The Heart Failure Readmission Intervention by Variable Early Follow-up (THRIVE) Study
WHAT IS KNOWN
-
Readmission rates within 30 days after hospitalization for heart failure remain high despite intensive efforts.
-
In-person clinic follow-up within 7 days after discharge from a heart failure hospitalization is associated with lower 30-day readmission.
-
In-person clinic follow-up within 7 days is resource intensive and difficult to achieve due to patient-level and health system obstacles, especially during the current coronavirus disease 2019 (COVID-19) pandemic.
WHAT THE STUDY ADDS
Introduction
See Editorial by Ezekowitz
Heart failure (HF) affects an estimated 5.7 million Americans1 and remains the leading cause of hospitalization among older adults in the United States.2 Reducing hospitalization for HF and subsequent readmissions is a nationally recognized priority, as ≥20% of Medicare beneficiaries hospitalized for HF are readmitted within 30 days and readmission rates have remained high despite various efforts.3–5 Hospitals with higher than expected risk-adjusted readmission rates for HF are penalized through reduced Medicare payments,6 and publicly reported ratings are heavily weighted against hospitals with higher than expected 30-day readmission rates.
Given the scarcity of strong clinical trial evidence to guide interventions, insurers, and health systems often spend significant resources on various transitional care strategies without knowing their net effectiveness.7 A recent meta-analysis reported that home visitation and multidisciplinary clinics may be associated with lower readmission risk at 3 to 6 months, but data are lacking for 30-day readmission, and included studies had various limitations.8 Aside from type of follow-up, the timing of postdischarge interventions may also be important, as previous studies suggest that early postdischarge follow-up may reduce subsequent readmission rates, but results have been mixed.9–14 We previously reported results from a retrospective cohort study that follow-up in the first 7 days postdischarge—mostly done through in-person clinic visits—was independently associated with a 19% lower risk of readmission compared with later follow-up among hospitalized patients with HF treated within an integrated healthcare delivery system.15 Importantly, early follow-up telephone visits mostly performed by nonphysician providers were also associated with a trend toward lower readmission rates, which has potentially important implications because phone-based contacts with nonphysicians may be more convenient for patients, enhance completion of early follow-up and be more practical and cost-effective when implemented on a large scale, especially during the current coronavirus disease 2019 (COVID-19) pandemic.16
To evaluate the effectiveness and efficiency of timely, structured telephone visits, we performed a pragmatic randomized clinical trial comparing 2 strategies of early follow-up within 7 days after discharge for HF. The usual care strategy was an initial in-person clinic visit with a physician, with any additional follow-up guided by the physician at that visit. The alternative strategy was an initial structured telephone visit with a nonphysician who followed protocol-driven next steps under physician supervision. We evaluated for differences in 30-day risks of readmission and death, along with overall clinic and telephone appointment utilization in the first 7 days after discharge.
Methods
Individual participant data that underlie the results reported in this article (text, tables, figures, and appendices) are available after deidentification from Dr Alan S. Go upon reasonable request between 9 months and up to 36 months following article publication. This study was registered on ClinicalTrials.gov.
Study Population
We first identified all patients aged ≥21 years who were hospitalized in 16 hospitals between January 15, 2017 and March 31, 2018, within Kaiser Permanente Northern California, a large integrated healthcare delivery system currently providing comprehensive inpatient and outpatient care for a sociodemographically diverse population of >4.4 million persons in Northern California.17
This study was approved by the Kaiser Permanente Northern California institutional review board. We obtained a waiver of informed consent given the minimal risk involved, as both treatment arms included follow-up options available as part of usual care.
Study Eligibility
We identified hospitalized adults with HF in participating facilities by conducting daily searches of International Classification of Diseases, Tenth Edition diagnosis codes for HF as the primary hospital problem, or an International Classification of Diseases, Tenth Edition diagnosis code for a HF-related sign or symptom as the primary hospital problem in combination with a HF-specific diagnosis code as a secondary problem within a comprehensive electronic health record (EHR) system. Under the supervision of board-certified physicians, a trained research nurse then manually reviewed medical records to ensure HF was the primary diagnosis based on Framingham criteria,18 supportive laboratory results, and diagnostic impressions of treating physicians. We excluded patients who were not health plan members, died before discharge, were discharged to a location other than home, enrolled in home hospice, were receiving chronic dialysis, or if a treating inpatient physician requested the patient not be randomized.
Randomization Approach
Within each participating hospital, a SAS software-based algorithm was used to randomly assign eligible patients in a 1:1 ratio to one of 2 strategies of early follow-up within 7 days after discharge: (1) initial follow-up guided by in-person clinic appointment with a primary care provider or (2) initial follow-up guided by a structured telephone appointment with a nurse or pharmacist trained in HF care supported by established treatment protocols and supervising physicians. Randomization was performed before hospital discharge; as such, all patients were discharged with knowledge and written documentation of their initial scheduled appointment type. Neither researchers nor clinical staff could be blinded to treatment assignment because treatment appointments were documented in the integrated EHR and clinical staff delivered treatment.
