The current Health Information Systems offer an opportunity to implement advanced methodology to generate evidence within the context of the healthcare activity. This evidence could complement that of the clinical trials, often challenged due to poor external validity. This use of real world data (RWD) has also its limitations, mainly due to selection bias, confusion, and the comprehensiveness and quality of the data itself. Congestive Heart Failure (CHF) is one of the most prevalent and lethal diseases in Europe. It is currently increasing in frequency and complexity and, therefore, the burden over the Health System is also growing. These elements make CHF a good touchstone to learn to manage these data and to explore its epidemiological and predictive possibilities. Our objectives are to describe:

–the clinical and demographic characteristics of the population diagnosed of CHF over a 5-year period (2011–2015);

–the epidemiology of the disease;

–the burden on the health system in terms of use of services.

Subjects: people of 40 years or older, alive after a first hospitalization by CHF in the public hospitals of the Basque Country (population 2,100,000). Sources of data: business Intelligence Platform, which includes all the information collected in hospitals, primary care, emergency rooms, day-hospitals of the Basque Health Service. Also, mortality data and socioeconomic data provided by the Basque Health Department. The variables we looked for were: demographics; use of health services during the 12 months previous to the index event; comorbidity at baseline (Charlson Index); socioeconomic status; mortality, no programmed hospitalizations - including potentially preventable hospitalizations - during one year after the index event. Analysis: descriptive, variation estimators (systematic coefficient of variation - SCV), rates of events, temporal trends. Analyses were conducted using the R version 3.3.2.

Baseline characteristics: 15,489 patients were included in the cohort, 7679 (49.6%) men and 7810 (50.4%) women. Women were older than men (82±9 vs. 77±11), but the Charlson index was greater in men (3.5±2.1 vs. 2.8±1.7 in women). Among the patients, 5640 (36%) had been hospitalized during the previous year. Among the patients, 4354 (28%) were in hospital and received a cardiovascular disease (CVD) diagnosis, other than CHF. Among the subjects, 12,755 (82%) had had a consultation in an outpatient clinic and in 6381 cases (41% overall) the consultation was with a cardiologist. Finally, the persons in the cohort visited their general practitioner an average of 14 (sd=11) times and the nurse an average of 16 (sd=13) times. Epidemiology: as of 31 of December 2015, the number of cases with the disease registered was 15,152 (prevalence in>40 years=1.2%). During the five years under study, the age-standardized incidence of first hospitalizations by CHD varied between 2.2 and 2.4 per thousand persons/year (women) and between 3.6 and 3.8 per thousand persons/year (men), without an identifiable temporal pattern. Regarding the geographical areas of the Basque Country, the variation was small (SCV<0.10 in all the years). Burden of the disease to the health system: 365 days after the discharge, 3864 (25%) patients had died - 820 (5%) within the 30 first days. Another 7899 had been hospitalized 14,744 times: 4714 (32%) due to CHF; 1976 (13%) due to other CVD; 8054 (55%) due to other diagnosis unrelated to CVD. Of these hospitalizations, 5143 were labelled as potentially preventable: 43% (CHD); 25% (CVD-no CHF); 33% (no CVD hospitalizations).

Our results replicate those obtained by other studies developed with very different methodologies, which means that RWD, if adequately managed and analysed, is a good method to do clinical research. As this data is collected in a continuous way, rather than point-to-point, it will allow using all the information of a given patient, without the restrictions of the ad-hoc research.