Data Resource Profile: Danish Heart Statistics
Morten Schmidt, Lisbeth Vestergaard Andersen, Søren Friis, Knud Juel, Gunnar Gislason · International Journal of Epidemiology · 2017
The Danish nationwide registries are acknowledged for their continuously updated high-quality health data,1,2 allowing for studies of long-term trends in drug use3 and incidence and outcome of diseases.4–6 Whereas strict data permissions are needed for the vast majority of such registries, a few freely available online data resources exist.7 These data sources have allowed examinations of, for example, long-term trends in use of antithrombotic,8 glucose-lowering9 or non-steroidal anti-inflammatory drugs,10 but to a lesser extent disease trends. Since 2015, a new online data source has facilitated free access to nationwide data on drug use and disease epidemiology. This data source, the Danish Heart Statistics, available at: [www.hjertetal.dk], was launched by the Danish Heart Foundation and the National Institute of Public Health at the University of Southern Denmark. As a unique feature among online resources, the Danish Heart Statistics provides nationwide comprehensive cardiovascular health data in Denmark since 2006, including data on diagnosis, examinations, therapy and prognosis. Thereby, the Danish Heart Statistics provides information of profound value for researchers, politicians, administrators, media and others. Moreover, within the same search engine, the Danish Heart Statistics holds linkages to other nationwide registries, allowing complete follow-up of death and subgroup analyses according to demography, education, ethnic origin and region. Herein, we review the Danish Heart Statistics as a data resource and potential research avenue. As this study was based solely on registry data and did not involve any contact with patients, no approval was required from the Danish Scientific Ethical Committee. The Danish Heart Statistics is based on five Danish nationwide registries: the National Patient Registry,1 the National Prescription Registry,2 the Civil Registration System,11 the Registry of Causes of Death12 and the Population Education Register.13,Table 1 provides summary descriptions of these registries. All registry data sources are linked by the unique Central Personal Register (CPR) number assigned to all Danish residents.11 The Danish National Health Service provides tax-supported health care, guaranteeing uniform access to general practitioners and hospitals, and partial reimbursement for prescribed medications, including all cardiovascular drugs.14 Data sources linked and incorporated into Danish Heart Statistics Diseases Examinations Procedures Surgery Validity is generally high for cardiovascular measures within the study period (see Table 2)1,17,18 Completeness has only been accessed for a few measures. However, the nationwide coverage ensures a high coverage for examinations, procedures and surgery, and also the majority of diseases treated at hospitals1 Prescribed drugs Drug costs Validity considered high Completeness considered high for all cardiovascular drugs, although a small proportion of low-dose aspirin is sold over the counter2,10 Sex and age Vital status Migration status Region of residence All-cause mortality and migration data are considered valid and complete11 Cardiovascular deaths Based on subjective judgement Caution should therefore be exercised when interpreting the data Level of education Validity considered high Completeness is high (> 96% of Danish inhabitants aged 15–69 years) Diseases Examinations Procedures Surgery Validity is generally high for cardiovascular measures within the study period (see Table 2)1,17,18 Completeness has only been accessed for a few measures. However, the nationwide coverage ensures a high coverage for examinations, procedures and surgery, and also the majority of diseases treated at hospitals1 Prescribed drugs Drug costs Validity considered high Completeness considered high for all cardiovascular drugs, although a small proportion of low-dose aspirin is sold over the counter2,10 Sex and age Vital status Migration status Region of residence All-cause mortality and migration data are considered valid and complete11 Cardiovascular deaths Based on subjective judgement Caution should therefore be exercised when interpreting the data Level of education Validity considered high Completeness is high (> 96% of Danish inhabitants aged 15–69 years) Data sources linked and incorporated into Danish Heart Statistics Diseases Examinations Procedures Surgery Validity is generally high for cardiovascular measures within the study period (see Table 2)1,17,18 Completeness has only been accessed for a few measures. However, the nationwide coverage ensures a high coverage for examinations, procedures and surgery, and also the majority of diseases treated at hospitals1 Prescribed drugs Drug costs Validity considered high Completeness considered high for all cardiovascular drugs, although a small proportion of low-dose aspirin is sold over the counter2,10 Sex and age Vital status Migration status Region of residence All-cause mortality and migration data are considered valid and complete11 Cardiovascular deaths Based on subjective judgement Caution should therefore be exercised when interpreting the data Level of education Validity considered high Completeness is high (> 96% of Danish inhabitants aged 15–69 years) Diseases Examinations Procedures Surgery Validity is generally high for cardiovascular measures within the study period (see Table 2)1,17,18 Completeness has only been accessed for a few measures. However, the nationwide coverage ensures a high coverage for examinations, procedures and surgery, and also the majority of diseases treated at hospitals1 Prescribed drugs Drug costs Validity considered high Completeness considered high for all cardiovascular drugs, although a small proportion of low-dose aspirin is sold over the counter2,10 Sex and age Vital status Migration status Region of residence All-cause mortality and migration data are considered valid and complete11 Cardiovascular deaths Based on subjective judgement Caution should therefore be exercised when interpreting the data Level of education Validity considered high Completeness is high (> 96% of Danish