Intervention
Patients were enrolled and assigned to the appropriate initial follow-up by a research nurse. Patients assigned to follow-up guided by an initial in-person clinic appointment were scheduled primarily with their primary care physician who provided usual care that included uniform access to preexisting evidence-based HF guidelines but did not include structured HF care protocols. Patients assigned to follow-up guided by an initial telephone appointment were called by a nurse or pharmacist who were previously trained and experienced using a structured HF protocol. The protocol included directions for titrating diuretics and other HF-related medications and ordering of laboratory tests, based on reported HF symptoms, weights, and vital signs (Appendix I in the Data Supplement). They also had immediate access to supervising physicians for support and could arrange for expedited subsequent in-person physician appointments as deemed necessary, along with scheduling further follow-up telephone appointments if considered clinically necessary.
Coordination of care after the initial visit followed usual Kaiser Permanente Northern California practice guidelines for HF. It was anticipated that after the initial postdischarge patient contact, whether by a nonphysician on the telephone or by a physician in clinic, patients would receive various types of subsequent follow-up appointments and at different time intervals, as needed. For example, patients with reduced left ventricular ejection fraction (LVEF <40%) were typically scheduled for follow-up by a cardiologist through an in-person clinic visit and also enrolled longitudinally over time in the existing HF care management program usually administered by telephone. For patients with preserved or mid-range LVEF, follow-up was determined at the discretion of their primary care physician.
Outcomes
Our primary outcome was readmission for HF within 30 days after discharge using the same algorithm used to identify index HF hospitalizations based on comprehensive EHR data. Secondary outcomes included 30-day readmission for any cause, 30-day all-cause death, successful completion of early follow-up, and 7-day clinic and telephone appointment utilization. These outcomes were chosen because 30-day readmission and death along with 7-day follow-up are generally accepted quality metrics for postdischarge patients with HF. Deaths were systematically ascertained using data from manual EHR review, health plan administrative databases, and state death certificates.19 Patients who died within 30 days postdischarge were censored on the day of death in the readmission analyses. No participants disenrolled from the health plan within 30 days after discharge, so there was no loss to follow-up.
Covariates
Age, gender, and self-reported race/ethnicity were obtained from administrative databases. Educational attainment and household income were estimated using residential block-level data from United States census data.20 Relevant cardiovascular conditions and risk factors were defined using EHR data based on diagnosis or procedure codes, laboratory results and medications using previously validated algorithms.20–23 Body mass index and blood pressure were based on the most recent preadmission clinic measurements. Estimated glomerular filtration rate was calculated from outpatient serum creatinine concentration.24 Peak BNP (B-type natriuretic peptide) and lowest hemoglobin levels during the index hospitalization were identified from laboratory databases. The American Heart Association Get With The Guidelines–Heart Failure risk score25 was also calculated.
Classification of HF with preserved (HFpEF), reduced, or mid-range LVEF was based on quantitative and qualitative assessments of left ventricular function from results of echocardiograms, radionuclide scintigraphy, other nuclear imaging modalities, and left ventriculography using available EHR information closest to the index admission. HFpEF was defined as LVEF ≥50% or a physician’s qualitative assessment of preserved or normal systolic function; HF with mid-range LVEF was defined as LVEF 41% to 49% or a physician’s qualitative assessment of mildly reduced systolic function, and HF with reduced LVEF defined as LVEF ≤40% or based on a physician’s qualitative assessment of moderate, moderate to severe, or severe systolic dysfunction.
Statistical Approach
We used SAS, version 9.4 (Cary, NC) for all analyses. Baseline characteristics were compared using standardized differences. The cumulative incidence for each outcome by group was plotted using Kaplan-Meier product limit estimates and compared using a log-rank test. Based on a conservative effect size threshold of ≥0.10 for potential differences in baseline characteristics between randomized groups, we also conducted Cox proportional hazards regression with additional adjustment for preadmission body mass index, preadmission diastolic blood pressure, lowest hemoglobin during index hospitalization, and month of enrollment. Finally, we performed prespecified subgroup analyses by age (<80 versus ≥80 years), gender, and baseline LVEF (≤40% or >40%). In original power calculations, we estimated a 15% 30-day readmission rate for HF in the usual care arm (ie, follow-up guided by an initial in-person physician visit). We targeted a sample size of 2544 to detect a minimum detectable absolute difference of 3.75% between groups (ie, 15.0% versus 11.25%), given a 2-sided α=0.05 and power of 80%. Reducing readmission rates after HF hospitalizations has been known to be difficult, so an absolute reduction of this magnitude would be considered clinically significant. The analysis is modified from that which was preregistered in clinicaltrials.gov due to suggestions from peer reviewers. The results using the a priori analytic approach yielded similar results to those presented herein and are available in Appendix II in the Data Supplement.
Results
Baseline Characteristics
During the study period, we identified 3104 confirmed HF admissions that met other eligibility criteria. We excluded 732 patients who were discharged before follow-up scheduling could be completed or were already booked for both types of follow-up and 281 patients who were previously enrolled and subsequently readmitted with HF (Figure 1). Thus, 2091 eligible patients were randomly assigned to follow-up guided by an initial structured telephone visit with a nonphysician (N=1027) or follow-up guided by an initial in-person physician clinic visit (N=1064), with recruitment stopped due to funding constraints. Given the actual enrolled sample size and observed 30-day HF readmission rate, we had 80% power to detect a minimum absolute difference between groups of 3.48% for our primary outcome.
