Table 4.

Included studies of systems of care for heart failure (excluding systematic reviews, meta-analyses and policy documents)

Authors, year	Sample	Study design	Intervention	Outcomes
Workforce
Zuily, 2010 [15]	3200 patients admitted to hospital with ADHF from 1997 to 2007	Pre and post-test design	A heart failure unit was implemented in 2002. All patients received an outpatient appointment with the unit within one month post-discharge. The visits included patient education, assessment with the Cardiologist and up-titration of medications. Patients were followed up monthly with six weekly education sessions.	heart failure related readmissions were reduced from 21.7 % in 2002 to 15.6 % in 2007 (p < 0.0001)
Boom, 2012 [13]	7634 patients hospitalized for ADHF who were participating in the EFFECT trial	Retrospective cohort study	Patients were divided as to whether they received cardiologist, general practitioner, or general practitioner with cardiology consultation	Patients treated by general practitioners alone had higher risk of 30-day (OR 1.50, 95 % CI 1.18–1.91) and 1–year mortality (OR 1.29, 95 % CI 1.10–1.50)
NICOR, 2012 [12]	36 504 patients admitted to hospital with heart failure	UK national audit consisting of retrospective review of medical records	Not applicable	In-patient mortality rates: -Cardiology 8 % -Gen Med 14 % Outpatient clinics mortality -Cardiology 18 % -non Cardiology 31 % Community follow-up mortality -Heart failure nurse 22 % -non heart failure nurse 27 %
Comin-Colet, 2014 [16]	2083 patients admitted with ADHF to the hospital with a heart failure service compared to 54 659 patients admitted to hospitals with no heart failure service	Retrospective cohort study	Implemented a health service wide heart failure program encompassing an inpatient service, community service and a heart failure unit including a multidisciplinary specialist heart failure team.	Patients admitted to the hospital with the heart failure service had a lower risk of death (hazard ratio 0.92, 95 % confidence interval, 0.86–0.97), 29 % less likely to experience a readmission for any cause (95 % confidence interval, 0.66–0.76), and 14 % lower risk of heart failure readmissions (95 % confidence interval, 0.80–0.94)
Primary care
Lee, 2010 [18]	10 599 patients who presented with heart failure and were discharged from an ED in Ontario	Retrospective cohort study	Patients were divided into one of three groups: collaborative follow-up with a cardiologist and primary care physician, primary care physician only follow-up, and no follow-up	Collaborative follow up with a cardiologist and primary care physician reduced 30 day mortality compared to primary care physician only (HR 0.79; 95 % CI 0.63 to 1.00).
Rosstad, 2013 [17]	19 clinicians participated in focus groups.	Qualitative study, focus interviews	27 clinicians were identified as clinical champions to facilitate the implementation of clinical pathways. Focus groups were conducted to discuss the implementation of the care pathway.	A disease-orientated care pathway was not sustainable or appropriate to use in primary care.
In-hospital studies
Williams, 2010 [21]	Patients admitted to hospital with HFrEF. 50 patients were allocated to the historical group and 47 to the transitional care group	Quasi-experimental design. All patients were followed up for 18 weeks.	In-hospital education and follow-up arrangements either an appointment at the nurse-led clinic or home visits by the community heart failure nurse.	30 day readmissions were lower in the transitional care group at 8 % vs 14 % in the historical group.
Tuso, 2014 [22]	2076 hospital readmissions within 30 days post discharge from a hospitalisation for heart failure	Prospective cohort study	Implemented a heart failure “bundle” that included inpatient heart failure education, a home visit within 48 h of discharge, a follow-up appointment with a physician and a follow-up phone call from a heart failure care manager within 7 days of discharge.	Readmissions rates were reduced from 19 to 15 % over 30-days (p = 0.03).
