Implementing heart failure disease management in primary care: a mixed-methods analysis of the IMPACT-B study

Abstract

Objectives: Heart failure is an important health problem and patients are generally older with several comorbidities. Multidisciplinary heart failure care is therefore recommended. However, there is little evidence in real-world settings on how to involve primary care health professionals and how to evaluate such programmes. The main objective of this study is to integrate and evaluate several disease management interventions in a primary care setting.

Design: Prospective, non-randomised, observational implementation study with a mixed-methods process evaluation conducted over 3 years (2020-2022).

Setting: Primary care practices and two regional hospitals (one tertiary, one secondary) in the Leuven region, Belgium, serving approximately 100 000 inhabitants.

Participants: 100 general practitioners (GPs) from 19 practices participated. A total of 96 patients were included in the disease management programme. Inclusion criteria for patients included high-risk status for heart failure (HF) readmission, based on clinical criteria. Exclusion criteria were not explicitly defined but participation required informed consent.

Interventions: Four interventions were implemented: (1) online HF education for GPs, (2) reimbursed natriuretic peptide (NP) testing, (3) patient education by trained primary care HF educators and (4) a structured transitional care protocol posthospital discharge.

Primary and secondary outcome measures: Primary outcomes included GP self-efficacy in HF management, NP testing rates, HF registration in electronic health records and patient self-efficacy (9-item European Heart Failure Self-Care Behaviour Scale (EHFScB-9)). Secondary outcomes included patient quality of life (Short Form-12 questionnaire (SF-12)), hospital readmission rates and provider satisfaction.

Results: GPs felt more competent in the management of HF after an online education (eight point increase in self-efficacy score after 6 months follow-up, (CI 2.9 to 13, p<0.001)). GPs conducted 314 reimbursed NP tests, half of which scored above age-specific thresholds for referral. After initiation, there was a significant increase in NP testing (Bayes Factor 64), with a total rate of 5.4 tests per 1000 patient years. The proportion of registered HF in patients aged 40 years or older in GPs' electronic health records increased from 2.27% to 2.57% between 2020 and 2022. 96 patients were included in the disease management programme, 75% of these inclusions followed the transitional protocol after discharge for HF. There was significant improvement in patients' self-efficacy after 6-month follow-up (an EHFScB-9 score change of 2.7 points, 95% CI 0.62 to 4.8), but no significant change in quality of life (an SF-12 score change of 0.15 points, 95% CI -1.1 to 1.4). Adherence to the transitional protocol depended on the presence of a specialist HF nurse during admission, 84% of all patients were seen within 14 days after discharge by their general practitioner. Time-trend analysis revealed an increase of HF as a secondary rather than primary cause of admission. Participating healthcare professionals reported satisfaction with the programme.

Conclusions: The IMPACT-B study demonstrated that an integrated disease management programme for HF could be implemented and assessed in routine clinical practice. The programme resulted in increased awareness and registration of HF in primary care, increased self-management of patients and improved follow-up after discharge, although these results should be interpreted cautiously given the uncontrolled pre-post study design.

Trial registration: Trial registration NCT04334447 (clinicaltrials.gov).

Keywords: Aged; Heart failure; Organisation of health services; Primary Health Care; Self-Management.

Figure 1. Timeline of the different stages in the IMPACT-B study implementation. EHR, electronic health record.

Figure 2. Longitudinal evolution on three different domains of perceived self-efficacy for general practitioners participating in the IMPACT-B study, colour-coded according to time of response.

Figure 3. Evolution in yearly number of natriuretic peptide tests requested in MCH Leuven (our clinical laboratory partner). Pictured in (A) are the total number of tests (dark blue) and (for 2020–2022) the number of reimbursed study tests (light blue). A formal Bayesian changepoint analysis of this time trend is visualised in (B). The grey lines indicate a sampling distribution of priors within a SD of 10 from the estimated changepoint. The blue lines indicate the changepoint probability.

Figure 4. Evolution in prevalence of registered heart failure in general practitioners' electronic health records of patients aged 40 years or older throughout the study period. HF, heart failure.

Figure 5. Evolution in patient-related outcome measures for patients who had an education by the heart failure educator. Panel A demonstrates the evolution in self-management scores as measured by the EHFScB-9 Scale. Panel B demonstrates the evolution of quality of life as measured by the SF-12 questionnaire. EHFScB-9, 9-item European Heart Failure Self-Care Behaviour; SF-12, Short Form-12 Questionnaire.

Figure 6. Evolution in yearly hospital admissions in Leuven and the larger region of Vlaams-Brabant with heart failure (HF) as a primary or secondary cause of admission between 2010 and 2022 (log-transformed scale). The evolution of admissions with chronic obstructive pulmonary disease (COPD), as primary or secondary cause of admission, is plotted for reference of trend evolution (grey dotted lines).