Figure 1. Cohort assembly of adults hospitalized for heart failure who were randomized to either a telephone visit with a trained nurse or pharmacist or an in-person physician clinic visit within 7 days post-discharge.
Mean age of enrolled participants was 78 years (N=933 for age ≥80 years), 56% were men, 41% were persons of color, 56% had HFpEF and 34% had HF with reduced LVEF (N=706 for LVEF <40%; Table). Overall, using a standardized mean difference threshold of >0.10, there were no differences in baseline characteristics between groups except for slightly higher mean nadir hemoglobin and a higher frequency of missing inpatient hemoglobin in those randomized to an initial in-person clinic visit (Table).
| Overall | Initial Structured Telephone Visit With Nonphysician | Initial In-Person Physician Clinic Visit | Effect Size | |
|---|---|---|---|---|
| (N=2091) | (N=1027) | (N=1064) | ||
| Index age | ||||
| Median (IQR) | 78.3 (68.7–85.6) | 78.3 (68.3–85.6) | 78.4 (69.0–85.7) | 0.06 |
| Gender | ||||
| Men | 1177 (56.3) | 563 (54.8) | 614 (57.7) | 0.03 |
| Women | 914 (43.7) | 464 (45.2) | 450 (42.3) | |
| Race | ||||
| White | 1264 (60.4) | 615 (59.9) | 649 (61.0) | 0.05 |
| Black | 180 (8.6) | 77 (7.5) | 103 (9.7) | |
| Asian/Pacific Islander | 246 (11.8) | 124 (12.1) | 122 (11.5) | |
| Other/unknown | 401 (19.2) | 211 (20.5) | 190 (17.9) | |
| Hispanic ethnicity | 303 (14.5) | 166 (16.2) | 137 (12.9) | 0.05 |
| Low educational attainment | 305 (14.6) | 157 (15.3) | 148 (13.9) | 0.02 |
| Low annual household income | 114 (5.5) | 57 (5.6) | 57 (5.4) | 0.00 |
| Left ventricular ejection fraction | ||||
| ≥50% | 1173 (56.1) | 572 (55.7) | 601 (56.5) | 0.03 |
| 41%–49% | 195 (9.3) | 91 (8.9) | 104 (9.8) | |
| <40% | 706 (33.8) | 354 (34.5) | 352 (33.1) | |
| Unknown | 17 (0.8) | 10 (1.0) | 7 (0.7) | |
| GWTG-HF risk score | ||||
| Mean (SD) | 42.4 (7.8) | 42.4 (7.8) | 42.5 (7.7) | 0.01 |
| Medical history | ||||
| Diabetes mellitus | 991 (47.4) | 500 (48.7) | 491 (46.1) | 0.03 |
| Atrial fibrillation/flutter | 1103 (52.7) | 526 (51.2) | 577 (54.2) | 0.03 |
| Mitral or aortic valvular heart disease | 541 (25.9) | 246 (24.0) | 295 (27.7) | 0.04 |
| Chronic lung disease | 999 (47.8) | 478 (46.5) | 521 (49.0) | 0.02 |
| Tobacco use | ||||
| None | 873 (41.8) | 435 (42.4) | 438 (41.2) | 0.04 |
| Former | 1096 (52.4) | 542 (52.8) | 554 (52.1) | |
| Current | 122 (5.8) | 50 (4.9) | 72 (6.8) | |
| Systolic blood pressure, mm Hg | ||||
| Mean (SD) | 126.4 (20.1) | 126.5 (20.1) | 126.3 (20.2) | 0.01 |
| Diastolic blood pressure, mm Hg | ||||
| Mean (SD) | 67.6 (13.6) | 68.3 (13.6) | 66.9 (13.5) | 0.10 |
| Body mass index, kg/m2 | ||||
| Mean (SD) | 30.4 (8.4) | 30.9 (8.6) | 30.0 (8.2) | 0.10 |
| Lowest inpatient hemoglobin | ||||
| Mean (SD) | 11.0 (2.2) | 10.9 (2.2) | 11.2 (2.2) | 0.11 |
| Unknown | 468 (22.4) | 181 (17.6) | 287 (27.0) | 0.11 |
| Preadmission medication use | ||||
| Angiotensin converting enzyme inhibitor | 695 (33.2) | 339 (33.0) | 356 (33.5) | 0.00 |
| Angiotensin II receptor blocker | 475 (22.7) | 222 (21.6) | 253 (23.8) | 0.03 |
| β-blocker | 1450 (69.3) | 705 (68.6) | 745 (70.0) | 0.01 |
| Calcium channel blocker | 701 (33.5) | 349 (34.0) | 352 (33.1) | 0.01 |
| Diuretic | 1526 (73.0) | 749 (72.9) | 777 (73.0) | 0.00 |
| Aldosterone receptor antagonist | 168 (8.0) | 78 (7.6) | 90 (8.5) | 0.02 |
| α-blocker | 439 (21.0) | 204 (19.9) | 235 (22.1) | 0.03 |
| Nitrate | 417 (19.9) | 209 (20.4) | 208 (19.5) | 0.01 |
| Hydralazine | 370 (17.7) | 175 (17.0) | 195 (18.3) | 0.02 |
| Digoxin | 168 (8.0) | 75 (7.3) | 93 (8.7) | 0.03 |
| Statin | 1406 (67.2) | 689 (67.1) | 717 (67.4) | 0.00 |
| Nonstatin lipid-lowering agent | 56 (2.7) | 30 (2.9) | 26 (2.4) | 0.01 |
| Antiplatelet | 226 (10.8) | 103 (10.0) | 123 (11.6) | 0.02 |
| Anticoagulant | 738 (35.3) | 345 (33.6) | 393 (36.9) | 0.03 |
| Potassium supplement | 553 (26.4) | 277 (27.0) | 276 (25.9) | 0.01 |
| Nonsteroidal anti-inflammatory drug | 117 (5.6) | 55 (5.4) | 62 (5.8) | 0.01 |
Follow-Up Completion and Type
