact care received by the usual-care group was not investigated.
Randomization and Masking Eligible patients were randomly assigned to receive the intervention or usual care using a centralized computer-generated randomization scheme using permuted block sizes of 2, 4, or 6. Randomization was stratified by eGFR (,30 vs $30 mL/min/ 1.73 m2), presence of diabetes, and occurrence of a hospitalization in the past year. Randomization occurred over the telephone by an individual blinded to patient identity. Although participants could not be masked to their assignment, outcome assessors were blinded to participant study assignments.
Procedures Participants in the intervention group received in-home training regarding how to use the device (LifeView; AmericanTeleCare) 42
and all the peripherals (blood pressure cuff, scale, glucometer, pulse oximeter, stethoscope, and web camera) and how to contact the clinical team. An interprofessional team consisting of a nephrologist, nurse practitioner, nurses, clinical pharmacy specialist, psychologist, social worker, telehealth care technician, and dietician reviewed the health status of each intervention participant and developed a patient-specific treatment plan addressing short- and long-term goals. The goal of the intervention program was comprehensive care of CKD and comorbid conditions using components of the chronic care model.13 Specific issues addressed included the following: management of blood pressure, volume status, proteinuria, diabetes mellitus, lipid levels, and depression; health literacy; and patient activation.14 The program also addressed lifestyle modification (physical activity, diet, weight reduction, and smoking cessation). A customized education program was developed based on each patient’s comorbid conditions and was delivered over broadband to the device. Patients could interact with the educational modules at their own learning pace. Patients were also given a customized self-monitoring strategy based on their clinical condition. Vital signs were automatically measured by the device and transmitted to the study team. The clinical team met daily to discuss high-risk patients. The full interprofessional team met weekly to review intervention patients and their progress toward goals, hospitalizations, and risks for recurrent or adverse events. Nurses also reviewed all vital signs and module responses transmitted to the study team twice a day. Vital signs were categorized as within or beyond range based on individualized patient parameters. Nurses addressed vital signs that were beyond the prespecified range with the assistance of a provider when clinically indicated. This typically involved interactive video conferencing (a virtual visit) with the patient to review the clinical situation, and possibly a medication change. Medication management was conducted by nurses, a clinical pharmacy specialist, and providers for all intervention participants based on guidelines15 and self-measured values of blood pressure, blood glucose, and volume status. Routine and acute video visits were scheduled to address both long-term issues and acute changes. Patients were seen periodically in the clinic by study staff to address acute issues or obtain laboratory or imaging data to follow up on issues identified during video visits. The clinical team reviewed each hospitalization and emergency department visit for each patient. Patients were contacted immediately after an identified hospitalization to re-engage them with the health care system and perform medication reconciliation. The frequency of monitoring was intensified following hospitalizations until patients returned to their baseline states. All intervention patients were reviewed every 30 days to ensure interaction with their devices and progress toward their individual goals. Study staff contacted patients who were not adequately using the system to encourage participation. Every 3 months, members of the study team, blinded to study assignment, contacted all participants. During these quarterly assessments, participants were queried for outcome events and possible adverse events. For events reported to have occurred at non-Veterans Affairs (VA) facilities, we requested release of information permission from the participant and discharge information from the facility. At the end of the study, all patients received a final visit during which outcomes were assessed. In addition to the reported events, the VA electronic medical record was queried to identify events that occurred within the VA system or for which the VA paid a claim from a third party.
End Points The primary end point was a composite of death, hospitalization, emergency department visits, and admission to a skilled nursing facility. Secondary end points included each component of Am J Kidney Dis. 2016;68(1):41-49
Telehealth for Patients With CKD the primary end point and incidence of end-stage kidney disease. Only self-reported events identified by the blinded study assessor at each 3-month telephone call and results from the VA electronic medical record query were used to define outcomes in the intervention and usual-care groups.
Statistical Analysis Our study sample size, n 5 600, was selected to provide statistical power of 80% to detect a 27% relative reduction in the occurrence of the primary end point in the intervention group compared to the usual-care group, using the 3:1 ratio of assignment to intervention, assuming a primary end point rate of 50% in the usual-care group, 10% loss to follow-up during the course of the study, and a 2-sided 0.05 significance level. One participant withdrew consent prior to device installation, so an additional participant was recruited to the study; hence, the final sample size of 601. Characteristics of patients by treatment group were compared using Pearson c2 (or Fisher exact test) and Wilcoxon rank sum tests for categorical and continuous variables, respectively. Cox proportional hazards models and corresponding maximum partial likelihood–based Wald tests were used to examine the association between intervention and time to first occurrence of the primary outcome. Cox proportional hazards models were also used to examine an association between intervention and time to first occurrence of each component of the primary outcome and time to the second occurrence of hospitalizations or emergency department visits. Cox proportional hazards models were repeated to include as additional explanatory measures baseline characteristics with significant differences between study arms and examine variation in the effect of the intervention by urban/rural residence
6805 Patients charts screened for inclusion
4105 Eligible by chart review
2449 Contacted • 373 No response • 1416 Declined • 59 Ineligible after patient contact
of participants. All analyses were repeated confining analyses to outcomes ascertained from the electronic medical record. All analyses used an intention-to-treat approach. An independent data and safety monitoring board monitored safety and efficacy of the intervention, progress of the study with respect to recruitment and follow-up, adherence to the study protocol, and quality of the study data. The human subjects’ safety committee at the Minneapolis VAHCS approved the study. Data analysis for this report was generated using SAS/STAT software, version 9.2 (SAS Institute Inc).
RESULTS From March 2012 through October 2012, a total of 6,805 patients were selected from the CKD registry based on their most recent eGFR measurements. Of these, 2,700 were deemed ineligible after chart review (Fig 1). A total of 601 patients were randomly assigned; 451 were assigned to the intervention, and 150, to usual care. A total of 547 (91.0%) patients had complete follow-up at 12 months, with no differences between groups in rates of follow-up completion. An additional 23 (3.8%) patients had partial follow-up at 12 months; primary and secondary outcomes were ascertained for all patients. Baseline characteristics were similar between study arms except for race and systolic blood pressure (Table 1). Mean eGFR was 37 mL/min/1.73 m2. 2700 Were ineligible by chart review • 319 Age > 85 years • 290 Deceased • 285 On Dialysis • 340 Significant Mental Health Issue • 154 No Address • 225 Nursing Home/Hospice • 145 Single eGFR < 60 mL/min/1.73 m2 • 838 Primary VA not Minneapolis • 104 Other reason 1656 Not contacted after recruitment target met
601 Patients underwent randomization Allocation 451 Assigned to intervention
150 Assigned to usual care
Follow-Up at 1 Year 1 Withdrew consent Analysis 450 Included in primary analysis
150 Included in primary analysis
Figure 1. Flow diagram shows screening, randomization, and follow-up of study participants. Abbreviations: eGFR, estimated glomerular filtration rate; VA, Veterans Affairs. Am J Kidney Dis. 2016;68(1):41-49
43
Ishani et al Table 1. Baseline Characteristics of Study Population Total (N 5 601)
Intervention (n 5 451)
Usual Care (n 5 150)
Age, y Male sex White racea,b SBP, mm Hgb DBP, mm Hg SBP . 140 mm Hg eGFR, mL/min/1.73 m2 eGFR category ,30 mL/min/1.73 m2 30-44 mL/min/1.73 m2 $45-60 mL/min/1.73 m2
75.1 6 8.1 592 (98.5) 585 (97.3) 133.1 6 19.7 70.9 6 11.9 193 (32.1) 37 6 9
75.3 6 8.1 445 (98.7) 445 (98.7) 134.3 6 19.9b 71.0 6 12.0 151 (33.5) 37 6 9
74.3 6 8.1 147 (98.0) 140 (93.3) 129.6 6 18.8 70.3 6 11.4 42 (28.0) 38 6 8
108 (18.0) 363 (60.4) 130 (21.6)
81 (18.0) 274 (60.8) 96 (21.3)
27 (18.0) 89 (59.3) 34 (22.7)
Hospitalization in year prior ACR, mg/g LDL-C, mg/dL ACEi/ARB use Education #HS or did not graduate Some college $4-y degree
172 (28.6) 296 6 748 73 6 30 335 (55.7)
127 (28.2) 321 6 815 74 6 31 256 (56.8)
45 (30.0) 221 6 484 70 6 29 79 (52.7)
238 (39.6) 214 (35.6) 149 (24.8)
179 (39.7) 157 (34.8) 115 (25.5)
59 (39.3) 57 (38.0) 34 (22.7)
Health status good to excellenta Diabetes mellitus Diabetes medication use Coronary artery disease Congestive heart failure History of stroke Daily aspirin use Previously seen nephrologist Informed of CKD Rural residence
395 (65.7) 256 (42.6) 223 (37.1) 207 (34.4) 125 (20.8) 99 (16.5) 456 (75.9) 320 (53.2) 378 (62.9) 334 (55.6)
288 (63.9) 191 (42.4) 168 (37.3) 158 (35.0) 95 (21.1) 78 (17.3) 343 (76.1) 239 (53.0) 285 (63.2) 244 (54.1)
107 (71.3) 65 (43.3) 55 (36.7) 49 (32.7) 30 (20.0) 21 (14.0) 113 (75.3) 81 (54.0) 93 (62.0) 90 (60.0)
Characteristic
Note: Values for categorical variables are given as number (percentage); for continuous variables, as mean 6 standard deviation. Conversion factor for LDL-C in mg/dL to mmol/L, 30.02586. Abbreviations: ACEi, angiotensin-converting enzyme inhibitor; ACR, albumin-creatinine ratio; ARB, angiotensin receptor blocker; CKD, chronic kidney disease; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; HS, high school; LDL-C, lowdensity lipoprotein cholesterol; SBP, systolic blood pressure. a Self-reported by patients. b All comparisons between groups were not significant except for white race and SBP, for which P , 0.05.
