Table 3.

Summary of studies evaluating strategies for the implementation of heart failure (HF) clinical guidelines

Author (year) Country	Setting	Study design	Unit of recruitment/analysis (n)	Intervention and process of implementation (when described)	Process outcomes*	Clinical outcomes*
Professional interventions
Education
Thilly et al31 (2003) France	Tertiary care; inpatient	Cluster RCT	Hospitals (20)/patients (370)	Intervention: Cardiologists presented guidelines and discussed cases with colleagues. Educational aids and guideline booklets were supplied to physicians. Control: Usual care; no implementation intervention Process: Planning/assessment— prior to developing the educational intervention, a preliminary survey was conducted to identify specific guideline deviations in practice. Guidelines determined to be of particular concern were made the focus of the intervention.	Target ACEI +27%†, P=0.003
Asch et al32 (2005) USA	Tertiary care; inpatient	Controlled before-after	Patients (489)	Intervention: Provider teams attended three training sessions where national Quality Improvement and HF experts guided them in studying, testing and implementing systematic improvements in HF care processes. Control: Usual care; no implementation intervention	ACEI +18%‡, P<0.0001; β-blockers −2%‡, P=0.49; LVEF +3%‡, P=0.49
Audit and feedback
Kasje et al22 (2006) Netherlands	Primary care	Cluster RCT	Providers (57)/patients (508)	Intervention: Providers received patient-specific feedback on a sample of patients, and attended structured meetings to discuss guidelines and current management, identify problems and propose solutions for improving HF patient care. Control: Providers received education on management of type II diabetes. Process: Planning/assessment—Optimal intervention design was determined through literature review. Specific barriers to guideline adherence were identified by physicians during peer-review meetings as part of the intervention.	ACEI +5%†, P>0.05
Frijling et al23 (2003) Netherlands	Primary care	Cluster RCT	Practices (124)/patients (236)	Intervention: Physician assistants provided physicians with a practice-specific feedback report, identified areas needing improvement and provided guidance and resources for improvement. Control: Usual care; no implementation intervention	Education OR 0.85, P=0.636
Cancian et al24 (2013) Italy	Primary care	Before-after	Patients (1905)	Intervention: Performance data were aggregated across 21 health units. Project leaders reviewed data and identified barriers to unit leaders, who conveyed the data to all physicians involved. Control: Usual care; no implementation intervention Process: Access to information, training, education—Intervention was explained to participating physicians through two health unit training meetings	ACEI +3.6%†, P=0.008; β-blockers +10.8%†, P<0.0001
Matthews et al25 (2007) USA	Tertiary care; outpatient	Before-after	Patients (265)	Intervention: Following discharge of patients from the hospital, outpatient physicians were provided quality of care reports outlining services received in hospital and areas for HF care improvement. This included instructions for medication titration and detailed HF education. Control: Usual discharge information	ACEI +6.4%, P=0.042†; β-blockers −1.1%†, P=0.73; MRA +11.1%†, P=0.26
Reminders
Ansari et al26 (2003) USA	Primary care	RCT	Patients (115)	Intervention: In addition to education on β-blocker use, physicians received a list of their patients with HF eligible for β-blockers as well as electronic alerts when accessing patients’ EMRs for the first two visits after randomisation. Control: Education on the use of β-blockers via grand rounds presentations and guideline dissemination Process: Planning/assessment—The intervention was designed to address a barrier identified at baseline.	β-blockers −17%†, P>0.05; target β-blockers −8%†, P>0.05	HF-related hospitalisations +4%†, P>0.05; 1-year all-cause mortality −12%†, P=0.05
Braun et al28 (2011) Germany	Primary care	Before-after	Patients (190)	Intervention: Computer-based system displayed a pop-up window of a condensed version of the HF guidelines during clinical consultations. Control: Usual care; no implementation intervention	ACEI −4.4%†, P=0.3; β-blockers +12.3%†, P=0.03; MRA +9.2%†, P=0.04
Butler et al27 (2006) USA	Tertiary care university hospital; inpatient	Before-after	Patients (1275)	Intervention: Computerised physician order entry system provided point-of-care reminders for select quality measures and included a prescription writer function. Control: Usual order entry form without disease-specific prompts Process: Planning/assessment—The intervention was developed iteratively prior to the intervention phase of the study. The programme was modified based on institutional requirements, developer-initiated improvements and user feedback.	ACEI +13%†, P=0.10; education +53%†, P<0.001; LVEF +5%†, P=0.86