Completed follow-up of any kind within 7 days postdischarge was 92% in the initial nonphysician telephone guided group compared with 79% in the initial physician clinic visit guided group (P<0.001). Among participants assigned to initial nonphysician telephone-guided follow-up, within the first 7 days postdischarge, 44% (N=456) had a telephone visit only, 39% (N=409) had a telephone visit and also required a subsequent clinic visit, and 9% (N=98) had a clinic visit only. Among participants assigned to an initial in-person physician clinic-guided follow-up, within the first 7 days postdischarge, 68% (N=730) attended a clinic visit only, 9% (N=103) had a clinic visit followed by a telephone visit, and 2% (N=18) had a telephone visit only. Thus, the overall frequency of clinic visits during the first 7 days postdischarge was significantly lower in the group assigned to initial nonphysician telephone guided follow-up (48%) compared with those assigned to initial physician clinic-guided follow-up (77%; P<0.001). In the nonphysician telephone-guided arm, the crossover rate was 9% (ie, received a clinic visit without initial telephone visit within 7 days) and in the initial physician clinic-guided arm, the crossover rate was 2% (ie, received a telephone visit without initial clinic visit within 7 days).
Thirty-Day Readmission for HF
The risk of being readmitted for HF within 30 days was 10.6% in those assigned to initial physician clinic-guided follow-up compared with 8.6% in those assigned initial nonphysician telephone guided follow-up (P=0.11; Figure 2A). In multivariable analysis that adjusted for potential differences in baseline characteristics, assignment to the nonphysician telephone guided arm was also not significantly associated with a difference in HF-specific readmission risk (adjusted hazard ratio [aHR], 0.81 [95% CI, 0.59–1.11]). No significant associations were seen in subgroup multivariable analyses: men (aHR,0.89 [95% CI, 0.58–1.36]), women (aHR, 0.71 [95% CI, 0.44–1.14]), age <80 years (aHR, 0.85 [95% CI, 0.56–1.29]), age ≥80 years (aHR, 0.73 [95% CI, 0.44–1.19]), LVEF ≤40% (aHR, 0.72 [95% CI, 0.44–1.18]), and LVEF >40% (aHR, 0.87 [95% CI, 0.57–1.32]).
Figure 2. Cumulative incidence of 30-day all-cause outcomes by treatment arm.A, Heart failure-specific readmission. B, All-cause readmission. C, All-cause mortality.
Thirty-Day Readmission for Any Cause
The risk of readmission for any cause within 30 days postdischarge was 20.6% in the initial physician clinic-guided arm compared with 18.8% in the initial nonphysician telephone-guided arm (P=0.30; Figure 2B). In a multivariable analysis that adjusted for potential differences in baseline characteristics, there was no significant difference in 30-day readmission (aHR, 0.82 [95% CI, 0.66–1.02]) for assignment to the nonphysician telephone-guided arm versus the physician clinic-guided arm. In subgroup analyses, assignment to the initial nonphysician telephone appointment arm was associated with a significantly lower adjusted risk of all-cause readmission in men (aHR, 0.70 [95% CI, 0.53–0.94]) but not women (aHR, 0.98 [95% CI, 0.70–1.38]). Assignment to the initial nonphysician telephone appointment arm was associated with a significant difference in all-cause readmission in those aged ≥80 years (aHR, 0.70 [95% CI, 0.50–0.97]) but not in those aged <80 years (aHR, 0.89 [95% CI, 0.67–1.20]). Finally, assignment to the initial nonphysician telephone appointment arm was not associated with a significant difference in adjusted risk for patients with LVEF >40% (aHR, 0.75 [95% CI, 0.57–1.00]) or LVEF ≤40% (aHR, 0.93 [95% CI, 0.65–1.33]).
Thirty-Day Mortality
Thirty-day all-cause mortality was 4.6% in patients assigned to a physician clinic-guided arm compared with 4.0% in those assigned to the nonphysician telephone guided arm (P=0.49; Figure 2C). After adjustment for potential differences in baseline characteristics, all-cause mortality was not significantly different for those assigned to initial nonphysician telephone guided follow-up (aHR, 0.75 [95% CI, 0.48–1.18]). Similarly, there were no significant adjusted effects of assignment to nonphysician telephone guided follow-up in prespecified subgroups: men (aHR, 0.70,[95% CI, 0.38–1.28]), women (aHR, 1.09 [95% CI, 0.57–2.06]), age <80 years (aHR, 0.86 [95% CI, 0.41–1.79]), age ≥80 years (aHR, 0.82 [95% CI, 0.49–1.36]), LVEF ≤40% (aHR, 0.66 [95% CI, 0.33–1.30]), or LVEF >40% (aHR, 0.81 [95% CI, 0.45–1.45).
