This study was a retrospective, non-interventional observational study conducted within Region Halland, which is situated in the southwestern region of Sweden with a population of 340,000 residents. The healthcare infrastructure within this region is comprised of three acute care hospitals, 40 inpatient wards, two emergency departments, 30 outpatient specialty clinics and 46 primary care clinics. The study included 23 primary care clinics operating under private administration alongside an equal number under public administration.

Region Halland possesses comprehensive access to pseudo-anonymized data through the Regional Healthcare Information Platform (RHIP) [18]. This dataset includes clinical, operational capacity, and financial information pertaining to all individuals who have received treatment since 2011 across all publicly-operated healthcare facilities within Region Halland. RHIP was the primary data source for this study, and a similar methodology has been previously used to investigate the population with heart failure within Region Halland [19, 20]. The data from RHIP includes primary care, emergency department care, hospital admissions and hospital outpatient care as well as inpatient care. It includes the complete Region Halland patient population linking clinical, operational, and cost information at the patient encounter level, together with system resource and capacity data (e.g., full-time equivalent nurses/physicians; hospital bed occupancy). RHIP also contains data concerning deceased patients including the date of death and therefore contains sufficient data to analyze all-cause mortality in the cohort. Information on dispensed drug treatment was retrieved from the National Drug Registry.

The study included individuals aged ≥ 40 years who’d had a T2D diagnosis between 2009 and 2019 and received a new cardiovascular diagnosis sometime between 2016 and 2019 upon inclusion to this study. Patients who had a CVD diagnosis prior to 2016 were excluded. ICD-codes for cardiovascular diagnosis are displayed in Supplementary Table 1. The index date for each patient was when they were first diagnosed with a new cardiovascular diagnosis. All participants were living in Region Halland during the study period.

The follow-up duration extended for one year from the index date, or until the patient’s death if that occurred earlier. The study procedure is displayed in Supplementary Fig. 1. Each patient was enrolled at the onset of their new cardiovascular diagnosis and concurrently when the new cardiovascular diagnosis was documented. The point at which the patient received the diagnosis was categorized into the following settings: hospital inpatient and outpatient care, primary care-, and the emergency department. Visits to the emergency department, outpatient care or primary care that led to a hospital admission within 48 h were defined as hospital inpatient care. The diagnosis defining the CVD and T2D is presented in Supplementary Table 1.

A Kaplan-Meier analysis depicting trends over time for the mortality and the events of emergency department visit/hospitalization, based and the mortality and events patterns over time for in-patient care, out-patient care, primary care, and the emergency department

Gathered at index were age at the index date, sex, comorbidities, specific cardiovascular diagnoses, estimated glomerular filtration rate (eGFR), glycated hemoglobin (HbA1c) levels, cholesterol values, and recorded blood pressure measurements. The specific diagnoses for comorbidities and cardiovascular diagnosis are listed in Supplementary Table 1. All laboratory results and blood pressure readings were collected over the entire study period and then averaged. The follow-up of laboratory values was those within three months before the study period ended. HbA1c values were categorized into four groups: <52 mmol/mol, 52–57 mmol/mol, 58–70 mmol/mol, and > 70 mmol/mol. Total cholesterol was classified as either ≥ 4.5 mmol/L or < 4.5 mmol/L, while LDL cholesterol was categorized as either ≥ 2.5 mmol/L or < 2.5 mmol/L. Renal function was stratified into three eGFR levels: >60 mL/min/1.73 m², 30–60 mL/min/1.73 m², and < 30 mL/min/1.73 m² [21, 22]. The urine albumin-to-creatinine ratio (UACR) was grouped into three categories: normal albuminuria (< 3 mg/mmol), microalbuminuria (3–30 mg/mmol), and macroalbuminuria (> 30 mg/mmol). Systolic blood pressure was divided into three ranges: <130 mm Hg, 131–139 mm Hg, and > 140 mm Hg.

The pharmacotherapies for blood pressure, diabetes and cholesterol were retrieved and the ATC codes are specified in Supplementary Table 1. The total number of days under care, hospitalizations, outpatient care visits, primary care visits, and emergency department visits were documented for each patient. Outpatient care and primary care visits were further categorized based on whether they involved consultations with a physician, nurse, or paramedical personnel. The study also recorded the number of follow-up visits in outpatient care, primary care, the emergency department, and the possibility of subsequent readmissions.

Visits to the emergency department, hospitalization and the date of death were registered. Number of days, until the first event occurred, was also registered. The Individual visits to hospital outpatient care or primary care visit to a physician or nurse within 30 days were registered.

Descriptive statistics was used to describe characteristics of the study population, including age, sex, comorbidities, laboratory tests and medication. Continuous variables were presented with mean and Standard Deviation (SD) and categorical variables were presented with frequency and percentages. The comparison of continuous variables was performed using Student’s t-test and when several groups were compared, the ANOVA analysis was used. All statistical tests were 2-sided, unless otherwise specified, and p < 0.05 was used to identify significant differences.

Relative risks for an emergency department visit/hospitalization and mortality were estimated by Hazard ratios (HRs) with 95% confidence intervals (CI), which were calculated in two Cox regression analyses with adjustments made for age, sex, atherosclerotic CVD, glucose regulation, lipid levels, kidney function, blood pressure levels and pharmacotherapy for diabetes, blood pressure and cholesterol. Two Kaplan-Meier plots for the outcome of emergency department visits/hospitalizations and for the mortality were performed. Statistical analyses were performed in IBM SPSS Statistics 29.

This study received ethical approval from the Swedish Ethical Review Board, at the Gothenburg Department of Medicine, under registration number 2020–05769. In this retrospective observational cohort study, the requirement for informed consent was waived, as it complied with the approvals granted by the Swedish Ethical Review Board. The methods and procedures employed in this research adhered to the relevant research guidelines and regulations.