inhabitants aged 15–69 years) The Danish National Patient Registry1 contains detailed data on all somatic admissions in Denmark since 1978 and on all emergency room or outpatient (ambulatory) hospital visits, and psychiatric admissions, since 1995. To ensure a minimum 10-year hospital history, the Danish Heart Statistics provides statistics on cardiovascular variables from 2006 onwards. Each hospital discharge or outpatient contact is recorded in the Patient Registry with one primary diagnosis and, when relevant, one or more secondary diagnoses. In the Danish Heart Statistics, only atrial fibrillation and heart failure are included as secondary diagnoses (Table 2). Diagnoses are included from both ongoing and completed outpatient contacts. Diagnoses have been coded according to the International Classification of Diseases, Tenth Revision (ICD-10) since 1994. Since 1996, surgical procedures have been coded according to the Danish version of the Nordic Medico-Statistical Committee (NOMESCO) Classification of Surgical Procedures.1 The discharging physician records all diagnoses at the time of hospital discharge or at the end of an outpatient contact. The attending physician/surgeon registers treatment and examination codes immediately after their completion. Thus, each treatment and examination are assigned a unique date, independent of the dates of admission and discharge. Registry codes Registry codes The Danish Heart Statistics incorporates information on inpatient admissions, outpatient contacts, and procedures performed during 1994–2014 (Table 2). Furthermore, the Heart Statistics includes data on number of bed days and costs of admissions, calculated from the Diagnosis Related Group Payment system.1 Transfers between hospitals within the same day (independent of cause) are considered part of the same admission. The Danish National Prescription Registry2 has recorded data on all prescriptions dispensed from community pharmacies in Denmark since 1995. In brief, for each filled prescription, the Prescription Registry records the patient’s CPR number, number of pharmaceutical units (e.g. tablets) and packages, the Anatomical Therapeutic Chemical code,15 volume, package name, active substance, dosage and date of dispensing. The dosing schedules and indications are not recorded, and no information is available on drugs used at hospital level. From the Prescription Registry, the Danish Heart Statistics retrieves the prescription data on all individuals prescribed cardiovascular drugs (ATC codes: C and B01) and, additionally, the cost of the redeemed medication. The Danish Civil Registration System11 has recorded continuously all changes in vital and migration status for the entire Danish population since 1968. The Civil Registration System administers the CPR number and keeps track of changes in address, migration and death in the entire Danish population. This provides the Danish Heart Statistics with information on sex, age, vital status, migration status, and municipality and region of residence of all residents in Denmark at 1 January each year. Moreover, the Civil Registration System also records country of birth of immigrants and descendants, enabling the Danish Heart Statistics to categorize ethnic origin as Denmark, Western countries, Middle Eastern and North African countries, Far East countries or ‘other countries’. The Danish Registry of Causes of Death12 contains information on deaths occurring in Denmark since 1943, including date and specific cause(s) of death coded according to ICD-10 (since 1994). The Danish Heart Statistics includes all diseased residents recorded with cardiovascular disease as the primary (underlying) cause of death since 2006. Causes of deaths are used in the estimation of disease incidence among residents with no hospital record of cardiovascular disease. The Population Education Registry,13 administered by Statistics Denmark, holds information on all completed educational programmes at Danish state schools and universities since 1981. In 2008, this registry reached education information on 96% of Danish residents aged 15–69 years. Also among immigrants the coverage is high (85–90%). The Danish Heart Statistics includes the highest level of education achieved, categorized as basic school, high school, medium length education, higher education, or unknown. The Danish Heart Statistics has nationwide coverage since 2006 and currently contains data through 2014. The data resource provides statistics on incidence, prevalence and prognosis of cardiovascular disease in the adult Danish population, as well as data on medical, interventional or surgical therapy performed for cardiovascular disease. Figure 1 depicts how to use the Danish Heart Statistics in a stepwise approach, including cardiovascular measures (i.e. disease, procedure or treatment) available for analysis (step 1), the outcomes available for each cardiovascular measure (step 2) and various options for stratification (step 3). Step 1 includes variables for the most common 12 cardiovascular diseases, eight procedures and four specific therapeutic groups. Table 2 provides a detailed description of all ICD, procedural and ATC codes. Information on a variety of different outcomes is available for each cardiovascular entity; however, as depicted by the colour coding in Figure 1, all outcomes are not available for all entities. For all diseases, information is available on incidence, prevalence, mortality, ambulatory contacts, inpatient admissions, readmissions, hospital bed days and 1- and 5-year cumulative survival proportions. In addition, short-term (30-day) mortality risk is available for acute coronary syndromes and stroke. For procedures, information is available on number of individual patients treated and the total number of procedures performed (as some patients may undergo several procedures). Information is also available on 1-year cumulative survival. For medical drug therapy, information is available on total number of users and medical costs per year. Descriptive statistics are by default provided for men and women separately, but stratification is also available by age (10-year intervals), education, ethnic origin, geographical region and calendar year (Table 3 and Figures 1–3). Incidence, prevalence and mortality