Prevalences of diabetes, coronary artery disease, and history of stroke were 42.6%, 34.4%, and 16.5%, respectively; 55.6% of patients were considered rural. Patients lived on average 98 miles from the Minneapolis VAHCS; 147 (24.5%) patients lived more than 200 miles away. Participants in the intervention group were engaged in the intervention; 96.2% completed at least 1 video visit. Mean numbers of measurements per participant per month were 14.9 6 10.9 (standard deviation) for blood pressure and 11.2 6 9.6 for weight, and on average, each intervention participant interacted with an educational module 5.8 6 7.9 days per month (Fig 2). A total of 850 vital signs triggered a patientspecific flag requiring an ad hoc patient contact (median, 1 per patient; maximum, 19 per patient). Mean numbers of visits per participant during the intervention period were 3.8 scheduled nurse, 1.9 ad hoc nurse, 4.7 provider, 1.2 clinical pharmacy 44
specialist, 1.4 dietician, 0.9 psychologist, and 0.3 social worker visits. At 1 year, the primary outcome of death, hospitalization, emergency department visits, or admission to a skilled nursing facility occurred for 70 (46.7%) patients in the usual-care group and 208 (46.2%) in the intervention group (P 5 0.9; Table 2). The unadjusted hazard ratio (HR) for the primary end point in the intervention group compared to the usual-care group was 0.98 (95% confidence interval [CI], 0.751.29; Table 2; Fig 3). When analyses were restricted to events only ascertained through the electronic medical record, none of the results were materially different. No significant differences between the 2 groups were seen for any of the secondary end points (Table 2). Median hospitalization durations for the first visit were also similar between the intervention and usual-care groups (3 days for the intervention Am J Kidney Dis. 2016;68(1):41-49
Telehealth for Patients With CKD
factors (HR, 1.02; 95% CI, 0.77-1.34). Similarly, adjustment for baseline imbalance did not alter our results for any of the secondary measures. Outcomes did not differ between the intervention and usualcare groups when stratified by key prespecified subgroups, including age, eGFR category, diabetes, hospitalization in the prior year, or congestive heart failure (P for interaction . 0.1 for each comparison; Table S1, available as online supplementary material). There was also no difference in the primary outcome by intensity of engagement with the system (above or below median for number of transmissions of blood pressure, weight, glucose level, etc; P . 0.1 for each comparison). In examining whether the effect of the intervention varied by urban/rural residence of participants, the interaction between rurality and treatment arm was not significant. However, among rural patients, the intervention group tended to have better primary outcomes (HR, 0.85; 95% CI, 0.58-1.22) compared to usual care, which was different directionally compared with results observed for urban patients (HR, 1.13; 95% CI, 0.76-1.69; Table S2). Results for secondary outcomes between rural and urban patients were mixed (Table S2). Rural participants in the intervention arm tended to have more transmissions of blood pressure, glucose, temperature, and weight measurements with the intervention device and more education module use with a statistically significantly greater number of weight transmissions compared to the usual-care group.
DISCUSSION
Figure 2. Distribution of average number of blood pressure readings, weight measurements, and days on which participants viewed an educational module per month. Values displayed in the box-and-whisker plots are the 5 number summary (minimum, first quartile, median, third quartile, and maximum) and mean 6 standard deviation.
group vs 2 for the usual-care group; P 5 0.8). Proportions with hospital readmissions were similar between groups (intervention vs usual care, 34.3% vs 40.0%; difference, 5.7%; 95% CI, 213.1% to 24.5%). There was no difference between groups in time to dialysis therapy initiation (Table 2). There was also no difference between the intervention and usual-care groups in control of intermediate variables such as blood pressure, diabetes, hyperlipidemia, or smoking (Table 3). At baseline, the intervention and usual-care groups were slightly imbalanced regarding race and systolic blood pressure. There was no difference in time to first event after adjusting for these baseline Am J Kidney Dis. 2016;68(1):41-49
In our randomized controlled trial of patients with moderate to severe CKD, we found that delivery of health care by an interprofessional team using telehealth could be effectively implemented for both rural and urban patients, but did not reduce the risk for death, hospitalization, emergency department visits, or admission to skilled nursing facilities compared to usual care. Similarly, there was no reduction in number of hospitalizations or median duration of hospitalizations. These results were seen despite significant engagement by patients in using the device and numerous virtual visits by the interprofessional team. Although our study was focused on the management of patients with CKD, we viewed CKD as a marker for significant chronic disease burden. A total of 42.6% of patients had diabetes; 34.4%, coronary artery disease; 20.8%, congestive heart failure; and 16.5%, prior stroke. Telehealth in its various forms has been used for the management of chronic diseases including congestive heart failure, chronic obstructive pulmonary disease, coronary artery disease, diabetes, 45
Ishani et al Table 2. Clinical End Points by Treatment Arm With Summary of Cox Proportional Hazards Analysis of Time-to-Event Variables End Point
Intervention (n 5 450)
Usual Care (n 5 150)
208 (46.2)
70 (46.7)
0.98 (0.75-1.29)
13 134 164 18 11
3 40 58 2 2
1.46 1.15 0.92 3.07 1.86
Primary Secondary Death Hospitalization Emergency department visits Admission to skilled nursing facility Initiation of dialysis Othera Second hospitalizationb Second emergency department visitb No. of days of first hospital visitc No. of days of all hospitalizationsc No. of hospitalizations within yeara 0 1 $2
(2.9) (29.8) (36.4) (4.0) (2.4)
(2.0) (26.7) (38.7) (1.3) (1.3)
46 (34.3) 64 (39.0) 3.4 6 3.5 [3] 5.7 6 8.6 [3]
16 (40.0) 26 (44.8) 3.6 6 3.9 [2] 6.2 6 6.5 [3]
316 (70.2) 88 (19.6) 46 (10.2)
110 (73.3) 24 (16.0) 16 (10.7)
HR (95% CI)
(0.42-5.11) (0.80-1.63) (0.68-1.24) (0.71-13.24) (0.41-8.39)
0.89 (0.50-1.57) 0.87 (0.55-1.37)
Note: Values for categorical variables are given as number (percentage); for continuous variables, as mean 6 standard deviation [median]. No significant differences were seen between groups in any of the measured outcomes. Abbreviations: CI, confidence interval; HR, hazard ratio. a Pearson c2 test for the categorized number of hospital admissions within the first year after randomization. b Percentage of those with a first visit. c Wilcoxon rank sum tests for duration of hospitalization.