In-hospital clinical audits/registries/quality improvement initiatives
Boutwell, 2011 [33] (State Action on Avoidable Rehospitalizations initiative- STAAR program)	148 hospitals throughout the US	Quality improvement program.	Hospitals work in partnership with providers and community services that the hospital frequently uses to collaborate in improving communication and coordination during transition from the hospital to the next setting of care.	No results published to date
Heidenreich 2012 [26] (GWTG-HF program)	over 5 million patients and over 2093 US hospitals participate at least once in GWTG throughout the US	Quality improvement program.	Registered hospitals receive access to GWTG toolkit specific for heart failure. The toolkit comprises of: initiation of evidence-based medications, implantation of appropriate device therapies, discharge education about heart failure, evaluation of left ventricular function, and post discharge follow-up appointment. Data is then entered into a web based system and each participating hospital receives site level patient data to identify those at risk of readmission.	Hospitals participating in GWTG-HF had significantly higher documentation of the left ventricular ejection fraction (93.4 % versus non-participating hospitals (89 %), use of ACEI or angiotensin receptor antagonist (88.3 % versus 86.6 %), and discharge instructions (74.9 % versus 70.5 %). After discharge, all-cause readmission at 30 days was 24.5 % and mortality at 30 days after admission was 11.1 %. 30-day readmission was lower for GWTG hospitals (−0.33 %; 95 % CI, −0.53 % to −0.12 %).
Hansen, 2013 [32] BOOST program	11 hospitals throughout the US. In Feb 2014, 180 hospitals were participating. Patient numbers are not mentioned	Quality improvement program. Pre and post implementation design.	The BOOST intervention consisted of a toolkit which contained: an implementation guide, project management tools, such as the Teach Back Training Curriculum, and PICO guidelines to evaluate the intervention; face‐to‐face training and 12 months of expert mentoring and coaching and assistance to build a culture that supports organisational change to reduce hospital readmissions, also linking with other participating sites and data management. Each site also received project benchmark data and site level patient data.	The average rate of 30-day rehospitalization prior to implementation was 15 % and 13 % 12 months later. This was an absolute reduction of 2 % and a relative reduction of 14 %.
H2H National Quality Improvement Initiative, 2015 [34] (H2H program)	No data available	Quality improvement program.	The Hospital to Home (H2H) Initiative provides a toolkit to clinicians assist them in implementing evidence-based care from clinical guidelines, within their organisation.	No data available
Transitional care
Jaarsma, 2008 [44]	1023 patients were enrolled post hospitalisation for ADHF. Patients were assigned to 1 of 3 groups: a usual care group, a HF nurse follow-up post-discharge and intensive support by a HF nurse.	RCT	The usual care group consisted of follow-up with a Cardiologist within two months post-discharge and then six monthly. Patients in group 2 had nine clinic visits with a HF nurse post-discharge, in addition to the usual care visits. Education about HF and self-management strategies were provided during the HF nurse clinic visits. Group 3 received the same visits as Group 2 and then also received one home visit and weekly telephone contact during the first month post-discharge. After the first month, they also received two additional home visits and two visits with the multidisciplinary team. All patients were follow up for 18 months.	Neither moderate nor intensive follow up by a HF nurse reduced the combined end points of HF death and hospitalization compared with usual care. At 18 months, 411 patients (40 %) were readmitted because of HF or died from any cause: 42 % in the control group, and 41 % and 38 % in the basic and intensive support groups, respectively (hazard ratio, 0.96 and 0.93, respectively; P = .73 and P = .52, respectively). All-cause mortality occurred in 29 % of patients in the control group, and there was a trend toward lower mortality in the intervention groups combined (P = .18).
Driscoll, 2011 [45]	Thirty-three community-based heart failure program coordinators recruited 484 patients diagnosed with systolic dysfunction and >1 earlier hospitalization for ADHF	Cohort study. All patients were followed up for six months.	Patient outcomes in programs with nurse-led titration (NLT) of beta-blockers were compared with those in programs that did not allow such titration.	At 6 months, 47 % of patients participating in UC programs had no change in dosage from baseline to 6 months, compared with 39 % of patients participating in NLT programs. Patients in NLT programs were also more likely to be prescribed at target dose (48 % NLT vs 36 % UC). The composite of all-cause hospitalizations and mortality was lower in patients participating in programs allowing NLT (HR 0.58, 95 % CI 0.42–0.81).
