Study setting and recruitment

The study started in October 2023 and will include all CR centres (n = 78) which report to the SWEDEHEART-CR registry. The sole exclusion criterion for CR centres is unwillingness to participate. The inclusion criteria for patients are 1) having a diagnosis of a type 1 MI (caused by atherosclerotic plaque rupture or coronary artery thrombosis) and 2) age 18–79 years at discharge from MI hospitalization. There are no exclusion criteria for patients.

Study design

The study is an open-label cluster-randomized effectiveness-implementation hybrid trial. The effectiveness-implementation hybrid design allows for testing an implementation strategy while observing the intervention´s impact on patient outcomes [20]. For the implementation support, the Consolidated Framework for Implementation Research (CFIR) model will be used to guide the design [21, 22]. Normalization Process Theory (NPT) will be used to guide the exploration of the implementation process [20].

Randomization

Centres will be randomized 1:1:1 to three clusters (A, B and C). Randomization will be stratified by geographical healthcare district (two districts in each stratum). Randomization will be performed by an independent organisation not involved in the study to avoid site-selection bias.

The interventions

First, baseline structure and process variables will be administered through the SWEDEHEART-CR registry at all CR centres. Second, two clusters will receive one of two interventions (A and B) and one cluster will receive no intervention (control cluster) (C) (Fig. 1):

A)

Reprting CR structure and process variables through the registry every six months and being offered structured implementation support (audit intervention + implementation support intervention).

B)

Reporting CR structure and process variables through the registry every six months but no structured implementation support being offered (audit intervention).

C)

No reporting of CR structure and process variables, and no structured implementation support offered (control).

Fig. 1

Overview, process and timeline of the study

The audit intervention

The audit intervention (clusters A and B) involves, on centre-level, reporting 30 variables on CR structure and processes, measuring adherence to the National Guidelines, to the SWEDEHEART registry. The complete list of CR structure and process variables is available in Additional file 1.

The implementation support intervention

For the implementation support intervention (cluster A) facilitators from the research team will provide hands-on support and guidance to the CR team as they work to implement changes [23]. Facilitators are physicians and nurses experienced in the field of cardiac rehabilitation. In the recruitment step, the centre directors or key stakeholders will be contacted by the principal investigator to offer study participation by means of an e-mail. Upon acceptance to participate, the time to start the intervention will be determined, a timeline established, and the CR centre director will be asked to allocate time in the schedule for the relevant CR staff to work on the study. The implementation support will be conducted in four steps: 1) evaluation of centre practice, 2) identification of areas in need of improvement, 3) implementation of change, and 4) follow-up (Table 1). The implementation objects are work routines or CR programme components listed in Additional file 1 identified as sub-optimally implemented at the respective CR centre. Implementation strategies (defined by results from the Expert Recommendations for Implementing Change (ERIC) project) used will be identifying and preparing champions, identify barriers and facilitators, facilitation, distribution of educational materials, conducting educational meetings, organizing clinician implementation team meetings, and, providing local and technical assistance [23,24,25,26,27,28].

Table 1 The steps of the implementation support intervention

The implementation support will be provided during a two-day visit from the facilitators, two follow-up phone calls during the first month after the visit, up to three educational meetings (number as requested by centre personnel), and a digital follow-up meeting 4 months after the initial visit. Additional phone or video calls, and emails will be provided as needed.

Primary effectiveness outcome

The primary endpoint is an “adherence score”, measuring adherence to the National Guidelines on Secondary Prevention [15]. The adherence score is partly derived from the 30 variables capturing guideline-directed CR structure and processes incorporated into the SWEDEHEART-CR registry in October 2023. Permissible values are yes/partly/no/unknown. Additionally, the adherence score will include nine process variables that measure a) patient attendance in various CR components (number of patients attending a CR programme component divided by the number of patients eligible for participation), and b) time between MI hospitalization discharge and start of different CR programme components, already audited in SWEDEHEART:

a)

Attendance in CR components

1.

Proportion of patients attending an initial CR assessment (nurse visit).

2.

Proportion of patients attending an individual visit to a physiotherapist after discharge before starting an exercise-based CR (EBCR) programme (a pre-exercise screening visit).

3.

Proportion of patients completing a 3-month EBCR programme.

4.

Proportion of patients attending an individual close-out visit to a physiotherapist after completing an EBCR programme (a post-exercise assessment visit).

5.

Proportion of patients attending a patient education programme.

6.