Qian et al29 (2011) USA	Tertiary care university hospital; inpatient	Before-after	Patients (5000)	Intervention: Computer program flagged eligible patients not receiving ACEI/ARB. Pharmacists verified the flags and notified the medical team via EMR. Patients were reflagged if no action was taken within 24 hours. Control: Usual care; no implementation intervention Process: Planning— Comprehensive plan-do-study-act cycle occurred over a period of 1 year prior to the intervention phase. Problems were identified in the system’s operating process and adjusted to increase work flow efficiency.	ACEI +9.2%†, P<0.002
Gravelin et al30 (2011) USA	Cardiology clinics; outpatient	Before-after	Patients (6632)	Intervention: EMR screening tool identified patients with left ventricular ejection fraction <35% and prompted cardiologists to refer to electrophysiologist for consideration of ICD and/or CRT. Control: Usual care; no implementation intervention	ICD/CRT referral: site 1 +47%†, P<0.02; site 2 +40%†, P<0.001
Organisational interventions
Changes in medical records systems
Reingold and Kulstad37 (2007) USA	Tertiary care university hospital; inpatient	Before-after	Patients (171)	Intervention: Existing HF order sets were modified to be more succinct and visually organised, with the addition of narrative information to encourage utilisation. Control: Routine order sets Process: Planning/assessment—The improvement process was initiated 5 years in advance of intervention phase, and the intervention was developed based on staff feedback.	ACEI +58%†, P=0.008
Oujiri et al38 (2011) USA	Tertiary care university hospital	Before-after	Patients (153)	Intervention: A discharge face sheet embedded into the EMR reminded physicians of evidence-based measures and required physicians to indicate reasons for unmet measures. Control: Computerised order entry form included reminders to address each diagnosis, but no prompts to follow treatment guidelines. Discharge orders were not easily accessible within the EMR, making it difficult to assess adherence to HF quality measures. Process: Planning/assessment—The institution’s admission and discharge processes were reviewed extensively to identify barriers to guideline adherence at baseline, and these were addressed in the intervention design.	ACEI +18%†, P<0.01; education +5%†, P>0.05; LVEF +12%†, P>0.05
Baker et al36 (2011) USA	Primary care	ITS	Patients (276)	Intervention: Pre-visit paper reminders of outstanding quality deficits were printed and placed outside the patient’s examination room to supplement existing electronic reminders within the EMR. Control: Electronic system offered point-of-care reminders, captured contraindications and patient refusals, and generated lists of patients not receiving essential medications. Process: Planning/assessment—Following earlier introduction of an electronic reminder system, physician adherence to guideline recommendations was evaluated. Reasons for gaps were identified among a subset of physicians and addressed in the design of the paper intervention.	ACEI +0% per year§, P=0.95; β-blockers +2.9% per year§, P=0.004
Persell et al35 (2011) USA	Primary care	ITS	Patients (not clear)	Intervention: An existing reminder system was updated to be minimally intrusive and include standardised means to capture contraindications. Control: EMR generated interruptive ‘pop-up’ reminders at point of care, and did not possess a mechanism to record contraindications. Process: Planning/assessment—Limitations in the EMR system were identified at baseline and addressed in the system redesign.	ACEI +5.3% per year§, P<0.001; β-blockers +5.7% per year§, P<0.001
Clinical multidisciplinary team
McCarren et al41 (2013) USA	Tertiary care; outpatient	Cluster RCT	Hospitals (12)/patients (220)	Intervention: Pharmacists were asked to invent methods to improve prescribing practices. Pharmacists received data on facility guideline adherence, along with a list of patients with suboptimal HF therapy. Control: Pharmacists were asked to invent methods to improve prescribing practices. Pharmacists received data on facility guideline adherence. Process: Planning—Intervention methods were designed to be pragmatic (ie, data collection and presentation required by each pharmacist was minimal to promote participation).	Target β-blockers +1%†, P>0.05
