Development of data standards

When deciding on the data standards, the methodology used for the development of international clinical data standards for the European Unified Registries for Heart Care Evaluation and Randomised Trials (EuroHeart) was followed [17]. The domains covered were the structure and processes applied within CR, as recommended in the National Guidelines on Secondary Prevention for patients with CCS [16]. The process of producing the data standards is summarized in Fig 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schematic diagram of the development process for constructing data standards for cardiac rehabilitation processes and structures within the SWEDEHEART-CR registry.

CR, cardiac rehabilitation; SWEDEHEART, The Swedish Web-system for Enhancement and Development of Evidence-based care in Heart disease Evaluated According to Recommended Therapies.

https://doi.org/10.1371/journal.pone.0293840.g001

A Data Science Group comprised of three members from the SWEDEHEART-CR Working Group (M.L., M.B. and E.H.). M.L. had the role of project manager. The Data Science Group prepared a list of candidate variables (S1 Table). The candidate variables were based on recommendations in the National Guidelines [16].The National Guidelines were in turn based on the European Society of Cardiology (ESC) Guidelines on cardiovascular disease prevention in clinical practice [2]; diabetes, pre-diabetes, and cardiovascular diseases [18]; the diagnosis and management of CCS [19]; and the management of arterial hypertension [20], adapted to the structure of CR as provided in Sweden. Standards in the National Guidelines were also defined in accordance with recommendations on standardization and quality improvement of secondary prevention through CR programmes in Europe from the European Association of Preventive Cardiology (EAPC) [4, 5]. Variables covering central recommendations which were already implemented at all or close to all (>95%) CR centres in Sweden, were removed at this stage [21]. These included for example assessment of dietary habits, level of physical activity, and tobacco use being included in routine follow-up, and that the CR centre reports to a quality registry (S1 Table).

The Data Science Group also suggested permissible values for the variables. Three values were suggested: Yes, Partly and No. The value Partly (Swedish “Delvis”) was included to reduce the risk of a ceiling effect with dichotomous answers. For the human resources domain, the Partly value enables a more nuanced answer than Yes/No if the CR team only has a part-time service from one of the professions mentioned. This mostly applies to dieticians, psychologists, and social workers. For the other variables this permissible value denotes if the routine is partly implemented/followed (i.e., using registry data continuously for improving quality of care) or if a part of the staff has the respective qualification (i.e., some of the nurses in the CR team having delegation to titrate medication).

The SWEDEHEART-CR Working Group had the role of selecting key variables and assuring face validity. The working group comprised 13 members: six cardiologists, one primary care physician, two physiotherapists, two nurses, and two psychologists. All members have extensive clinical and/or research experience within CR and between 1–12 years of experience working with quality registries. Candidate variables were presented by the Data Science Group to the Working Group which were asked to select key variables based on the following criteria:

• The variable is included in the National Guidelines
• The variable is included in ESC guidelines or EAPC consensus documents
• The variable is associated with patient outcomes
• The data standard has room for improvement within Swedish CR
• The variable is clearly written and objective
• Clear format and structure of the information to be captured within a variable (permissible value)

To be accepted as key variables, candidate variables should fulfil at least 4 out of these 6 criteria. The Working Group came to a joint conclusion after three discussion rounds (two e-mail discussion rounds and one face-to-face meeting). Phrasing was further adapted after the final discussion round.

A Reference Group was comprised of six cardiologists, ten nurses, and five physiotherapists from eight CR centres in Sweden. The Reference Group had the role of peer reviewing, ascertaining content validity and test-retest reliability. The group had a wide geographical representation and included representatives from small (less than 50 post MI patients followed per year, n = 3), medium-sized (50–150 patients followed per year, n = 3), and large (more than 150 patients followed per year, n = 2) CR centres across the country. Additionally, five nurses specialized in cardiac care who work for the SWEDEHEART registry as service contacts for registry users were included in the Reference Group. Key variables were reviewed by the Reference Group, providing critical feedback on the selection of variables (using the same criteria as listed above), as well as applicability and generalizability. The final set of key variables can be seen in Table 1. Process variables already included in the SWEDEHEART-CR registry are shown in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. New data standards for auditing structure and processes in cardiac rehabilitation.

https://doi.org/10.1371/journal.pone.0293840.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Data variables already included in the SWEDEHEART-CR registry auditing cardiac rehabilitation processes.

https://doi.org/10.1371/journal.pone.0293840.t002

Setting and populations

The SWEDEHEART-CR registry was started in 2005 [12]. Since 2018 all CR centres in the country report data to the registry (n = 78). All CR centres reporting to SWEDEHEART-CR will be subject to the new data standards. The patient population included in SWEDEHEART-CR has been previously described in detail [12]. In short, all patients with a type-1 MI discharged alive from coronary care units around the country are eligible for CR follow-up and are consequently eligible for registration in SWEDEHEART-CR.