Probability of Freedom from First Occurrence of Primary Outcome
obesity, depression and other mental health problems, and chronic pain, with mixed results.2-8 Part of the reason for the lack of consensus on the value of
1.0 Usual Care Intervention 0.8
0.6
0.4
0.2
Log−rank p=0.902 0.0 0
1
2
3
4
5
6
7
8
9
10
11
12
Months Since Enrollment No. at risk Usual 150 143 133 125 121 112 106 102 97 89 85 81 Care Intervention 450 432 409 385 357 341 321 302 289 274 261 255
Figure 3. Kaplan-Meier time-to-event occurrences for the primary end point of a composite of death, hospitalization, emergency department visits, and admission to a skilled nursing facility. The hazard ratio for first occurrence with intervention was 0.98 (95% confidence interval, 0.75-1.29; P 5 0.90). 46
telehealth relates to different levels of intervention intensity16,17 and the clinical relevance of the studied outcomes, including both surrogate outcomes and hard clinical end points.18,19 Our results are similar to other studies that have evaluated case management in patients with CKD, although our study involved a greater component of interprofessional care than both these studies and also included home video monitoring and education.20-23 However, despite the intensity and broad scope of our intervention, we were unable to detect an improvement in clinical outcomes between groups. Patients within rural areas have a number of barriers to optimal health care, including higher prevalences of obesity, hypertension, diabetes, and tobacco use.24 Rural areas also have a scarcity of nephrologists25 and these patients are less likely to visit a nephrologist.25 Previous studies have suggested that patients with CKD stages 3 to 4 in rural areas have poorer process-based markers of care compared with urban patients, such as measurement of either hemoglobin A1c in patients with diabetes, assessment of urinary albumin excretion, or use of angiotensinconverting enzyme inhibitors or angiotensin receptor blockers among patients with diabetes or significant proteinuria.26 Rural patients with CKD also have a greater risk for hospitalization and all-cause mortality compared with urban patients.26 This pattern of poorer care appears to have been reversed in our study, with rural intervention patients having a trend toward a lower incidence of the primary outcome. The lack of statistical significance in this group is Am J Kidney Dis. 2016;68(1):41-49
Telehealth for Patients With CKD Table 3. Intermediate Outcomes at Study End by Treatment Group Outcome
Total
Intervention
Usual Care
Pa
No. of participants with complete intermediate outcome data No. with SBP # 140 mm Hg, LDL-C # 100 mg/dL, and HbA1c # 8% at study end
444
340 185 (54.4)
104 53 (51.0)
0.5
No. of participants with baseline SBP . 140 mm Hg and complete data No. with SBP # 140 mm Hg at study end
174
135 72 (53.3)
39 20 (51)
0.8
No. of participants with baseline LDL-C . 100 mg/dL and complete data No. with LDL-C # 100 mg/dL at study end
76
61 31 (51)
15 8 (53)
0.9
No. of participants with baseline HbA1c . 8% and complete data No. with HbA1c # 8% at study end
48
33 14 (42)
15 3 (33)
0.6
No. of participants who were smokers at baseline No. no longer smoking at study end
52
40 9 (23)
12 5 (42)
0.3
Note: Values are given as number (percentage). Conversion factor for LDL-C in mg/dL to mmol/L, 30.02586. Abbreviations: HbA1c, hemoglobin A1c; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure. a Pearson c2 test, except for Fisher exact test for smoking status due to small expected cell counts.
likely because the study was not powered to evaluate outcomes within subgroups. Although the overall study was negative, rural patients may be a selected subgroup in which the use of telehealth and interprofessional care may offer benefits, particularly in areas that are scarce in subspecialty care. This subgroup needs to be specifically tested in future studies. There are a number of possible reasons that our study was unable to demonstrate an overall benefit in outcomes. First, aggressive home monitoring and interprofessional care may simply be ineffective in altering patient outcomes. Alternatively, the team may have identified a greater number of issues that may be ignored by patients without constant monitoring.27 Another possible explanation of negative results is that patients in the intervention group did not engage in using the system. Although a small number of patients (3.8%) did not engage in the system, most were fully engaged after 1 year of follow-up. Adherence rates in our study were likely significantly better than can be achieved in usual clinical care given the resources available during the trial. Another possibility is that the duration of follow-up was too short. It may be that differences between groups would be observed with longer follow-up. Finally, it could be that individuals in the usual-care group received very aggressive therapy compared with those receiving traditional usual care. We did not track interventions provided to the usual-care group. However, during the course of our study, the VA was changing the model of primary care to that of a patient-centered medical home, called PACT (Patient Aligned Care Team), along with implementation of a structured telehealth program. Implementation of the PACT and home telehealth have independently been associated with reductions in emergency department visits, hospitalizations,21-23,28-30 and improved control of Am J Kidney Dis. 2016;68(1):41-49
hypertension.24,31 As an example, 51.3% of patients in the usual-care group had incident control of hypertension, an incident control rate that is substantially greater than community standards. Aggressive implementation of therapy in the usual-care group (PACT, telehealth, and case management) may have been why we were unable to demonstrate a difference in either intermediate or clinical outcomes between groups. Another consideration is that disease progression does not always follow a pattern of progressive decline, but instead, patients may suddenly deteriorate. Therefore, early intervention to prevent clinical deterioration may not always be possible. Also, the specific monitored physiologic parameters may not be those that best predict disease progression. We powered the study assuming a 50% event rate in the usualcare group, with an observed event rate of 46.3%. Our loss to follow-up was significantly smaller than anticipated. Our study had sufficient power for effects of this magnitude, but not for smaller effects. Although our study results were negative, the ranges of the CIs for our effect estimates do not rule out a clinically significant benefit for a similar intervention. Our study had numerous strengths. We examined relevant and clinically important primary end points and enrolled a patient population at high risk for adverse outcomes. We attempted to target numerous facets of care through the use of a diverse interprofessional team. We sought to engage and activate patients in their care, provide in-home education, facilitate visits using video, manage chronic diseases, and identify and address acute issues. The intervention included medication management and comprehensive yet simple-to-understand educational modules. We included both urban and rural patients and randomly assigned patients in their homes, potentially enhancing the generalizability of our 47
Ishani et al
findings.10 Finally, all our outcome assessors were blinded to study assignment. A potential weakness of our study is that patients in the intervention group were likely queried more often than those in the usual-care group about hospitalizations and emergency department visits, likely enhancing their recall of these events during the outcome assessment calls and potentially biasing our results toward more adverse outcomes reported in the intervention group. However, no difference between groups was seen when we confined our analysis to only VA electronic medical record2ascertained events. We performed a complex intervention with many different aspects. It is unclear whether any individual component of our intervention could be associated with benefit that is overwhelmed by the overall study results. However, our goal was to deliver a comprehensive intervention to determine whether a fully integrated model of care could alter patient outcomes. Given the overall negative study findings, it is unlikely that individual components of care would result in a different outcome. The generalizability of our results to other settings and other patient populations remains to be determined. In summary, a strategy of telehealth by an interprofessional team could be effectively implemented in both rural and urban patients but did not alter outcomes in patients with moderate to severe CKD compared with usual care. Strategies for improving the care of this population are needed.
ACKNOWLEDGEMENTS The Center for Innovative Kidney Care, which is based at Minneapolis VAHCS, comprises Melissa Atwood, Ann Bangerter, Theresa Borah, Olga Brusilovsky, Andrea Cutting, Mia Dobbs, Wendy Elofson, Kori Geinert, Kanika Gupta, Kendra Hackney, Ron Howell, Miles Jarzyna, Quinn Kellerman, Elaine Kroska, LauraJean Krueger, Narlina Lalani, Shannon McCutcheon, Alec Otteman, Megan Plumstead, Bethany Roberts, Keri Rowe, Jackie Rust, Rita Schmitt, Herb Stockley, Jaclyn Strutt, Melvis Tambi, and Callen Weispfennig. The members of Data Safety Monitoring Board for this trial were Kristine E. Ensrud, MD, MPH, Minneapolis VAHCS; Jeffrey Connaire, MD, Minneapolis VAHCS; and Brent Taylor, PhD, VAHCS. Nan Booth, MSW, MPH, ELS, provided manuscript editing. Portions of this work were presented at the 2014 American Society of Nephrology Kidney Week, November 11-16, Philadelphia, PA. Support: The project was funded by grants from the VA Center for Innovation to Dr Ishani and Dr Rosenberg. The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. Financial Disclosure: Dr Rosenberg has received honoraria from Wolter Kluwer Health for serving as an author for UpToDate and from the American Society of Nephrology for service as education director. The other authors declare that they have no relevant financial interests. Contributions: Study concept and design: AI, MER; data acquisition, analysis or interpretation: AI, JC, DP, SO, BC, SN, 48
DN, MER; statistical analysis: BC, DN; obtainment of funding: AI, MER; administrative, technical, or material support: AI, JC, DP, SO, BC, SN, DN, MER; study supervision: AI, MER. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved. AI takes responsibility that this study has been reported honestly, accurately, and transparently; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned and registered have been explained. Peer Review: Evaluated by 2 external peer reviewers, a Statistical Editor, a Co-Editor, and the Editor-in-Chief.