Mejhert et al39 (2004) Sweden	Tertiary university hospital; outpatient	RCT	Patients (208)	Intervention: A nurse monitored patients after discharge and adjusted their medications under the supervision of a senior cardiologist. Control: Conventional follow-up in primary care	Target ACEI +14%†, P<0.05; ACEI −5%†, P>0.05; β-blockers −6%, P>0.05	4-year all-cause mortality +7%†, P>0.05; 4-year all-cause readmissions +0%†, P>0.05
Kasper et al40 (2002) USA	Tertiary university hospital; outpatient	RCT	Patients (200)	Intervention: In the intervention group, HF nurses closely followed up with patients after discharge and implemented the cardiologist-developed treatment algorithm. The control group received care from the primary physician alone. Control: Conventional follow-up in primary care	ACEI +12.3%†, P=0.07; β-blockers +8.1%†, P=0.27
Ansari et al26 (2003) USA	Primary care at a university hospital; outpatient	RCT	Patients (105)	Intervention: In addition to receiving education on β-blocker use, NPs, under physician supervision, were responsible for initiating, titrating and maintaining eligible patients with HF on β-blockers. Control: All providers received education on the use of β-blockers via grand rounds presentations and guideline dissemination. Process: Planning/assessment— The intervention was designed to address a barrier identified at baseline.	β-blockers +32%†, P<0.001; target β-blockers +33%†, P<0.001	HF-related hospitalisations −1%†, P=0.66 1-year all-cause mortality −5%†, P=0.05
Güder et al44 (2015) Germany	Tertiary university hospital; outpatient	RCT	Patients (390)	Intervention: HF specialist nurses closely followed up with patients after discharge and uptitrated medications under cardiologist supervision. Control: Conventional follow-up in primary care	ACEI +4.9%†, P<0.05; target ACEI +25.1%†, P<0.001; β-blockers +7.4%†, P<0.05; target β-blockers +23.9%†, P<0.001; MRA +5.7%†, P>0.05; target MRA +0.3%, P>0.05
Warden et al42 (2014) USA	Tertiary care; inpatient	Before-after	Patients (150)	Intervention: Pharmacists reviewed patients’ records, addressed prescription concerns to the primary care team and made suggestions for medication treatment and monitoring. Control: Usual care; medication reconciliation and patient management by physicians and nurses	ACEI +13%†, P=0.02; education +17%†, P=0.007	30-day HF-related readmissions −12%†, P=0.11 30-day all-cause readmissions −21%†, P=0.02
Martinez et al43 (2013) USA	HF clinic; outpatient	Before-after	Patients (144)	Intervention: Pharmacists managed a clinic in which they initiated and adjusted medication dosages based on clinical characteristics. Control: Usual care; medication titration conducted by cardiologists Process: Planning/assessment—The intervention was introduced to address previously identified gaps in HF care.	Target ACEI +21.9%†, P=0.007; target β-blockers +24.3%†, P=0.012
Crissinger et al45 (2015) USA	HF clinic; outpatient	Cohort	Patients (899)	Intervention: Nurse practitioners and pharmacists adjusted medication dosages based on clinical characteristics under HF physician supervision. Control: Patients were managed by general cardiologists.	ACEI +6%†, P>0.05; >50% target ACEI +10%†, P<0.0167; β-blockers +44%†, P<0.0167; >50% target β-blockers +43%†, P<0.0167
Clinical pathways
Panella et al46 (2005) Italy	Tertiary care; inpatient	RCT	Patients (68)	Intervention: An integrated care pathway displayed patient care goals and provided the sequence and timing of actions necessary to achieve goals. Control: Usual care; no implementation intervention Process: Information, training and education—The intervention group received training to use the pathway. Planning/assessment - There was a 6-month planning period prior to the intervention phase to build work teams, review practices, develop the pathway and perform ongoing evaluation and improvement.	ACEI +8.28%†, P>0.05; education +27.7%†, P<0.01; LVEF +35.4%†, P<0.01	30-day all-cause readmissions −4.36%†, P>0.05 30-day all-cause mortality −7.33%†, P<0.05
Garin et al47 (2012) Switzerland	Tertiary care; inpatient	Before-after	Patients (363)	Intervention: A computerised clinical pathway included order sets for each stage of the hospital stay and required specific evaluation, treatment and education criteria to be met prior to the next stage. Control: Usual care; no implementation intervention	Target ACEI +0.2%†, P=0.97; β-blockers +14.3%†, P=0.006; LVEF +16%†, P=0.002	30-day all-cause mortality −0.4%†, P=0.8; 90-day all-cause mortality −0.8%†, P=0.11; 30-day all-cause readmissions −6.6%†, P=0.11; 90-day all-cause readmissions −8.2%†, P=0.11