Discussion
In a large, diverse community-based population receiving care within an integrated healthcare delivery system, early follow-up after HF hospitalization guided by an initial nonphysician structured telephone visit with an option to schedule an expedited follow-up clinic visit and additional telephone calls was not associated with significantly different 30-day HF or any-cause readmission and mortality rates compared with being assigned to usual care of early follow-up guided by an initial in-person primary care physician clinic visit. However, the initial nonphysician telephone follow-up strategy effectively increased the rate of successful early follow-up and decreased the number of physician clinic visits in the first 7 days after discharge, which may be more efficient for healthcare delivery systems and more convenient for patients with HF, especially during the current COVID-19 pandemic and resulting quarantine orders.25
Patients most vulnerable to adverse outcomes may be those who find it most difficult to keep in-person appointments soon after hospital discharge due to underlying frailty, lack of social support, caregiver fatigue, or transportation challenges. Telephone appointments allow patients to be conveniently contacted in their own location and repeat calls were made as necessary. In contrast, rescheduling of missed clinic visits can often lead to delayed follow-up until after the initial target 7-day postdischarge period. Importantly, those assigned to initial telephone visit completed 7-day follow-up more often (92%) than those assigned to initial clinic visits (79%). Our study thus supports that access to care within 7 days after discharge with telephone appointments is superior to in-person clinic visits.
Telephone follow-up providers were nurses and pharmacists, while clinic visit providers were physicians, yet there were no significant differences in outcomes between both groups for readmission and death. A possible reason is that telephone appointment providers, although nonphysicians, were trained in HF-specific care, while primary care physicians providing in-clinic follow-up had variable comfort levels treating HF. Telephone providers followed uniform structured protocols for titrating HF medications, ordering laboratory studies, and reinforcing discharge lifestyle instructions. Telephone providers also had direct access to supervising physicians who were primarily cardiologists, served as entry points to social work, and other specialized care teams (eg, diabetes mellitus, anticoagulation management) and would often directly book primary care appointments and cardiology referrals as needed within our integrated healthcare delivery system framework.
In-person clinic providers were predominantly the patients’ own primary care physicians trained in general internal medicine or family medicine. While these physicians had uniform access to the same regional HF treatment guidelines, primary care clinics at participating medical centers did not systematically employ specific HF management protocols. In-clinic follow-up with a primary care physician is consistent with contemporary usual care for most health systems pre-COVID-19, especially given it is not necessarily feasible to have early cardiologist in-person follow-up for all patients. Our results suggest that trained nonphysician providers who are experienced in HF-specific care and supported through an integrated EHR, structured protocols, and physician backup may be able to achieve similar outcomes as physicians in recently discharged patients with HF. However, our results may not be achievable with less well-trained and empowered nonphysician telephone providers or in a health system without comprehensive care, a fully integrated EHR, or reduced ability to quickly escalate care, if needed.
Among patients randomly assigned to initial nonphysician telephone guided follow-up, 92% completed the telephone visit within 7 days as planned. As anticipated, a significant proportion (39%) also received a subsequent early in-person physician clinic visit. These patients should not be regarded as crossover between treatment arms, because clinic visits were an important and intentional part of the telephone-guided follow-up strategy. In-person clinic visits after initial telephone follow-up likely reflected escalation of care by nonphysician telephone providers to address early clinical decompensation and was an important tool in selected patients in the initial telephone-guided arm. Our study shows that telephone-guided follow-up can significantly reduce early clinic visits by serving as a means to triage patients, which led to more efficient care downstream without sacrificing clinical outcomes.
A recent retrospective study suggested that reduction in 30-day HF readmissions among Medicare beneficiaries since 2010 was associated with increased 30-day mortality in the same population.26 These findings suggest that readmission reduction programs could have unintended negative consequences, emphasizing that both readmission and mortality need to be considered for any intervention. In contrast, we observed no statistically significant differences in 30-day HF or any-cause readmission or death. This consistency supports our major finding that a strategy of initial nonphysician telephone-guided follow-up with the ability to expeditiously escalate care to clinic visits as needed is comparable to uniformly scheduling initial physician clinic-guided follow-up without unintended clinical consequences.
The current standard of HF transitional care is to provide in-person physician clinic follow-up within 7 days of hospital discharge based on current guidelines.27,28 As a result, national quality metrics (eg, American Heart Association Get With The Guidelines–Heart Failure) require in-person follow-up within 7 days.29 Yet, many health systems have difficulty providing systematic early clinic follow-up among the growing HF population. In one large US study, the median percentage of HF patients who received follow-up within 7 days of hospital discharge was only 38.3%.11 The resources required to consistently provide a clinic appointment within a few days of hospital discharge are significant and may not be practically attainable in all patients and health systems, and this has been exacerbated by COVID-19.25 Our study demonstrates that an initial telephone guided follow-up strategy is feasible and effective within an integrated healthcare delivery model and can reduce clinic visits in the first 7 days after discharge.