Proportion of patients attending a close-out CR visit with a nurse at one-year after MI.

b)

Time (days) to start of different components of CR

7.

Time from hospital discharge to initial CR assessment (nurse visit).

8.

Time from hospital discharge to the pre-exercise screening visit to a physiotherapist.

9.

Time from pre-exercise screening visit to start of EBCR programme.

The responses will be pooled at each site into an adherence score for each CR centre. The contribution of each variable to the score will be weighed depending on the importance of each measurement, based on level of recommendation in European guidelines [1].

Secondary effectiveness outcomes

Table 2 displays secondary outcomes, encompassing both short- and long-term patient outcomes and implementation outcomes. Implementation outcomes will be based on responses gathered through the customized Normalisation Measure Development (NoMAD) questionnaire [29] and insights obtained from focus group interviews, as well as a cost-effectiveness analysis.

Table 2 Secondary outcomesSample size calculations

As the number of CR centres is fixed, sample size calculations have been performed to estimate the difference in mean adherence score the study will be able to reveal. For this purpose, in a feasibility analysis the CR structure and process variables were collected from 8 CR centres of different sizes and geographical locations. The mean adherence score was 30.8 (standard deviation [± 2.4]) out of a maximum available score of 39. For the audit intervention, given the following presumptions:

The study will have the power to identify a difference of ± 2.0 in adherence score between centres in cluster B (randomized to having variables on CR structure and processes incorporated in the SWEDEHEART registry) and cluster C (no new variables incorporated).

For the implementation support intervention, only centres in the lower 2 tertiles of the adherence score at baseline (approximately 17 centres) will be offered implementation support. Given the following presumptions:

The study will have the power to identify a difference of ± 2.5 in adherence score between centres in cluster A (randomized to receiving implementation support) and cluster B (no implementation support).

The timeline

The audit intervention started in October 2023 and will continue for 3 years. An interim analysis will be conducted two years after the start of the intervention. If the interim analysis shows the primary endpoint to be met (a difference of at least ± 2.0 in adherence score) the intervention will be terminated. Otherwise, the intervention will be continued until October 2026 and thereby uphold.

The implementation support intervention will start Q2 2024. Centres randomized to the implementation intervention will receive implementation support consecutively over a period of 18 months. The order in which centres will receive implementation support will depend on the centres’ possibilities and the research team´s capacity. The study outline is displayed in Table 3.

Table 3 Flow chart for the studyData analysis

In the primary outcome analysis of the audit intervention, CR centres in clusters B and C will be compared. All CR centres that have responded to the CR structure and process variables will be included in the analysis. For the primary outcome analysis of the implementation support intervention CR centres in clusters A and B will be compared. All CR centres that have i) provided answers to the structure and process variables and that ii) have scores in the lower two tertiles of the adherence score at baseline will be compared. Our analyses will assume intention-to-treat principles by treating intervention assignment as randomized regardless of whether the intervention had uptake within the practice. If not all centres in cluster A accept implementation assistance, a per-protocol analysis will be performed including only those centres in study group A that accept implementation assistance.

For the secondary analyses on patient outcomes, all patients that attended at least two follow-up visits within CR will be included. For the outcome analysis of the audit intervention patients followed at CR centres in clusters B and C will be compared. For the outcome analysis of the implementation support intervention patients followed at CR centres in clusters A and B will be compared. A per-protocol analysis will be performed including only patients belonging to CR centres in cluster A that accept implementation assistance.

Quantitative data

For baseline characteristics descriptive statistics will be used (means +/-standard deviation, medians [quartile 1, quartile 3], proportions [%] and ranges). For the primary outcome analysis, given randomized treatment assignment, total adherence scores at end of follow-up as well as change in scores between baseline and follow-up will be compared using linear regression analysis. In the case of unequal randomization concerning CR centre size (small < 75 patients/year, medium 75–150 patients/year or large > 150 patients per year), geography (6 geographical districts in Sweden, two in each stratum) or the centres belonging to a university hospital (yes/no), adjusted multivariable analysis will be performed. Results will be reported as relative treatment effects (odds ratios) with 95% confidence intervals. For secondary outcome analysis on short-term patient outcomes, the same statistical methods will be used, applying linear (continuous) or logistic (binary) regression analyses. For long-term outcomes (major adverse cardiovascular events [MACE] and total mortality) Cox proportional hazards regression models will be performed, reporting hazard ratios with 95% confidence intervals. In case of missing data, imputation will be considered. For all quantitative analyses a two-sided test of statistical significance will be used with an alpha level of 0.05.