Whellan et al48 (2001) USA	HF clinic; outpatient	Before-after	Patients (117)	Intervention: Based on predefined protocols and severity of the patient’s illness, a follow-up schedule for clinic visits and telephone calls was initiated at the time of enrolment. Control: Usual care; no implementation intervention Process: Access to information, training and education—Pre-enrolment, internal medicine house staff and primary care physicians in the network were presented an outline of the programme; pocket cards with inclusion criteria and referral phone numbers were also provided for all nursing stations at the hospital. Planning/assessment—The programme was designed by adapting practices from other disease management programmes to the needs of the local health system.	β-blockers +24%†, P<0.01; target β-blockers +7%†, P<0.01; ACEI +1%†, P=0.75	1.5 (control) vs 0 (intervention) all-cause hospitalisations per patient-year, P<0.01
McCue et al49 (2009) USA	Tertiary care; inpatient	Cohort	Patients (6013)	Intervention: A clinical pathway comprised an order sheet, clinical outcomes monitoring checklist, explanations for nursing and disease-specific patient education forms. Control: Usual care; no implementation intervention Process of implementation: Planning/assessment—Design of the clinical pathway was dynamic; practitioner feedback was continuously sought and incorporated into pathway design throughout the intervention period.	ACEI +17.2%†, P<0.001; LVEF +10.6%†, P<0.001
Ranjan et al50 (2003) USA	Tertiary care; inpatient	Cohort	Patients (371)	Intervention: A clinical pathway for HF care was implemented. Control: Usual care; no implementation intervention	ACEI +33%†, P<0.001
Continuity of care
Hickey et al51 (2016) Australia	HF clinic; outpatient	Cohort	Patients (335)	Intervention: HF disease management clinic facilitates communication between hospital and primary care by means of a comprehensive medication titration form outlining recommended target dose of medications, the order of titration and primary clinician responsible for managing titration. Control: Discharge titration form was available, but rarely used to facilitate patient transition from hospital to community. Process: Planning/assessment—A steering committee comprising cardiologists, general practitioners, pharmacists and nurses met quarterly to refine the implementation intervention in an iterative PDSA cycle. Barriers and solutions were developed by interviewing physicians and practice managers.	Target ACEI +11%† (2010), +18%† (2011), P=0.051; target β-blockers −5%† (2010), +13%† (2011), P=0.045
Financial interventions
Provider incentives
Esse et al52 (2013) USA	Tertiary care; inpatient	Cohort	Patients (4304)	Intervention: Primary physicians responsible for patients in the Medicare Advantage Prescription Drug Plan were financially compensated for utilisation of evidence-based HF therapy. Control: Usual care; no implementation intervention	ACEI −1.85%†, P=0.244; β-blockers −0.06%†, P=0.972	All-cause hospitalisations: acute visits +2.58%†, P=0.100; ER visits +0.62%†, P=0.675
Institutional incentives
Lindenauer et al53 (2007) USA	Tertiary care; inpatient	Controlled before-after	Patients (50 678)	Intervention: Hospitals submitted data on 33 HF quality measures. Those performing in the top decile for a given year received a 2% bonus payment in addition to usual Medicare reimbursement. Control: Usual care; no implementation intervention	ACEI +2%‡, P=0.34; LVEF +5.1%‡, P<0.001
Sutton et al54 (2012) England	Tertiary care; inpatient	Controlled before-after	Patients (not clear)	Intervention: Hospitals submitted data on 28 HF quality measures. At the end of the first year, hospitals that reported quality scores in the top quartile received a 4% bonus. Control: Usual care; no implementation intervention	ACEI +1.4%‡; LVEF +8.1%‡, no P values reported; education +15.2%‡	30-day all-cause mortality −0.6%†, P=0.3
Combined interventions
Peters-Klimm et al33 (2008) Germany	Primary care	Cluster RCT	Providers (37)/patients (168)	Intervention: Physicians engaged in four didactic, interdisciplinary educational meetings with primary care physicians, cardiologists and psychosomatic specialists; and received pharmacotherapy feedback (% target dose) on individual patients. Control: Physicians received a standard lecture on guideline-recommended treatment of HF. Process: Information, training and education—Physicians received initiation visit, which included an introduction to the intervention and a handout of the trial investigator file. Opinion leaders—Education component of the intervention was provided by a senior cardiologist with didactic expertise.	ACEI +8.7%†, P=0.15; target ACEI +12.3%†, P=0.04; β-blockers −4.8%†, P=0.67; target β-blockers +1.7%†, P=0.26