SUPPLEMENTARY MATERIAL Table S1: Clinical end points by strata at randomization with summary of proportional hazards analysis of time-to-event variables. Table S2: Clinical end points stratified by setting at randomization with summary of proportional hazards analysis of time-toevent variables. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2016.01.018) is available at www.ajkd.org
REFERENCES 1. Berwick DM, Nolan TW, Whittington J. The triple aim: care, health, and cost. Health Aff (Millwood). 2008;27(3):759-769. 2. Wennberg DE, Marr A, Lang L, O’Malley S, Bennett G. A randomized trial of a telephone care-management strategy. N Engl J Med. 2010;363(13):1245-1255. 3. Chaudhry SI, Mattera JA, Curtis JP, et al. Telemonitoring in patients with heart failure. N Engl J Med. 2010;363(24):2301-2309. 4. Kroenke K, Krebs EE, Wu J, Yu Z, Chumbler NR, Bair MJ. Telecare collaborative management of chronic pain in primary care: a randomized clinical trial. JAMA. 2014;312(3):240-248. 5. Goldstein RS, O’Hoski S. Telemedicine in COPD: time to pause. Chest. 2014;145(5):945-949. 6. Gellis ZD, Kenaley BL, Ten HT. Integrated telehealth care for chronic illness and depression in geriatric home care patients: the Integrated Telehealth Education and Activation of Mood (I-TEAM) study. J Am Geriatr Soc. 2014;62(5):889-895. 7. Wade VA, Karnon J, Elshaug AG, Hiller JE. A systematic review of economic analyses of telehealth services using real time video communication. BMC Health Serv Res. 2010;10(10):233. 8. McLean S, Sheikh A, Cresswell K, et al. The impact of telehealthcare on the quality and safety of care: a systematic overview. PLoS One. 2013;19(8):e71238. 9. Margolis KL, Asche SE, Bergdall AR, et al. Effect of home blood pressure telemonitoring and pharmacist management on blood pressure control: a cluster randomized clinical trial. JAMA. 2013;310(1):46-56. 10. Rifkin DE, Abdelmalek JA, Miracle CM, et al. Linking clinic and home: a randomized controlled clinical effectiveness trial of real-time, wireless blood pressure monitoring for older patients with kidney disease and hypertension. Blood Press Monit. 2013;18(1):8-15. 11. Gordon EJ, Fink JC, Fischer MJ. Telenephrology: a novel approach to improve coordinated and collaborative care for chronic kidney disease. Nephrol Dial Transplant. 2013;28(4):972-981. 12. US Census Bureau. 2010 Census Urban and Rural Classification and Urban Area Criteria. https://www.census.gov/geo/ reference/ua/urban-rural-2010.html. Accessed August 1, 2015. Am J Kidney Dis. 2016;68(1):41-49
Telehealth for Patients With CKD 13. Wagner EH, Austin BT, Von KM. Organizing care for patients with chronic illness. Milbank Q. 1996;74(4):511-544. 14. Hibbard JH, Stockard J, Mahoney ER, Tusler M. Development of the Patient Activation Measure (PAM): Conceptualizing and Measuring Activation in Patients and Consumers. Health Serv Res. 2004;39(4 Pt 1):1005-1026. 15. Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl. 2013;3:1-150. 16. Feltner C, Jones CD, Cene CW, et al. Transitional care interventions to prevent readmissions for persons with heart failure: a systematic review and meta-analysis. Ann Intern Med. 2014;160(11):774-784. 17. Inglis SC, Clark RA, McAlister FA, Stewart S, Cleland JG. Which components of heart failure programmes are effective? A systematic review and meta-analysis of the outcomes of structured telephone support or telemonitoring as the primary component of chronic heart failure management in 8323 patients: abridged Cochrane review. Eur J Heart Fail. 2011;13(9):1028-1040. 18. Omboni S, Gazzola T, Carabelli G, Parati G. Clinical usefulness and cost effectiveness of home blood pressure telemonitoring: meta-analysis of randomized controlled studies. J Hypertens. 2013;31(3):455-467. 19. Worswick J, Wayne SC, Bennett R, et al. Improving quality of care for persons with diabetes: an overview of systematic reviews - what does the evidence tell us? Syst Rev. 2013;7(2):1-14. 20. Barrett BJ, Garg AX, Goeree R, et al. A nurse-coordinated model of care versus usual care for stage 3/4 chronic kidney disease in the community: a randomized controlled trial. Clin J Am Soc Nephrol. 2011;6(6):1241-1247. 21. Hopkins RB, Garg AX, Levin A, et al. Cost-effectiveness analysis of a randomized trial comparing care models for chronic kidney disease. Clin J Am Soc Nephrol. 2011;6(6): 1248-1257. 22. Van Zuilen AD, Bots ML, Dulger A, et al. Multifactorial intervention with nurse practitioners does not change cardiovascular
Am J Kidney Dis. 2016;68(1):41-49
outcomes in patients with chronic kidney disease. Kidney Int. 2012;82(11):710-717. 23. Peeters MJ, van Zuilen AD, van den Brand JA, et al. Nurse practitioner care improves renal outcome in patients with CKD. J Am Soc Nephrol. 2014;25(2):390-398. 24. Agency for Healthcare Research and Quality. 2010 National Healthcare Quality Report. January 2013. Rockville, MD. http://www.ahrq.gov/research/findings/nhqrdr/nhqr10/index.html. Accessed August 1, 2015. 25. Yan G, Cheung AK, Ma JZ, et al. The associations between race and geographic area and quality-of-care indicators in patients approaching ESRD. Clin J Am Soc Nephrol. 2013;8(4):610-618. 26. Rucker D, Hemmelgarn BR, Lin M, et al. Quality of care and mortality are worse in chronic kidney disease patients living in remote areas. Kidney Int. 2011;79(2):210-217. 27. Weinberger M, Oddone EZ, Henderson WG. Does increased access to primary care reduce hospital readmissions? Veterans Affairs Cooperative Study Group on Primary Care and Hospital Readmission. N Engl J Med. 1996;334(22):1441-1447. 28. Hebert PL, Liu CF, Wong ES, et al. Patient-centered medical home initiative produced modest economic results for Veterans Health Administration, 2010-12. Health Aff (Millwood). 2014;33(6):980-987. 29. Nelson KM, Helfrich C, Sun H, et al. Implementation of the patient-centered medical home in the Veterans Health Administration: associations with patient satisfaction, quality of care, staff burnout, and hospital and emergency department use. JAMA Intern Med. 2014;174(8):1350-1358. 30. Darkins A, Kendall S, Edmonson E, Young M, Stressel P. Reduced cost and mortality using home telehealth to promote selfmanagement of complex chronic conditions: a retrospective matched cohort study of 4,999 veteran patients. Telemed J E Health. 2015;21(1):70-76. 31. O’Neill JL, Cunningham TL, Wiitala WL, Bartley EP. Collaborative hypertension case management by registered nurses and clinical pharmacy specialists within the Patient Aligned Care Teams (PACT) model. J Gen Intern Med. 2014;29(suppl 2):S675S681.
49
Editorial, p. 5
S
trategies are needed to improve the care of patients with chronic disease. This care should reduce adverse health outcomes, provide a timely and convenient care experience for patients regardless of their
From the 1Section of Renal Diseases and Hypertension, Minneapolis Veterans Affairs Health Care System; 2Division of Renal Diseases and Hypertension, University of Minnesota Medical School; and 3Center for Chronic Disease Outcomes Research, Minneapolis Veterans Affairs Health Care System, Minneapolis, MN. * A list of Center for Innovative Kidney Care members appears in the Acknowledgements. Received September 29, 2015. Accepted in revised form January 14, 2016. Originally published online March 1, 2016. Trial registration: www.ClinicalTrials.gov; study number: NCT01446029. Am J Kidney Dis. 2016;68(1):41-49
location, and be of high value with the potential to reduce overall health system costs.1 Telehealth is an example of such a strategy. It has been used, with or without case management, in various forms to manage patients with chronic illnesses such as chronic obstructive pulmonary disease, heart failure, and diabetes mellitus.2-6 Despite its growing use, telehealth outcomes have been variable Address correspondence to Areef Ishani, MD, MS, Department of Medicine, Nephrology Section, Minneapolis VA Medical Center, One Veterans Dr (111J), Minneapolis, MN 55417. E-mail: [email protected] Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. This is a US Government Work. There are no restrictions on its use. 0272-6386 http://dx.doi.org/10.1053/j.ajkd.2016.01.018
41
Ishani et al
and its expense is often considerable, emphasizing the need to carefully assess its effectiveness.7,8 Previous studies have suggested that interprofessional case management or home monitoring could improve intermediate outcomes such as blood pressure in patients with chronic kidney disease (CKD).9-11 The feasibility and effectiveness of such interventions on patientcentered outcomes is unclear. We conducted a randomized controlled trial to determine whether an intervention consisting of telehealth with interprofessional team case management could be effectively implemented and whether it could improve the combined end point of death, hospitalization, emergency department visits, or admission to a nursing home in patients with moderate to severe CKD compared to usual care.