Study Limitations and Insights on Pragmatic Trials
This clinical trial was conducted in a pragmatic manner, which used fewer resources and eased implementation within existing care pathways. We were able to allocate the initial follow-up contact strategy as a means to dictate the direction of care after the first contact. However, we could not control how follow-up was delivered after the initial contact. While such control would have allowed for a definitive prescription of exactly how care differed between the 2 groups, implementation of such control was not practical in our study, given the complexity of decision making around the intensity of care required for individual patients after the initial follow-up contact.
While some crossover occurred, it only affected 2% of patients assigned to initial physician clinic-guided follow-up and 9% of patients assigned to initial nonphysician telephone-guided follow-up. In unadjusted and multivariable analyses, there were no statistically significant differences in the prespecified outcomes between groups. The secondary outcome of 30-day death is a competing risk for the primary outcome of 30-day all-cause readmission, as those who died could not be readmitted. However, because mortality was similar between groups, formally accounting for death as a competing risk would not materially change the results. We also acknowledge that for the primary outcome of HF readmission, the event rate in our control group was lower than originally projected, and we also did not reach our original target sample size due to funding limitations, which both contributed to reduced power. Our intervention was also implemented within a fully integrated healthcare delivery system supported by a comprehensive EHR system, which may affect generalizability of our results to other types of health systems that are less integrated or to uninsured patients relying on safety net systems.
The limitations that we encountered in this study highlight some of the challenges in conducting a pragmatic clinical trial in a healthcare delivery system with established and robust care pathways. Before enrollment started, national professional society guidelines already strongly promoted in-person clinic follow-up appointments within 7 days postdischarge as an important quality metric.29 Thus, despite the lack of definitive clinical trial evidence for this recommendation, this led to some initial skepticism by providers to support randomization of patients to an initial nonphysician telephone-guided follow-up strategy instead of automatically scheduling an initial in-person physician clinic appointment, which contributed to recruitment challenges. A pragmatic clinical trial often requires providers to deviate to a certain degree from their usual practice. Strategies to increase adherence to the study protocol would include: selecting interventions in areas where providers do not have strong preferences regarding treatment selection, spending more resources before study initiation to ensure more widespread provider support for the study intervention at all participating centers, and increasing the scale of patient enrollment to shorten the length of the enrollment period given that clinical care may further evolve over time. In a pragmatic trial of different care delivery strategies, variation in aspects of care between the study arms are often difficult to predict, given that it can be impractical to control for the many components of care within a complex population such as HF. Given this unpredictability, it is helpful to prospectively collect a broad range of data to characterize how patients may be treated differently in each group despite randomization. If all components of an intervention cannot be strictly controlled, then understanding the variation in how the intervention was implemented in each center can assist with study interpretation.
Conclusions
Early follow-up after hospitalization for HF guided by an initial nonphysician, structured telephone visit reduces the number of early in-person clinic visits required without any significant difference in 30-day readmission or mortality compared with follow-up guided by an initial physician clinic visit. Our study results were achieved with telephone care providers trained in HF care supported by physician supervision, accessing a comprehensive EHR integrated across all practice settings, and having the ability to change cardiac medications and arrange future clinic visits as needed. Our findings suggest that there may be an opportunity to improve the efficiency of HF transitional care by using adequately trained and resourced nonphysician telephone providers supported within an integrated health care framework.
Acknowledgments
We would like to thank all the heart failure care managers, cardiologists, internal medicine and family medicine physicians, hospital-based specialists, and support staff that provide high-quality care for our patients with heart failure, as well as Farzien Khoshniat-Rad, Ellen Boyer, and Sophia Walia who provided critical technical support for the study.
Sources of Funding
The study was funded by The Permanente Medical Group (TPMG) Delivery Science Research Program. The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the article; and decision to submit the article for publication.
Footnotes
References
- 1.
Mozaffarian D, Benjamin EJ, Go AS, Arnett DK, Blaha MJ, Cushman M, Das SR, de Ferranti S, Després JP, Fullerton HJ, Howard VJ, Huffman MD, Isasi CR, Jiménez MC, Judd SE, Kissela BM, Lichtman JH, Lisabeth LD, Liu S, Mackey RH, Magid DJ, McGuire DK, Mohler ER, Moy CS, Muntner P, Mussolino ME, Nasir K, Neumar RW, Nichol G, Palaniappan L, Pandey DK, Reeves MJ, Rodriguez CJ, Rosamond W, Sorlie PD, Stein J, Towfighi A, Turan TN, Virani SS, Woo D, Yeh RW, Turner MB ; Writing Group Members; American Heart Association Statistics Committee; Stroke Statistics Subcommittee. Heart disease and stroke statistics-2016 update: a report from the American Heart Association.Circulation. 2016; 133:e38–360. doi: 10.1161/CIR.0000000000000350LinkGoogle Scholar - 2.
McDermott K, Elixhauser A, Sun R Trends in Hospital Inpatient Stays in the United States, 2005–2014, Agency for Healthcare Research and Quality, 2017. Statistical Brief #225 https://www.hcup-us.ahrq.gov/reports/statbriefs/sb225-Inpatient-US-Stays-Trends.pdf.Google Scholar - 3.