Qualitative data

For the implementation support intervention (cluster A), semi structured focus groups interviews will be conducted on each intervention site. A focus group will include 3–4 members of the CR team and a facilitator with experience of qualitative interviews will conduct the interviews. The interviews will assess the perception of contributing organizational, contextual, and structural factors that impact successful/unsuccessful uptake of the guidelines. Depending on number of centres accepting participation, approximately 10–12 interviews will be performed (pre- and post-implementation support intervention). Interviews will be conducted face-to-face or virtually via videoconferencing depending on the CR team´s and the facilitator´s availability and preference. To ensure anonymity, any identifiable information shared by participants will be dissociated from individual identities before analysis. The interview questions will be designed using the CFIR interview guide [30]. The interviews will be audio-recorded, transcribed verbatim, and analysed with descriptive, qualitative content analysis with an inductive and manifest approach according to Graneheim and Lundman [31, 32]. CFIR definitions and coding guidelines will be used to assist with coding of qualitative data [26].

Cost and cost-effectiveness analysis

A comparative health economic evaluation of the implementation and the usual care models will be conducted. Generally, an economic evaluation serves to provide decision makers with relevant information about the value for money as to alternative treatment models. The economic evaluation will adopt generally accepted methods for such types of analyses, including the estimation of all relevant costs and benefits from both a health system and a societal perspective. Data on both direct and indirect costs of the two models will be collected through a survey where all centres will fill out the resources needed to ensure the implementation of the intervention.

Based on the cost estimates and the effect of the implementation assistance on patient outcomes, the economic evaluation will then be able to conduct a cost-effectiveness analysis (CEA). The CEA responds to the key policy question of the cost of the measured effects. Such information contributes to making informed priority decisions under fixed budget constraints in healthcare services. In addition, the incremental cost-effectiveness ratio (ICER) will be computed to assess the added costs relative to the added effects (benefits) of the intervention model compared with usual care:

$$\mathrm=\mathrmt-Costuc/Effectst-Effectsuc}$$

where t = treatment option and uc = usual care option. The ICER shows the additional (incremental) costs of implementing the treatment model compared with the usual care option. Consequently, the ICER responds to the related policy question of how much more will be achieved for how much more resources (costs) compared with the current situation. The effect measures include those identified above under Study objectives (Sect. 4): guidelines adherence and patient-level outcomes.

Ethical considerations and withdrawal criteria

The study will be performed in compliance with the study protocol, the Declaration of Helsinki, and current national and international regulations governing this clinical trial. The study has been approved by the Swedish Ethical Review Authority (Registration number: 2023-03217-01) and is registered at ClinicalTrials.gov (identifier: NCT05889416).

The SWEDEHEART registry is sanctioned by Swedish law, stating that all patients are informed of their inclusion and their right to opt-out and have their data erased at any time without a specific reason [13]. Opt-out is extremely rare, counting fewer than ten cases per year.

All centres report data to the SWEDEHEART registry on a voluntary basis. For the audit intervention an opt-out approach will be applied, i.e., CR centres not willing to provide answers to the new variables will be asked to convey this to the registry. Otherwise, if they submit answers to the new variables, they will be included in the analysis.

For the implementation support intervention, a perquisite for participation is a verbal consent from a centre director or key stakeholder followed by a signed Letter of Intent from the CR centre director.

Data protection

The SWEDEHEART registry data is collected through an interactive web-based IT-platform, developed and maintained by Uppsala Clinical Research centres (UCR), Uppsala, Sweden. The data is electronically transferred to UCR in encrypted format and stored on a central server.

To ensure correct data matching and analysis, centre-level data (structure and process variables) will be requested in an identifiable form (i.e., name of CR centre is linked to data). All patient data from the SWEDEHEART registry and other national registries is, however, delivered pseudonymized to researchers, only containing a study identification number for each patient. As patient data contains information on at which centre the patients had their follow-up, matching to centre-level data will be done by CR centre. A patient-level identification key (study identification number linked to personal identification number) will be stored at the National Board of Health and Welfare, to allow for delivery of long-term outcome data – in the case of this study for up to 5 years.

All electronic study data delivered to the research team will be stored in a locked data storage requiring double identification for access. Only members of the research team will have access to data. Data processing will be performed in accordance with the provisions of the General Data Protection Regulation (GDPR) and other relevant legislation. No data will be shared outside of Sweden.