METHODS Patients We randomly assigned 451 participants to receive the intervention and 150 participants to receive usual care from the Minneapolis Veterans Affairs Health Care System (VAHCS), the St. Cloud, MN, VAHCS, and their affiliated community-based clinics. A registry of patients with CKD (estimated glomerular filtration rate [eGFR] , 60 mL/min/1.73 m2 based on the most recent creatinine value within past year) was created. From March 2012 through October 2012, patients older than 18 years whose most recent eGFR (calculated using the IDMS-traceable 4-variable MDRD [Modification of Diet in Renal Disease] Study equation) within the last year was ,60 mL/ min/1.73 m2 were identified from the registry and invited to participate in the study. Patients from rural areas were oversampled.12 Patients who were unable to give consent, had life expectancy less than 1 year, lived in a skilled nursing facility, or had a primary care provider unwilling to allow participation were excluded. Potential participants were invited to be in the study by mail and then by telephone. Interested individuals were scheduled for an inhome visit, during which the written informed consent process was completed, baseline measurements were collected, and participants were randomly assigned to receive the intervention or usual care. Participants randomly assigned to the intervention received a second in-home visit during which the video monitoring device with peripherals and broadband (most commonly cellular connection) were installed. Participants in the usual-care group were invited to attend a CKD education class and asked to follow up with their primary care providers regarding management of their kidney disease. Participants assigned to usual care may or may not have been seen by a nephrologist during the course of the trial, at the discretion of their primary care provider. The exact care received by the usual-care group was not investigated.
Randomization and Masking Eligible patients were randomly assigned to receive the intervention or usual care using a centralized computer-generated randomization scheme using permuted block sizes of 2, 4, or 6. Randomization was stratified by eGFR (,30 vs $30 mL/min/ 1.73 m2), presence of diabetes, and occurrence of a hospitalization in the past year. Randomization occurred over the telephone by an individual blinded to patient identity. Although participants could not be masked to their assignment, outcome assessors were blinded to participant study assignments.
Procedures Participants in the intervention group received in-home training regarding how to use the device (LifeView; AmericanTeleCare) 42
and all the peripherals (blood pressure cuff, scale, glucometer, pulse oximeter, stethoscope, and web camera) and how to contact the clinical team. An interprofessional team consisting of a nephrologist, nurse practitioner, nurses, clinical pharmacy specialist, psychologist, social worker, telehealth care technician, and dietician reviewed the health status of each intervention participant and developed a patient-specific treatment plan addressing short- and long-term goals. The goal of the intervention program was comprehensive care of CKD and comorbid conditions using components of the chronic care model.13 Specific issues addressed included the following: management of blood pressure, volume status, proteinuria, diabetes mellitus, lipid levels, and depression; health literacy; and patient activation.14 The program also addressed lifestyle modification (physical activity, diet, weight reduction, and smoking cessation). A customized education program was developed based on each patient’s comorbid conditions and was delivered over broadband to the device. Patients could interact with the educational modules at their own learning pace. Patients were also given a customized self-monitoring strategy based on their clinical condition. Vital signs were automatically measured by the device and transmitted to the study team. The clinical team met daily to discuss high-risk patients. The full interprofessional team met weekly to review intervention patients and their progress toward goals, hospitalizations, and risks for recurrent or adverse events. Nurses also reviewed all vital signs and module responses transmitted to the study team twice a day. Vital signs were categorized as within or beyond range based on individualized patient parameters. Nurses addressed vital signs that were beyond the prespecified range with the assistance of a provider when clinically indicated. This typically involved interactive video conferencing (a virtual visit) with the patient to review the clinical situation, and possibly a medication change. Medication management was conducted by nurses, a clinical pharmacy specialist, and providers for all intervention participants based on guidelines15 and self-measured values of blood pressure, blood glucose, and volume status. Routine and acute video visits were scheduled to address both long-term issues and acute changes. Patients were seen periodically in the clinic by study staff to address acute issues or obtain laboratory or imaging data to follow up on issues identified during video visits. The clinical team reviewed each hospitalization and emergency department visit for each patient. Patients were contacted immediately after an identified hospitalization to re-engage them with the health care system and perform medication reconciliation. The frequency of monitoring was intensified following hospitalizations until patients returned to their baseline states. All intervention patients were reviewed every 30 days to ensure interaction with their devices and progress toward their individual goals. Study staff contacted patients who were not adequately using the system to encourage participation. Every 3 months, members of the study team, blinded to study assignment, contacted all participants. During these quarterly assessments, participants were queried for outcome events and possible adverse events. For events reported to have occurred at non-Veterans Affairs (VA) facilities, we requested release of information permission from the participant and discharge information from the facility. At the end of the study, all patients received a final visit during which outcomes were assessed. In addition to the reported events, the VA electronic medical record was queried to identify events that occurred within the VA system or for which the VA paid a claim from a third party.
End Points The primary end point was a composite of death, hospitalization, emergency department visits, and admission to a skilled nursing facility. Secondary end points included each component of Am J Kidney Dis. 2016;68(1):41-49
Telehealth for Patients With CKD the primary end point and incidence of end-stage kidney disease. Only self-reported events identified by the blinded study assessor at each 3-month telephone call and results from the VA electronic medical record query were used to define outcomes in the intervention and usual-care groups.
Statistical Analysis Our study sample size, n 5 600, was selected to provide statistical power of 80% to detect a 27% relative reduction in the occurrence of the primary end point in the intervention group compared to the usual-care group, using the 3:1 ratio of assignment to intervention, assuming a primary end point rate of 50% in the usual-care group, 10% loss to follow-up during the course of the study, and a 2-sided 0.05 significance level. One participant withdrew consent prior to device installation, so an additional participant was recruited to the study; hence, the final sample size of 601. Characteristics of patients by treatment group were compared using Pearson c2 (or Fisher exact test) and Wilcoxon rank sum tests for categorical and continuous variables, respectively. Cox proportional hazards models and corresponding maximum partial likelihood–based Wald tests were used to examine the association between intervention and time to first occurrence of the primary outcome. Cox proportional hazards models were also used to examine an association between intervention and time to first occurrence of each component of the primary outcome and time to the second occurrence of hospitalizations or emergency department visits. Cox proportional hazards models were repeated to include as additional explanatory measures baseline characteristics with significant differences between study arms and examine variation in the effect of the intervention by urban/rural residence
6805 Patients charts screened for inclusion
4105 Eligible by chart review
2449 Contacted • 373 No response • 1416 Declined • 59 Ineligible after patient contact
of participants. All analyses were repeated confining analyses to outcomes ascertained from the electronic medical record. All analyses used an intention-to-treat approach. An independent data and safety monitoring board monitored safety and efficacy of the intervention, progress of the study with respect to recruitment and follow-up, adherence to the study protocol, and quality of the study data. The human subjects’ safety committee at the Minneapolis VAHCS approved the study. Data analysis for this report was generated using SAS/STAT software, version 9.2 (SAS Institute Inc).