Outpatient clinics
Fonarow, 2011 [51] IMPROVE program	15,381 patients with HFrEF from 167 US outpatient cardiology practices	Prospective clinical registry	No invention as it was a clinical registry	Adherence to a range of guideline-recommended heart failure therapies ranging from 30–80 %. An increase in adherence in performance measures was significantly associated with improved survival
Driscoll, 2014 [46]	13 patients diagnosed with HFrEF were randomised to usual care and 12 to the NLT clinic	RCT	Patients were randomised to optimisation of BB in a nurse-led titration (NLT) clinic, led by a nurse specialist with the support of a cardiologist in a heart failure clinic, or by their primary care physician	The time to maximum dose was shorter in the NLT group compared to the UC group (90 ± 14 vs 166 ± 8 days, p < 0.0005). At six months, in the NLT group there were 82 % on high dose and 9 % on low dose beta-adrenergic receptor blocker compared to the UC group with 42 % patients reaching maximum dose and 42 % patients on low dose.
Fenner, 2014 [50]	Patients admitted to hospital with ADHF.	No data was available.	Patients were seen in hospital and given education by a heart failure nurse. Patients had an appointment scheduled within three days post-discharge to attend the Heart Success Transition Clinic (HSTC). They were seen in the clinic for 4–6 weeks and then referred back to primary care. A telemedicine clinic was also available for patients living in rural and remote areas.	HSTC found a reduction in heart failure readmission rate from 17.92 % to 13.49 %. The telemedicine clinic reduced heart failure readmissions from 18 % to 10 % over a six month period
Telemonitoring programs
Piette, 2008 [60]	52 heart failure patients and their carers	Prospective cohort study	Telemonitoring using informal carers. The CarePartner Program included an automated telephonic heart failure assessment and behaviour change service. Patients received weekly calls from the system and reported information about their health and self-care using their touchtone telephone. Care nurse managers were notified when a patient reported an urgent medical condition.	75 % of patients had made changes in their self-care as a result of the intervention.
Schwarz, 2008 [59]	102 patients and their carers post discharge from hospital with ADHF	RCT 90 day follow-up	Participants were interviewed within 10 days post discharge and 90 days later. The patient recorded their weight and vital signs daily via the telemonitoring system and responded to questions about symptoms. The data from the telemonitoring system was monitored daily by a heart failure nurse. Usual care consisted of follow-up with their primary physician or cardiologist.	No significant difference in hospital readmissions or mortality
Woodend, 2008 [57]	121 patients admitted with HF and 128 patients admitted with angina	RCT 90 day follow-up	The telemonitoring system consisted of video conferencing and phone transmission of weight, vital signs and ECG. The patient was required to record their weight and vital signs via the telemonitoring system daily and weekly video conferencing with the heart failure nurse. Usual care consisted of follow-up with their primary physician or cardiologist.	No significant difference in hospital readmissions or mortality for heart failure patients
Chaudhry, 2010 [55]	1653 patients admitted with ADHF	RCT	The telemonitoring group was instructed to make daily, toll-free calls to the system. During each call, patients heard a series of questions about general health and heart-failure symptoms. The protocol required the sites to contact any patient whose response generated an alert.	No significant difference in hospital readmissions or mortality
Koehler, 2011 [52]	710 chronic heart failure patients	RCT	The telemonitoring program consisted of: portable devices for ECG, blood pressure, and body weight measurements. Patients were required to undertake daily self-assessments and the data was transferred to the telemonitoring data centre. Usual care consisted of follow-up with their primary physician or cardiologist.	No difference in mortality or heart failure hospitalisations between groups.
