Heart failure (HF) is an important health problem. Most chronic HF management occurs in primary care. Although guidelines exist, there is an important implementation gap in current HF care in Belgium. We will conduct a non-randomized, non-controlled prospective observational trial to implement guideline-recommended best clinical practices in primary care in Leuven, a region of ± 100.000 inhabitants. These best clinical practices include education of general practitioners, access to natriuretic peptides and audits in the electronic health record (EHR), training and implementation of HF educators in primary care and a structured post-discharge transition process . The main objective is to implement best clinical practices in a dynamic way and study the implementation process. We will evaluate the implementation of several guideline-recommended best clinical practices to optimize the diagnosis and treatment of heart failure in a real-world primary care setting.

The primary objective of this study is to establish a dynamic implementation environment for HF management in Leuven. It is not a primary objective to study the effect of various best clinical practices on the defined outcomes. We focus on the evolution in quality of HF management through an evaluation of the implementation process. Specific research questions are as follows:

1. Do HF education and a learning environment improve the GPs' self-efficacy?
2. Does reimbursement of the NT-proBNP test lead to implementation?
3. How do primary care professionals and patients experience the introduction of a primary care HF educator?
4. What is the adoption rate of a HF discharge checklist in hospitals' EHR? How do care professionals adhere to a checklist-supported structured discharge protocol?
5. What is the feasibility of implementing automated audits in GPs EHR?
6. Does this implementation study affect regional levels of HF hospitalization and readmissions?

Outcomes

Measurements at baseline We will register contact data (telephone number and/or mail address), and sociodemographic variables (age, sex and type of HF ) for patients who receive a HF educator contact and give informed consent. If it is an education-session post-discharge, we will ask if the patient had a contact with his/her GP.

We record patients' relevant risk criteria at discharge in the checklist.

We will record all yearly HF hospitalizations of patients in Groot-Leuven in the regional hospitals of UZ Leuven and Heilig Hart as our primary outcome for effectiveness (aggregated data). Secondary outcomes will be the number of yearly emergency room visits for HF and HF readmissions within 30 days. We will also extract these data for the previous 5 years, starting from 01.01.2014 to establish a time-series trend.

Process Evaluation

We will investigate the intervention implementation in a process evaluation. On a population level, we will look at yearly echocardiographies and cardiology consultations as a process indicator. In participating practices, we will monitor the number of registered HF diagnoses in the GPs EHR (coded or free-text). Process evaluation for specific guideline-recommended practices is as follows:

GP education GPs need to complete an online questionnaire before and after the web-based mandatory training session. The questionnaire assesses GP self-efficacy in the diagnosis and management of heart failure.

NT-proBNP We will look at the number of tests conducted per participating physician and the total number of tests performed in the largest regional lab (MCH Leuven).

HF educator We will register the number of conducted HF educator contacts and patient self-efficacy and quality of life at baseline and after 6 months as scored by the EHFScB-9 and SF-12 questionnaires respectively. We will perform a qualitative analysis of primary care professions' (GP , HF nurse, HF educator) and patients' experience of the introduction of this new role through semi-structured interviews.

Structured transitional protocol We will measure the total number of transitional contacts by proxy: the HF nurse in the hospital will fill in a standardized checklist in the EHR. Outcomes are the number of completed checklists and HF educator contacts after discharge. The HF educator verifies whether high-risk patients who gave informed consent saw their GP at the education session post-discharge.

Automated EHR audits We will record the proportion of HF patients with a registered left ventricular ejection fraction in the EHR, as well as the proportion of HFrEF patients with guideline-recommended medication (RAAS-blockade/B-blocker/MRA). We will look at the proportion of HF patients who had at least one contact with their GP in the last six months.

Data collection

At the patient level We will collect individual patient data for patients referred for a primary care HF educator consultation (by the GP or as part of the structured transmural protocol). Inclusion requires informed consent. The HF nurse will fill out a registration form with level of referral (GP or HF nurse in the hospital), contact information, sociodemographic data and an EHFScB-9 questionnaire assessing self-efficacy. We will repeat the questionnaire after 6 months via telephone or mail. The HF educator will record whether high-risk patients who gave informed consent saw their GP at the education session post-discharge.

At the care profession level We will collect the number of NT-proBNP tests at the participating lab. A yearly audit in the GPs EHR will record registered diagnoses and audit-specific outcomes.

At the population level (Groot-Leuven) Data collection for the primary outcome occurs at the hospital level through data already collected as part of the mandatory Minimale Ziekenhuisgegevens (MZG) registration for the population of Groot-Leuven (i.e. postal codes 3000, 3001, 3010, 3012, 3018, 3050). As part of the MZG registration, hospitals need to classify hospital stays according to diagnosis. Reimbursement of the hospital stay is dependent on the number of primary and secondary coded diagnoses. We define a HF hospitalization as a hospitalization with ICD-10 code I50 as the primary diagnosis. Data collection for echocardiographies and cardiology consultations occurs at the level of the Intermutual Agency (IMA). IMA is in charge of collecting administrative and medical data for all patients who have national health insurance.

For the transitional contact, we will collect the aggregated number of checklists and number of HF educator contacts after discharge. The number of completed checklists can be collected in the hospital EHR. The participating nursing organizations will collect the number of HF education sessions.

Data management

The PI and an assisting DMP coordinator will collect and enter data. All patient data is coded and pseudonymized. In pseudonymized data, we replace identifying particulars by an artificial identifier (i.e. a pseudonym). The list with pseudonyms, along with all other data is password-protected and stored in a secure encrypted cloud-drive (KU Leuven J-Drive). All other authors will have full access to the coded data (including the statistical reports and tables) in the study and will be able to take responsibility for the integrity of the data and the accuracy of the data analysis.

Data analysis

We will perform descriptive statistics regarding the baseline sociodemographic variables of patients with a HF nurse consultation. We will longitudinally plot all other outcomes measured at population level during the study. We will use chi-square tests to compare self-efficacy outcomes for GPs and HF patients who have a HF nurse contact. This is a descriptive study, hence we provide no sample size analysis. As stated previously, the goal is to study the implementation process, not the effectiveness of various interventions. Because of its descriptive nature, we plan to cautiously interpret the results.

Ethics

Before the study start, we will present the participating GPs with an informed consent form that outlines the intervention and the purpose of the study. Additionally, the HF educator will inform all the patients who receive an education session about the study and ask for informed consent. It is impossible to collect informed consents at the regional level (hospitalizations, echocardiographies, specialist consultations …), although this is not an issue since the data are aggregated and anonymous. We received approval from the University Hospitals Leuven Ethics Committee in September 2019 (S62838).

Inclusion:

- Diagnosis of HF

Exclusion

- No pre-specified exclusion criteria