Jencks SF, Williams MV, Coleman EA . Rehospitalizations among patients in the Medicare fee-for-service program.N Engl J Med. 2009; 360:1418–1428. doi: 10.1056/NEJMsa0803563CrossrefMedlineGoogle Scholar - 4.
Dharmarajan K, Hsieh AF, Lin Z, Bueno H, Ross JS, Horwitz LI, Barreto-Filho JA, Kim N, Bernheim SM, Suter LG, Drye EE, Krumholz HM . Diagnoses and timing of 30-day readmissions after hospitalization for heart failure, acute myocardial infarction, or pneumonia.JAMA. 2013; 309:355–363. doi: 10.1001/jama.2012.216476CrossrefMedlineGoogle Scholar - 5.
Fingar K, Washington R . Trends in Hospital Readmissions for Four High-Volume Conditions, 2009–2013: Statistical Brief #196. Healthcare Cost and Utilization Project (HCUP) Statistical Briefs [Internet], Rockville, MD: Agency for Healthcare Research and Quality (US), 2015. 2006https://www.hcup-us.ahrq.gov/reports/statbriefs/sb196-Readmissions-Trends-High-Volume-Conditions.jsp.Google Scholar - 6. Patient Protection and Affordable Care Act PLN-, §2702, 124 and Stat. 119. 111th Congress, 2nd Session, 2010. U.S. Government Publishing Office.Google Scholar
- 7.
Smith K, Fleming JP, Gros B . Editorial: transitional care clinics to reduce 30-day readmissions in heart failure patients.Cureus. 2018; 10:e2069. doi: 10.7759/cureus.2069MedlineGoogle Scholar - 8.
Feltner C, Jones CD, Cené CW, Zheng ZJ, Sueta CA, Coker-Schwimmer EJ, Arvanitis M, Lohr KN, Middleton JC, Jonas DE . Transitional care interventions to prevent readmissions for persons with heart failure: a systematic review and meta-analysis.Ann Intern Med. 2014; 160:774–784. doi: 10.7326/M14-0083CrossrefMedlineGoogle Scholar - 9.
Dhalla IA, O’Brien T, Morra D, Thorpe KE, Wong BM, Mehta R, Frost DW, Abrams H, Ko F, Van Rooyen P, Bell CM, Gruneir A, Lewis GH, Daub S, Anderson GM, Hawker GA, Rochon PA, Laupacis A . Effect of a postdischarge virtual ward on readmission or death for high-risk patients: a randomized clinical trial.JAMA. 2014; 312:1305–1312. doi: 10.1001/jama.2014.11492CrossrefMedlineGoogle Scholar - 10.
Hansen LO, Young RS, Hinami K, Leung A, Williams MV . Interventions to reduce 30-day rehospitalization: a systematic review.Ann Intern Med. 2011; 155:520–528. doi: 10.7326/0003-4819-155-8-201110180-00008CrossrefMedlineGoogle Scholar - 11.
Hernandez AF, Greiner MA, Fonarow GC, Hammill BG, Heidenreich PA, Yancy CW, Peterson ED, Curtis LH . Relationship between early physician follow-up and 30-day readmission among Medicare beneficiaries hospitalized for heart failure.JAMA. 2010; 303:1716–1722. doi: 10.1001/jama.2010.533CrossrefMedlineGoogle Scholar - 12.
Leppin AL, Gionfriddo MR, Kessler M, Brito JP, Mair FS, Gallacher K, Wang Z, Erwin PJ, Sylvester T, Boehmer K, Ting HH, Murad MH, Shippee ND, Montori VM . Preventing 30-day hospital readmissions: a systematic review and meta-analysis of randomized trials.JAMA Intern Med. 2014; 174:1095–1107. doi: 10.1001/jamainternmed.2014.1608CrossrefMedlineGoogle Scholar - 13.
Van Spall HGC, Lee SF, Xie F, Oz UE, Perez R, Mitoff PR, Maingi M, Tjandrawidjaja MC, Heffernan M, Zia MI, Porepa L, Panju M, Thabane L, Graham ID, Haynes RB, Haughton D, Simek KD, Ko DT, Connolly SJ . Effect of patient-centered transitional care services on clinical outcomes in patients hospitalized for heart failure: the PACT-HF randomized clinical trial.JAMA. 2019; 321:753–761. doi: 10.1001/jama.2019.0710CrossrefMedlineGoogle Scholar - 14.
McKay C, Park C, Chang J, Brackbill M, Choi JY, Lee JH, Kim SH . Systematic review and meta-analysis of pharmacist-led transitions of care services on the 30-day all-cause readmission rate of patients with congestive heart failure.Clin Drug Investig. 2019; 39:703–712. doi: 10.1007/s40261-019-00797-2CrossrefMedlineGoogle Scholar - 15.
Lee KK, Yang J, Hernandez AF, Steimle AE, Go AS . Post-discharge follow-up characteristics associated with 30-day readmission after Heart Failure Hospitalization.Med Care. 2016; 54:365–372. doi: 10.1097/MLR.0000000000000492CrossrefMedlineGoogle Scholar - 16.
Reza N, DeFilippis EM, Jessup M . Secondary impact of the COVID-19 pandemic on patients with heart failure.Circ Heart Fail. 2020; 13:e007219. doi: 10.1161/CIRCHEARTFAILURE.120.007219LinkGoogle Scholar - 17.