RESULTS From March 2012 through October 2012, a total of 6,805 patients were selected from the CKD registry based on their most recent eGFR measurements. Of these, 2,700 were deemed ineligible after chart review (Fig 1). A total of 601 patients were randomly assigned; 451 were assigned to the intervention, and 150, to usual care. A total of 547 (91.0%) patients had complete follow-up at 12 months, with no differences between groups in rates of follow-up completion. An additional 23 (3.8%) patients had partial follow-up at 12 months; primary and secondary outcomes were ascertained for all patients. Baseline characteristics were similar between study arms except for race and systolic blood pressure (Table 1). Mean eGFR was 37 mL/min/1.73 m2. 2700 Were ineligible by chart review • 319 Age > 85 years • 290 Deceased • 285 On Dialysis • 340 Significant Mental Health Issue • 154 No Address • 225 Nursing Home/Hospice • 145 Single eGFR < 60 mL/min/1.73 m2 • 838 Primary VA not Minneapolis • 104 Other reason 1656 Not contacted after recruitment target met
601 Patients underwent randomization Allocation 451 Assigned to intervention
150 Assigned to usual care
Follow-Up at 1 Year 1 Withdrew consent Analysis 450 Included in primary analysis
150 Included in primary analysis
Figure 1. Flow diagram shows screening, randomization, and follow-up of study participants. Abbreviations: eGFR, estimated glomerular filtration rate; VA, Veterans Affairs. Am J Kidney Dis. 2016;68(1):41-49
43
Ishani et al Table 1. Baseline Characteristics of Study Population Total (N 5 601)
Intervention (n 5 451)
Usual Care (n 5 150)
Age, y Male sex White racea,b SBP, mm Hgb DBP, mm Hg SBP . 140 mm Hg eGFR, mL/min/1.73 m2 eGFR category ,30 mL/min/1.73 m2 30-44 mL/min/1.73 m2 $45-60 mL/min/1.73 m2
75.1 6 8.1 592 (98.5) 585 (97.3) 133.1 6 19.7 70.9 6 11.9 193 (32.1) 37 6 9
75.3 6 8.1 445 (98.7) 445 (98.7) 134.3 6 19.9b 71.0 6 12.0 151 (33.5) 37 6 9
74.3 6 8.1 147 (98.0) 140 (93.3) 129.6 6 18.8 70.3 6 11.4 42 (28.0) 38 6 8
108 (18.0) 363 (60.4) 130 (21.6)
81 (18.0) 274 (60.8) 96 (21.3)
27 (18.0) 89 (59.3) 34 (22.7)
Hospitalization in year prior ACR, mg/g LDL-C, mg/dL ACEi/ARB use Education #HS or did not graduate Some college $4-y degree
172 (28.6) 296 6 748 73 6 30 335 (55.7)
127 (28.2) 321 6 815 74 6 31 256 (56.8)
45 (30.0) 221 6 484 70 6 29 79 (52.7)
238 (39.6) 214 (35.6) 149 (24.8)
179 (39.7) 157 (34.8) 115 (25.5)
59 (39.3) 57 (38.0) 34 (22.7)
Health status good to excellenta Diabetes mellitus Diabetes medication use Coronary artery disease Congestive heart failure History of stroke Daily aspirin use Previously seen nephrologist Informed of CKD Rural residence
395 (65.7) 256 (42.6) 223 (37.1) 207 (34.4) 125 (20.8) 99 (16.5) 456 (75.9) 320 (53.2) 378 (62.9) 334 (55.6)
288 (63.9) 191 (42.4) 168 (37.3) 158 (35.0) 95 (21.1) 78 (17.3) 343 (76.1) 239 (53.0) 285 (63.2) 244 (54.1)
107 (71.3) 65 (43.3) 55 (36.7) 49 (32.7) 30 (20.0) 21 (14.0) 113 (75.3) 81 (54.0) 93 (62.0) 90 (60.0)
Characteristic
Note: Values for categorical variables are given as number (percentage); for continuous variables, as mean 6 standard deviation. Conversion factor for LDL-C in mg/dL to mmol/L, 30.02586. Abbreviations: ACEi, angiotensin-converting enzyme inhibitor; ACR, albumin-creatinine ratio; ARB, angiotensin receptor blocker; CKD, chronic kidney disease; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; HS, high school; LDL-C, lowdensity lipoprotein cholesterol; SBP, systolic blood pressure. a Self-reported by patients. b All comparisons between groups were not significant except for white race and SBP, for which P , 0.05.
Prevalences of diabetes, coronary artery disease, and history of stroke were 42.6%, 34.4%, and 16.5%, respectively; 55.6% of patients were considered rural. Patients lived on average 98 miles from the Minneapolis VAHCS; 147 (24.5%) patients lived more than 200 miles away. Participants in the intervention group were engaged in the intervention; 96.2% completed at least 1 video visit. Mean numbers of measurements per participant per month were 14.9 6 10.9 (standard deviation) for blood pressure and 11.2 6 9.6 for weight, and on average, each intervention participant interacted with an educational module 5.8 6 7.9 days per month (Fig 2). A total of 850 vital signs triggered a patientspecific flag requiring an ad hoc patient contact (median, 1 per patient; maximum, 19 per patient). Mean numbers of visits per participant during the intervention period were 3.8 scheduled nurse, 1.9 ad hoc nurse, 4.7 provider, 1.2 clinical pharmacy 44
specialist, 1.4 dietician, 0.9 psychologist, and 0.3 social worker visits. At 1 year, the primary outcome of death, hospitalization, emergency department visits, or admission to a skilled nursing facility occurred for 70 (46.7%) patients in the usual-care group and 208 (46.2%) in the intervention group (P 5 0.9; Table 2). The unadjusted hazard ratio (HR) for the primary end point in the intervention group compared to the usual-care group was 0.98 (95% confidence interval [CI], 0.751.29; Table 2; Fig 3). When analyses were restricted to events only ascertained through the electronic medical record, none of the results were materially different. No significant differences between the 2 groups were seen for any of the secondary end points (Table 2). Median hospitalization durations for the first visit were also similar between the intervention and usual-care groups (3 days for the intervention Am J Kidney Dis. 2016;68(1):41-49
Telehealth for Patients With CKD
factors (HR, 1.02; 95% CI, 0.77-1.34). Similarly, adjustment for baseline imbalance did not alter our results for any of the secondary measures. Outcomes did not differ between the intervention and usualcare groups when stratified by key prespecified subgroups, including age, eGFR category, diabetes, hospitalization in the prior year, or congestive heart failure (P for interaction . 0.1 for each comparison; Table S1, available as online supplementary material). There was also no difference in the primary outcome by intensity of engagement with the system (above or below median for number of transmissions of blood pressure, weight, glucose level, etc; P . 0.1 for each comparison). In examining whether the effect of the intervention varied by urban/rural residence of participants, the interaction between rurality and treatment arm was not significant. However, among rural patients, the intervention group tended to have better primary outcomes (HR, 0.85; 95% CI, 0.58-1.22) compared to usual care, which was different directionally compared with results observed for urban patients (HR, 1.13; 95% CI, 0.76-1.69; Table S2). Results for secondary outcomes between rural and urban patients were mixed (Table S2). Rural participants in the intervention arm tended to have more transmissions of blood pressure, glucose, temperature, and weight measurements with the intervention device and more education module use with a statistically significantly greater number of weight transmissions compared to the usual-care group.
DISCUSSION
Figure 2. Distribution of average number of blood pressure readings, weight measurements, and days on which participants viewed an educational module per month. Values displayed in the box-and-whisker plots are the 5 number summary (minimum, first quartile, median, third quartile, and maximum) and mean 6 standard deviation.
group vs 2 for the usual-care group; P 5 0.8). Proportions with hospital readmissions were similar between groups (intervention vs usual care, 34.3% vs 40.0%; difference, 5.7%; 95% CI, 213.1% to 24.5%). There was no difference between groups in time to dialysis therapy initiation (Table 2). There was also no difference between the intervention and usual-care groups in control of intermediate variables such as blood pressure, diabetes, hyperlipidemia, or smoking (Table 3). At baseline, the intervention and usual-care groups were slightly imbalanced regarding race and systolic blood pressure. There was no difference in time to first event after adjusting for these baseline Am J Kidney Dis. 2016;68(1):41-49
In our randomized controlled trial of patients with moderate to severe CKD, we found that delivery of health care by an interprofessional team using telehealth could be effectively implemented for both rural and urban patients, but did not reduce the risk for death, hospitalization, emergency department visits, or admission to skilled nursing facilities compared to usual care. Similarly, there was no reduction in number of hospitalizations or median duration of hospitalizations. These results were seen despite significant engagement by patients in using the device and numerous virtual visits by the interprofessional team. Although our study was focused on the management of patients with CKD, we viewed CKD as a marker for significant chronic disease burden. A total of 42.6% of patients had diabetes; 34.4%, coronary artery disease; 20.8%, congestive heart failure; and 16.5%, prior stroke. Telehealth in its various forms has been used for the management of chronic diseases including congestive heart failure, chronic obstructive pulmonary disease, coronary artery disease, diabetes, 45
Ishani et al Table 2. Clinical End Points by Treatment Arm With Summary of Cox Proportional Hazards Analysis of Time-to-Event Variables End Point
Intervention (n 5 450)
Usual Care (n 5 150)
208 (46.2)
70 (46.7)
0.98 (0.75-1.29)
13 134 164 18 11
3 40 58 2 2
1.46 1.15 0.92 3.07 1.86
Primary Secondary Death Hospitalization Emergency department visits Admission to skilled nursing facility Initiation of dialysis Othera Second hospitalizationb Second emergency department visitb No. of days of first hospital visitc No. of days of all hospitalizationsc No. of hospitalizations within yeara 0 1 $2
(2.9) (29.8) (36.4) (4.0) (2.4)
(2.0) (26.7) (38.7) (1.3) (1.3)
46 (34.3) 64 (39.0) 3.4 6 3.5 [3] 5.7 6 8.6 [3]
16 (40.0) 26 (44.8) 3.6 6 3.9 [2] 6.2 6 6.5 [3]
316 (70.2) 88 (19.6) 46 (10.2)
110 (73.3) 24 (16.0) 16 (10.7)
HR (95% CI)
(0.42-5.11) (0.80-1.63) (0.68-1.24) (0.71-13.24) (0.41-8.39)
0.89 (0.50-1.57) 0.87 (0.55-1.37)
Note: Values for categorical variables are given as number (percentage); for continuous variables, as mean 6 standard deviation [median]. No significant differences were seen between groups in any of the measured outcomes. Abbreviations: CI, confidence interval; HR, hazard ratio. a Pearson c2 test for the categorized number of hospital admissions within the first year after randomization. b Percentage of those with a first visit. c Wilcoxon rank sum tests for duration of hospitalization.