Angermann, 2012 [39]	715 patients hospitalised with systolic heart failure	RCT All patients were followed up for 180 days.	The telemonitoring intervention consisted of: inhospital visit, structured telephone follow-up addressing heart failure symptoms, medications, health systems utilisation and psychological well-being; titration of heart failure medication, and increased access to specialist care. Contact was weekly for one month and then individualised based on NYHA class. Usual care consisted of follow-up with their primary physician or cardiologist.	No difference in mortality or hospitalisations between groups.
Dendale, 2012 [54]	160 patients hospitalised with ADHF from seven hospitals	RCT Six month follow up	Patients were asked to transmit their weigh and vital signs daily via the telemonitoring system. When these measurements exceeded preset limits for two consecutive days an automatic email alert was sent to their primary physician and heart failure clinic. Their primary physician was to contact the patient when they received an alert and the heart failure nurse would follow up with the patient 1–3 days post alert. Usual care and telemonitoring patients were all seen in the heart failure clinic 2 weeks post-discharge. Usual care consisted of follow-up with their primary physician post-discharge.	All-cause mortality was significantly lower in the TM group as compared with the UC group (5 % vs. 17.5 %, P = 0.01). The number of heart failure readmissions per patient showed a trend (0.24 vs. 0.42 hospitalizations/patient, P = 0.06) in favour of TM.
Pekmezaris, 2012 [56]	168 patients hospitalised with a primary or secondary diagnosis of heart failure	RCT All patients were followed up for 90 days.	The telehealth intervention consisted of two video-based nursing visits (including weighs and monitoring of vital signs) and one visit with a community heart failure nurse within the first two weeks post-discharge. The frequency of the telehealth visits was determined the heart failure nurse based on patient needs and continued for 90 days post-discharge. Usual care consisted of follow-up with a community heart failure nurse.	No significant difference in hospital readmissions or mortality
Baker, 2013 [61]	3534 patients with chronic heart failure, chronic obstructive pulmonary disease, or diabetes mellitus. Intervention group (n = 1,767) and in the matched control group (n = 1,767)	Retrospective matched cohort study. Two years of follow-up	The Health Buddy Program, which integrated a content-driven telehealth system with care management.	The Health Buddy Program had 15 % lower risk-adjusted all-cause mortality (HR 0.85, 95 % CI 0.74–0.98) and reductions in inpatient readmissions during the study period that were 18 % greater than those of matched controls during this same time period (−0.035 vs −0.003; difference-in-differences = −0.032, 95 % CI = −0.054 to −0.010).
Krum, 2013 [53] CHAT study	405 patients diagnosed with heart failure. 217 patients were randomised to usual care by their primary physican and 188 to the intervention group.	Cluster deign trial with randomisation at level of General Practitioner. All patients were followed up for 12 months	The intervention comprised of ongoing support by touchtone telephone using the ‘TeleWatch’ system. Patients were required to dial in monthly to receive advice about the management of their heart failure and to complete education modules about the management of heart failure at home. Patients also had access to heart failure specialist nurse via the Telewatch system.	Fewer patients hospitalised for any cause (74 versus 114, adjusted HR 0.67, 95 % CI 0.50–0.89) and who died or were hospitalised (89 versus 124, adjusted HR 0.70 (95 % CI 0.53–0.92), in the intervention group vs usual care group, respectively.
Black, 2014 [58]	No data available	Multicentre RCT	Patients in the intervention group will receive intensive patient education using the ‘teach-back’ method and receive instruction in using the telemonitoring equipment. Following hospital discharge, they will receive a series of nine scheduled health coaching telephone calls over 6 months from nurses located in a centralized call center. The nurses also will call patients and patients’ physicians in response to alerts generated by the telemonitoring system, based on predetermined parameters.	No published results to date

ADHF acute decompensated heart failure, OR odds ratio, HR hazard ratio, CI confidence intervals, ACEI angiotensin converting enzyme inhibitors, HFrEF heart failure with reduced ejection fraction, HTM home telemonitoring, NTS nurse telephone support