Gordon NP Sociodemographic and Health-Related Characteristics of Members in Kaiser Permanente’s Northern California Region, 2014/2015. Internal report, Oakland, CA: Division of Research, Kaiser Permanente Medical Care Program, 2017Google Scholar - 18.
McKee PA, Castelli WP, McNamara PM, Kannel WB . The natural history of congestive heart failure: the Framingham study.N Engl J Med. 1971; 285:1441–1446. doi: 10.1056/NEJM197112232852601CrossrefMedlineGoogle Scholar - 19.
Arellano MG, Petersen GR, Petitti DB, Smith RE . The California automated mortality linkage system (CAMLIS).Am J Public Health. 1984; 74:1324–1330. doi: 10.2105/ajph.74.12.1324CrossrefMedlineGoogle Scholar - 20.
Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY . Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization.N Engl J Med. 2004; 351:1296–1305. doi: 10.1056/NEJMoa041031CrossrefMedlineGoogle Scholar - 21.
Go AS, Iribarren C, Chandra M, Lathon PV, Fortmann SP, Quertermous T, Hlatky MA ; Atherosclerotic Disease, Vascular Function and Genetic Epidemiology (ADVANCE) Study. Statin and beta-blocker therapy and the initial presentation of coronary heart disease.Ann Intern Med. 2006; 144:229–238. doi: 10.7326/0003-4819-144-4-200602210-00004CrossrefMedlineGoogle Scholar - 22.
Go AS, Lee WY, Yang J, Lo JC, Gurwitz JH . Statin therapy and risks for death and hospitalization in chronic heart failure.JAMA. 2006; 296:2105–2111. doi: 10.1001/jama.296.17.2105CrossrefMedlineGoogle Scholar - 23.
Goldberg RJ, Gurwitz JH, Saczynski JS, Hsu G, McManus DD, Magid DJ, Smith DH, Go AS ; CVRN PRESERVE HF Investigators. Comparison of medication practices in patients with heart failure and preserved versus those with reduced ejection fraction (from the Cardiovascular Research Network [CVRN]).Am J Cardiol. 2013; 111:1324–1329. doi: 10.1016/j.amjcard.2013.01.276CrossrefMedlineGoogle Scholar - 24.
Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J ; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate.Ann Intern Med. 2009; 150:604–612. doi: 10.7326/0003-4819-150-9-200905050-00006CrossrefMedlineGoogle Scholar - 25.
Peterson PN, Rumsfeld JS, Liang L, Albert NM, Hernandez AF, Peterson ED, Fonarow GC, Masoudi FA ; American Heart Association Get With the Guidelines-Heart Failure Program. A validated risk score for in-hospital mortality in patients with heart failure from the American Heart Association get with the guidelines program.Circ Cardiovasc Qual Outcomes. 2010; 3:25–32. doi: 10.1161/CIRCOUTCOMES.109.854877LinkGoogle Scholar - 26.
Gupta A, Allen LA, Bhatt DL, Cox M, DeVore AD, Heidenreich PA, Hernandez AF, Peterson ED, Matsouaka RA, Yancy CW, Fonarow GC . Association of the hospital readmissions reduction program implementation with readmission and mortality outcomes in Heart Failure.JAMA Cardiol. 2018; 3:44–53. doi: 10.1001/jamacardio.2017.4265CrossrefMedlineGoogle Scholar - 27.
Ponikowski P, Voors AA, Anker SD, Bueno H, Cleland JGF, Coats AJS, Falk V, González-Juanatey JR, Harjola VP, Jankowska EA, Jessup M, Linde C, Nihoyannopoulos P, Parissis JT, Pieske B, Riley JP, Rosano GMC, Ruilope LM, Ruschitzka F, Rutten FH, van der Meer P . 2016 ESC guidelines for the diagnosis and treatment of acute and Chronic Heart Failure.Rev Esp Cardiol (Engl Ed). 2016; 69:1167. doi: 10.1016/j.rec.2016.11.005MedlineGoogle Scholar - 28.
Yancy CW, Jessup M, Bozkurt B, Butler J, Casey DE, Drazner MH, Fonarow GC, Geraci SA, Horwich T, Januzzi JL, Johnson MR, Kasper EK, Levy WC, Masoudi FA, McBride PE, McMurray JJ, Mitchell JE, Peterson PN, Riegel B, Sam F, Stevenson LW, Tang WH, Tsai EJ, Wilkoff BL ; American College of Cardiology Foundation; American Heart Association Task Force on Practice Guidelines. 2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines.J Am Coll Cardiol. 2013; 62:e147–e239. doi: 10.1016/j.jacc.2013.05.019CrossrefMedlineGoogle Scholar - 29.
Patel DB, Shah RM, Bhatt DL, Liang L, Schulte PJ, DeVore AD, Hernandez AF, Heidenreich PA, Yancy CW, Fonarow GC . Guideline-appropriate care and in-hospital outcomes in patients with heart failure in teaching and nonteaching hospitals: findings from get with the guidelines-heart failure.Circ Cardiovasc Qual Outcomes. 2016; 9:757–766. doi: 10.1161/CIRCOUTCOMES.115.002542LinkGoogle Scholar