Ethical considerations

The need for signed informed consent by patients for inclusion in Swedish quality registries has collectively been waived in Sweden. At hospital admission, MI patients are informed verbally and in writing about data being collected and entered in the registry. All patients have the right to deny registration and upon request to be removed from the registry at any time. As the data standards described herein will be reported at centre-level there is no need for additional informed consent at patient-level or ethical review board consent.

In an upcoming trial, however, we aim to evaluate whether auditing structure and processes of CR through the SWEDEHEART registry can increase adherence to national guidelines and improve patient outcomes. This trial has been evaluated and approved by the Swedish Ethical Review Board (Dnr 2023-03217-01; ClinicalTrials.gov ID: NCT 05889416).

Baseline and follow-up data

Patient-level data from consecutive patient follow-up visits at the CR centre is reported at fixed time points after the MI to SWEDEHEART-CR, with data collected from electronic medical records (i.e., comorbidities, current medication, and re-admissions), self-report (i.e., symptoms, physical activity, and dietary habits), or by measurement (i.e., blood pressure, lipid levels) (Fig 2). As the data standards proposed here are not on a patient-level, they will be administered on centre-level biannually. The variables will be accessible for one month at a time. To receive responses from all key CR professions (nurse, physiotherapist, and physician), the CR team will be encouraged to answer the questions together. The data standards will be launched during the fall 2023. To acquire a baseline status all CR centres will complete the questions for the first time. Thereafter, to enable evaluation of whether auditing service delivery can alter implementation of the national guidelines, centers will be randomized to report data bi-annually for three years or report no data.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The basic structure of visits within cardiac rehabilitation (CR) as applied at most CR centres in Sweden.

Visits at which data is collected in the SWEDEHEART registry are marked in light blue. CR, cardiac rehabilitation; EBCR, exercise-based cardiac rehabilitation; MI, myocardial infarction.

https://doi.org/10.1371/journal.pone.0293840.g002

Data capture and storage

SWEDEHEART-CR data is collected through an interactive web-based platform developed and maintained by Uppsala Clinical Research Centre (UCR; http://www.ucr.uu.se). The data is entered online by healthcare professionals at each CR centre. Patient-level data is transferred encrypted and stored in a central secure server. Data for benchmarking is continuously accessible online and is published annually in the SWEDEHEART annual reports (http://www.ucr.uu.se/swedeheart).

Data quality, endpoints, and linkage to other data

As the primary aim of the data standards is to audit implementation of national guidelines, no systematic literature review was performed when preparing the candidate variables. In some instances, when converting a recommendation from the guidelines to a candidate variable, the phrasing was adapted for clarity and to increase objectiveness.

Data entry to SWEDEHEART is voluntary, and centres get no reimbursement for doing so. By entering data, however, the centres and their patients have continuous online access without direct cost to local and national benchmarking. As for all other data entered to SWEDEHEART, a user manual with definitions of each variable will be provided to ensure correct registration of data. Also, the service contacts are available to aid healthcare personnel responsible for data entry if needed. The data platform has error checking for range and consistency. Since the start of the registry, SWEDEHEART data has been internally audited every other year to certify the correctness of the data entered. The auditing has consistently confirmed >95% agreement with data from medical records [12].

Since every patient for whom data is entered into the SWEDEHEART registry has a unique personal identification number and is linked to the CR centre responsible for their follow-up, the data standards described here can be linked to patient data at each centre. This allows auditing not only of adherence to guidelines (data standards proposed) but also patient outcomes. Linkage with other nationwide registries in Sweden which collect data on hospital admissions, diagnoses, causes of death, pharmacotherapy, and socioeconomic data can also be performed using the personal identification number.

Future plans

Whether auditing service delivery of CR within a national quality registry can improve adherence to guidelines and patient outcomes has to our knowledge not been scientifically evaluated. Therefore, as a next step we aim to evaluate whether auditing structure and processes of CR in a national registry can increase adherence to national guidelines and improve patient outcomes.