Probability of Freedom from First Occurrence of Primary Outcome
obesity, depression and other mental health problems, and chronic pain, with mixed results.2-8 Part of the reason for the lack of consensus on the value of
1.0 Usual Care Intervention 0.8
0.6
0.4
0.2
Log−rank p=0.902 0.0 0
1
2
3
4
5
6
7
8
9
10
11
12
Months Since Enrollment No. at risk Usual 150 143 133 125 121 112 106 102 97 89 85 81 Care Intervention 450 432 409 385 357 341 321 302 289 274 261 255
Figure 3. Kaplan-Meier time-to-event occurrences for the primary end point of a composite of death, hospitalization, emergency department visits, and admission to a skilled nursing facility. The hazard ratio for first occurrence with intervention was 0.98 (95% confidence interval, 0.75-1.29; P 5 0.90). 46
telehealth relates to different levels of intervention intensity16,17 and the clinical relevance of the studied outcomes, including both surrogate outcomes and hard clinical end points.18,19 Our results are similar to other studies that have evaluated case management in patients with CKD, although our study involved a greater component of interprofessional care than both these studies and also included home video monitoring and education.20-23 However, despite the intensity and broad scope of our intervention, we were unable to detect an improvement in clinical outcomes between groups. Patients within rural areas have a number of barriers to optimal health care, including higher prevalences of obesity, hypertension, diabetes, and tobacco use.24 Rural areas also have a scarcity of nephrologists25 and these patients are less likely to visit a nephrologist.25 Previous studies have suggested that patients with CKD stages 3 to 4 in rural areas have poorer process-based markers of care compared with urban patients, such as measurement of either hemoglobin A1c in patients with diabetes, assessment of urinary albumin excretion, or use of angiotensinconverting enzyme inhibitors or angiotensin receptor blockers among patients with diabetes or significant proteinuria.26 Rural patients with CKD also have a greater risk for hospitalization and all-cause mortality compared with urban patients.26 This pattern of poorer care appears to have been reversed in our study, with rural intervention patients having a trend toward a lower incidence of the primary outcome. The lack of statistical significance in this group is Am J Kidney Dis. 2016;68(1):41-49
Telehealth for Patients With CKD Table 3. Intermediate Outcomes at Study End by Treatment Group Outcome
Total
Intervention
Usual Care
Pa
No. of participants with complete intermediate outcome data No. with SBP # 140 mm Hg, LDL-C # 100 mg/dL, and HbA1c # 8% at study end
444
340 185 (54.4)
104 53 (51.0)
0.5
No. of participants with baseline SBP . 140 mm Hg and complete data No. with SBP # 140 mm Hg at study end
174
135 72 (53.3)
39 20 (51)
0.8
No. of participants with baseline LDL-C . 100 mg/dL and complete data No. with LDL-C # 100 mg/dL at study end
76
61 31 (51)
15 8 (53)
0.9
No. of participants with baseline HbA1c . 8% and complete data No. with HbA1c # 8% at study end
48
33 14 (42)
15 3 (33)
0.6
No. of participants who were smokers at baseline No. no longer smoking at study end
52
40 9 (23)
12 5 (42)
0.3
Note: Values are given as number (percentage). Conversion factor for LDL-C in mg/dL to mmol/L, 30.02586. Abbreviations: HbA1c, hemoglobin A1c; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure. a Pearson c2 test, except for Fisher exact test for smoking status due to small expected cell counts.
likely because the study was not powered to evaluate outcomes within subgroups. Although the overall study was negative, rural patients may be a selected subgroup in which the use of telehealth and interprofessional care may offer benefits, particularly in areas that are scarce in subspecialty care. This subgroup needs to be specifically tested in future studies. There are a number of possible reasons that our study was unable to demonstrate an overall benefit in outcomes. First, aggressive home monitoring and interprofessional care may simply be ineffective in altering patient outcomes. Alternatively, the team may have identified a greater number of issues that may be ignored by patients without constant monitoring.27 Another possible explanation of negative results is that patients in the intervention group did not engage in using the system. Although a small number of patients (3.8%) did not engage in the system, most were fully engaged after 1 year of follow-up. Adherence rates in our study were likely significantly better than can be achieved in usual clinical care given the resources available during the trial. Another possibility is that the duration of follow-up was too short. It may be that differences between groups would be observed with longer follow-up. Finally, it could be that individuals in the usual-care group received very aggressive therapy compared with those receiving traditional usual care. We did not track interventions provided to the usual-care group. However, during the course of our study, the VA was changing the model of primary care to that of a patient-centered medical home, called PACT (Patient Aligned Care Team), along with implementation of a structured telehealth program. Implementation of the PACT and home telehealth have independently been associated with reductions in emergency department visits, hospitalizations,21-23,28-30 and improved control of Am J Kidney Dis. 2016;68(1):41-49
hypertension.24,31 As an example, 51.3% of patients in the usual-care group had incident control of hypertension, an incident control rate that is substantially greater than community standards. Aggressive implementation of therapy in the usual-care group (PACT, telehealth, and case management) may have been why we were unable to demonstrate a difference in either intermediate or clinical outcomes between groups. Another consideration is that disease progression does not always follow a pattern of progressive decline, but instead, patients may suddenly deteriorate. Therefore, early intervention to prevent clinical deterioration may not always be possible. Also, the specific monitored physiologic parameters may not be those that best predict disease progression. We powered the study assuming a 50% event rate in the usualcare group, with an observed event rate of 46.3%. Our loss to follow-up was significantly smaller than anticipated. Our study had sufficient power for effects of this magnitude, but not for smaller effects. Although our study results were negative, the ranges of the CIs for our effect estimates do not rule out a clinically significant benefit for a similar intervention. Our study had numerous strengths. We examined relevant and clinically important primary end points and enrolled a patient population at high risk for adverse outcomes. We attempted to target numerous facets of care through the use of a diverse interprofessional team. We sought to engage and activate patients in their care, provide in-home education, facilitate visits using video, manage chronic diseases, and identify and address acute issues. The intervention included medication management and comprehensive yet simple-to-understand educational modules. We included both urban and rural patients and randomly assigned patients in their homes, potentially enhancing the generalizability of our 47
Ishani et al
findings.10 Finally, all our outcome assessors were blinded to study assignment. A potential weakness of our study is that patients in the intervention group were likely queried more often than those in the usual-care group about hospitalizations and emergency department visits, likely enhancing their recall of these events during the outcome assessment calls and potentially biasing our results toward more adverse outcomes reported in the intervention group. However, no difference between groups was seen when we confined our analysis to only VA electronic medical record2ascertained events. We performed a complex intervention with many different aspects. It is unclear whether any individual component of our intervention could be associated with benefit that is overwhelmed by the overall study results. However, our goal was to deliver a comprehensive intervention to determine whether a fully integrated model of care could alter patient outcomes. Given the overall negative study findings, it is unlikely that individual components of care would result in a different outcome. The generalizability of our results to other settings and other patient populations remains to be determined. In summary, a strategy of telehealth by an interprofessional team could be effectively implemented in both rural and urban patients but did not alter outcomes in patients with moderate to severe CKD compared with usual care. Strategies for improving the care of this population are needed.