Fonarow et al/Gheorghiade et al 55 56 (2010/2012) USA	Cardiology clinic; outpatient	Before-after	Patients (15 177)	Intervention: The intervention consisted of a guideline-based clinical decision support tool kit, educational materials, practice-specific data reports, benchmarked quality of care reports and structured educational opportunities. Control: Usual care; no implementation intervention Process: Information, training and education—A 1-day workshop for practice personnel provided overview of study goals and tool kit. Planning/assessment—A steering committee was appointed to follow a structured, rigorous, guideline-driven process to develop the pathways and tools prior to the intervention phase. Opinion leaders—The educational component of the intervention included expert opinions regarding best practices in HF care.	ACEI +6.7%†, P<0.001; target ACEI +1.8%, P=0.053†; β-blockers +7.4%†, P<0.001; target β-blockers +9.8%, P≤0.001; MRA +27.4%†, P<0.001; target MRA +4.1%, P=0.107; education +9.1%†, P<0.001; ICD referral +30.3%†, P<0.001
Goff et al34 (2005) USA	Primary care	Before-after	Patients (3141)	Intervention: Physicians received performance audit and feedback, aggregated across a multicounty health service area; and patient-specific chart reminders regarding medications and education. Control: Usual care; no implementation intervention Process: Planning—The intervention planning team identified and addressed barriers at provider and patient levels. Patients—The intervention planning team developed an educational brochure based on results of focus groups with patients with HF.	ACEI −2.7%†, P=0.26; β-blockers +15.2%†, P<0.0001; LVEF +4.3%†, P<0.0001
Riggio et al57 (2009) USA	Tertiary care; inpatient	Before-after	Patients (4728)	Intervention: The intervention consisted of a computerised discharge checklist with electronic prompts on medication use, LVEF assessment and discharge instructions; personalised resident performance reports; financial bonus for residents achieving a threshold of quality compliance; lectures on hospital/state/nation quality performance. Control: Usual care; no implementation intervention Process:Planning—The intervention planning team received and incorporated ongoing feedback from residents and physicians in developing the reminder system prior to the intervention phase.	ACEI +15.7%†, P<0.001; education +55.8%†, P<0.001; LVEF −0.2%†, P=0.78
Scott et al58 (2004) Australia	Mixed; tertiary and primary care practices	Before-after	Patients (904)	Intervention: The in-hospital component consisted of: reminders on patient charts; clinical pathways for emergency chest pain assessment and management; educational presentations as grand rounds, seminars, workshops and case-based meetings; briefing of hospital and primary care physicians by clinical pharmacists. The discharge planning component consisted of standardised discharge referral summaries with personal treatment targets; medication lists forwarded to community pharmacists; pharmacist counselling of patients about lifestyle changes, drug therapy and risk factor modification; postdischarge telephone follow-up by clinical pharmacists of high-risk patients. Control: Usual care; no implementation intervention Process: Planning/assessment— Intervention was designed to address several implementation barriers that were identified through literature review.	ACEI +15%†, P=0.04; β-blockers +21%†, P=0.01; LVEF +9%†, P=0.06	30-day HF-related readmissions +0.8%†, P>0.05 All-cause mortality: 30 days −2.9%†, P<0.04; 6 months −7.6%†, P<0.001; 1 year +10.4%†, P=0.005
Dykes et al59 (2005) USA	Tertiary care; inpatient	Before-after	Patients (314)	Intervention: This involved a clinical pathway in EMR; an HF self-management education tool; and ongoing performance feedback. Control: Usual care; no implementation intervention	Medication prescription +6.4%†, P=0.389; education +64.9%†, P=0.000

*Statistically significant results are shown in bold letters.

†Absolute risk difference reported as (intervention group − control group).

‡Difference in difference (controlled before/after studies) reported as [intervention group (Time 2 − Time 1) − control group (Time 2 − Time 1)].

§Difference in rate of change (ITS studies) reported as (intervention group rate of change − control group rate of change).

ACEI, ACE inhibitor; ARB, angiotensin II receptor blocker; CRT, cardiac resynchronisation therapy; EMR, electronic medical records; ER, emergency room; HF, heart failure; ICD, internal cardioverter defibrillator; ITS, interrupted time series; LVEF, left ventricular ejection fraction; MRA, mineralocorticoid receptor antagonist; NP, nurse practitioner; PDSA, plan-do-study-act cycle; RCT, randomised controlled trial.