ACKNOWLEDGEMENTS The Center for Innovative Kidney Care, which is based at Minneapolis VAHCS, comprises Melissa Atwood, Ann Bangerter, Theresa Borah, Olga Brusilovsky, Andrea Cutting, Mia Dobbs, Wendy Elofson, Kori Geinert, Kanika Gupta, Kendra Hackney, Ron Howell, Miles Jarzyna, Quinn Kellerman, Elaine Kroska, LauraJean Krueger, Narlina Lalani, Shannon McCutcheon, Alec Otteman, Megan Plumstead, Bethany Roberts, Keri Rowe, Jackie Rust, Rita Schmitt, Herb Stockley, Jaclyn Strutt, Melvis Tambi, and Callen Weispfennig. The members of Data Safety Monitoring Board for this trial were Kristine E. Ensrud, MD, MPH, Minneapolis VAHCS; Jeffrey Connaire, MD, Minneapolis VAHCS; and Brent Taylor, PhD, VAHCS. Nan Booth, MSW, MPH, ELS, provided manuscript editing. Portions of this work were presented at the 2014 American Society of Nephrology Kidney Week, November 11-16, Philadelphia, PA. Support: The project was funded by grants from the VA Center for Innovation to Dr Ishani and Dr Rosenberg. The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. Financial Disclosure: Dr Rosenberg has received honoraria from Wolter Kluwer Health for serving as an author for UpToDate and from the American Society of Nephrology for service as education director. The other authors declare that they have no relevant financial interests. Contributions: Study concept and design: AI, MER; data acquisition, analysis or interpretation: AI, JC, DP, SO, BC, SN, 48
DN, MER; statistical analysis: BC, DN; obtainment of funding: AI, MER; administrative, technical, or material support: AI, JC, DP, SO, BC, SN, DN, MER; study supervision: AI, MER. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved. AI takes responsibility that this study has been reported honestly, accurately, and transparently; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned and registered have been explained. Peer Review: Evaluated by 2 external peer reviewers, a Statistical Editor, a Co-Editor, and the Editor-in-Chief.
SUPPLEMENTARY MATERIAL Table S1: Clinical end points by strata at randomization with summary of proportional hazards analysis of time-to-event variables. Table S2: Clinical end points stratified by setting at randomization with summary of proportional hazards analysis of time-toevent variables. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2016.01.018) is available at www.ajkd.org
REFERENCES 1. Berwick DM, Nolan TW, Whittington J. The triple aim: care, health, and cost. Health Aff (Millwood). 2008;27(3):759-769. 2. Wennberg DE, Marr A, Lang L, O’Malley S, Bennett G. A randomized trial of a telephone care-management strategy. N Engl J Med. 2010;363(13):1245-1255. 3. Chaudhry SI, Mattera JA, Curtis JP, et al. Telemonitoring in patients with heart failure. N Engl J Med. 2010;363(24):2301-2309. 4. Kroenke K, Krebs EE, Wu J, Yu Z, Chumbler NR, Bair MJ. Telecare collaborative management of chronic pain in primary care: a randomized clinical trial. JAMA. 2014;312(3):240-248. 5. Goldstein RS, O’Hoski S. Telemedicine in COPD: time to pause. Chest. 2014;145(5):945-949. 6. Gellis ZD, Kenaley BL, Ten HT. Integrated telehealth care for chronic illness and depression in geriatric home care patients: the Integrated Telehealth Education and Activation of Mood (I-TEAM) study. J Am Geriatr Soc. 2014;62(5):889-895. 7. Wade VA, Karnon J, Elshaug AG, Hiller JE. A systematic review of economic analyses of telehealth services using real time video communication. BMC Health Serv Res. 2010;10(10):233. 8. McLean S, Sheikh A, Cresswell K, et al. The impact of telehealthcare on the quality and safety of care: a systematic overview. PLoS One. 2013;19(8):e71238. 9. Margolis KL, Asche SE, Bergdall AR, et al. Effect of home blood pressure telemonitoring and pharmacist management on blood pressure control: a cluster randomized clinical trial. JAMA. 2013;310(1):46-56. 10. Rifkin DE, Abdelmalek JA, Miracle CM, et al. Linking clinic and home: a randomized controlled clinical effectiveness trial of real-time, wireless blood pressure monitoring for older patients with kidney disease and hypertension. Blood Press Monit. 2013;18(1):8-15. 11. Gordon EJ, Fink JC, Fischer MJ. Telenephrology: a novel approach to improve coordinated and collaborative care for chronic kidney disease. Nephrol Dial Transplant. 2013;28(4):972-981. 12. US Census Bureau. 2010 Census Urban and Rural Classification and Urban Area Criteria. https://www.census.gov/geo/ reference/ua/urban-rural-2010.html. Accessed August 1, 2015. Am J Kidney Dis. 2016;68(1):41-49
Telehealth for Patients With CKD 13. Wagner EH, Austin BT, Von KM. Organizing care for patients with chronic illness. Milbank Q. 1996;74(4):511-544. 14. Hibbard JH, Stockard J, Mahoney ER, Tusler M. Development of the Patient Activation Measure (PAM): Conceptualizing and Measuring Activation in Patients and Consumers. Health Serv Res. 2004;39(4 Pt 1):1005-1026. 15. Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl. 2013;3:1-150. 16. Feltner C, Jones CD, Cene CW, et al. Transitional care interventions to prevent readmissions for persons with heart failure: a systematic review and meta-analysis. Ann Intern Med. 2014;160(11):774-784. 17. Inglis SC, Clark RA, McAlister FA, Stewart S, Cleland JG. Which components of heart failure programmes are effective? A systematic review and meta-analysis of the outcomes of structured telephone support or telemonitoring as the primary component of chronic heart failure management in 8323 patients: abridged Cochrane review. Eur J Heart Fail. 2011;13(9):1028-1040. 18. Omboni S, Gazzola T, Carabelli G, Parati G. Clinical usefulness and cost effectiveness of home blood pressure telemonitoring: meta-analysis of randomized controlled studies. J Hypertens. 2013;31(3):455-467. 19. Worswick J, Wayne SC, Bennett R, et al. Improving quality of care for persons with diabetes: an overview of systematic reviews - what does the evidence tell us? Syst Rev. 2013;7(2):1-14. 20. Barrett BJ, Garg AX, Goeree R, et al. A nurse-coordinated model of care versus usual care for stage 3/4 chronic kidney disease in the community: a randomized controlled trial. Clin J Am Soc Nephrol. 2011;6(6):1241-1247. 21. Hopkins RB, Garg AX, Levin A, et al. Cost-effectiveness analysis of a randomized trial comparing care models for chronic kidney disease. Clin J Am Soc Nephrol. 2011;6(6): 1248-1257. 22. Van Zuilen AD, Bots ML, Dulger A, et al. Multifactorial intervention with nurse practitioners does not change cardiovascular
Am J Kidney Dis. 2016;68(1):41-49
outcomes in patients with chronic kidney disease. Kidney Int. 2012;82(11):710-717. 23. Peeters MJ, van Zuilen AD, van den Brand JA, et al. Nurse practitioner care improves renal outcome in patients with CKD. J Am Soc Nephrol. 2014;25(2):390-398. 24. Agency for Healthcare Research and Quality. 2010 National Healthcare Quality Report. January 2013. Rockville, MD. http://www.ahrq.gov/research/findings/nhqrdr/nhqr10/index.html. Accessed August 1, 2015. 25. Yan G, Cheung AK, Ma JZ, et al. The associations between race and geographic area and quality-of-care indicators in patients approaching ESRD. Clin J Am Soc Nephrol. 2013;8(4):610-618. 26. Rucker D, Hemmelgarn BR, Lin M, et al. Quality of care and mortality are worse in chronic kidney disease patients living in remote areas. Kidney Int. 2011;79(2):210-217. 27. Weinberger M, Oddone EZ, Henderson WG. Does increased access to primary care reduce hospital readmissions? Veterans Affairs Cooperative Study Group on Primary Care and Hospital Readmission. N Engl J Med. 1996;334(22):1441-1447. 28. Hebert PL, Liu CF, Wong ES, et al. Patient-centered medical home initiative produced modest economic results for Veterans Health Administration, 2010-12. Health Aff (Millwood). 2014;33(6):980-987. 29. Nelson KM, Helfrich C, Sun H, et al. Implementation of the patient-centered medical home in the Veterans Health Administration: associations with patient satisfaction, quality of care, staff burnout, and hospital and emergency department use. JAMA Intern Med. 2014;174(8):1350-1358. 30. Darkins A, Kendall S, Edmonson E, Young M, Stressel P. Reduced cost and mortality using home telehealth to promote selfmanagement of complex chronic conditions: a retrospective matched cohort study of 4,999 veteran patients. Telemed J E Health. 2015;21(1):70-76. 31. O’Neill JL, Cunningham TL, Wiitala WL, Bartley EP. Collaborative hypertension case management by registered nurses and clinical pharmacy specialists within the Patient Aligned Care Teams (PACT) model. J Gen Intern Med. 2014;29(suppl 2):S675S681.
49