rates of all cardiovascular diseases combined are furthermore available by municipality, in formats of both tables and maps (Figure 4). Number and age-standardizeda incidence rates of first-time hospitalization for myocardial infarction among men (M) and women (W) in 2014, stratified by age, education, ethnicity and region aStandardized (European Standard Population) rate per 100 000 person-years. bGenerated from the Danish Heart Statistics in correlation with the graphical output shown in Figure 1. Number and age-standardizeda incidence rates of first-time hospitalization for myocardial infarction among men (M) and women (W) in 2014, stratified by age, education, ethnicity and region aStandardized (European Standard Population) rate per 100 000 person-years. bGenerated from the Danish Heart Statistics in correlation with the graphical output shown in Figure 1. Content and stepwise use of the Danish Heart Statistics. The colours symbolize the search options available within each step: all outcomes and subgroups depicted with a solid red line are available for each cardiovascular measure in Step 1 with a similar solid red line. Similar for dotted lines and other colours. To limit chance findings due to small numbers, all stratified analyses on education, ethnic origin, geographical region and municipality are calculated as average figures for the preceding three years. Consequently, the total number of individuals may vary between the various stratified analyses. Moreover, persons with unknown country of origin are excluded from analyses of ethnic origin. To protect patient confidentiality, results are not provided for combinations involving less than 10 patients. Output from the Danish Heart Statistics is available as tables, bar charts or maps (Figures 2–4). All data can be saved as spreadsheets and tables, and data can also be converted to comma- or tabulator-separated files. The Danish Heart Statistics application (‘SAS Mobile BI’) is free for download to iPad and Android. The Danish Heart Statistics operates with a set of predefined definitions for the various outcome measures (Table 2). Incidence is defined as a first-time diagnosis of cardiovascular disease in the hospital or cause-of-death registries within 10 years, i.e. excluding any patients with previous in- or outpatient diagnoses of the same disease within 10 years; and prevalence is defined as number of persons who at 1 January each year are recorded with one or more in- or outpatient diagnoses within the past 10 years. Incident and prevalent diagnoses are provided as number of events per year and as rates per 100 000 person-years. Mortality is calculated as rate per 100 000 person-years, 30-day mortality risk (%) and 1- and 5-year survival proportions (%). Readmissions are defined as a new inpatient admission within 30 days after the first admission among 30-day survivors. Hospital bed days are defined as the total and average number of days for each inpatient admission. All rates in the Danish Heart Statistics are standardized to the age distribution of the European Standard Population, allowing international comparison.16 Data from the Danish National Prescription Registry2 and Civil Registration System (CPR number, migration data, death and country of birth)11 are accurate and complete. Educational data from the Population Education Registry13 are also almost complete, except for immigrants (as described above). Moreover, follow-up of all patients is virtually complete owing to the continuous registration of migration and vital status in the Danish Civil Registration System.11 The underlying (primary) cause of death, as recorded in the Danish Registry of Causes of Death, is typically based on subjective clinical judgment and is therefore prone to misclassification.12 Consequently, caution should be exerted in the interpretation of cause-of-death data. The Patient Registry1 covers all hospital admissions and contacts in Denmark and is often used as reference standard for other Danish registries. The validity of diagnoses and procedures recorded in the Patient Registry have recently been examined systematically and overall found to be high (Table 2).17,18 In general, cardiovascular procedures and diseases (e.g. acute myocardial infarction, stroke, heart valve disease, aortic aneurysm, deep venous thrombosis, pulmonary embolism) will always hospital admission and will therefore be by the However, of information on deaths hospital may to of both disease incidence and This is for acute cardiovascular disease, such as myocardial infarction and pulmonary should be a is not considered in Denmark a physician has of Thus, all patients in an or at a hospital are also and in the Patient Registry when no is at the of the cardiovascular diseases heart atrial also be to be at some at hospital admission or outpatient contact and be recorded in the Patient However, to the extent some of the by the Danish Heart Statistics are treated only by general this may a of disease in the Heart Statistics. Table provides an of the and of Danish Heart Statistics as a research data the nationwide coverage and the outcomes and stratification options since 2006. data are only available in and studies will information on and prescription and to other The coverage period of Danish Heart Statistics (since is also than of the underlying Patient Registry and of other nationwide registry data of detailed data on and drug and of Danish Heart Statistics as a research data source online data on all heart diseases and procedures at all Danish hospitals during online data on use of and drugs in Denmark since 2006 to and in study the Danish population has a and the uniform health small hospital from of specific hospitals, health or age and virtually complete follow-up of all patients no is the Danish Civil Registration System records vital status and on a The have been found high for all procedures, drug and most diagnoses some of will almost always be and should be into when interpreting the As an example, the codes for atrial fibrillation also the codes for atrial the However, cardiovascular events or treated only by general practitioners are not based on disease, procedure and drug data, the data are only available in Data is therefore not available on age, sex, education, ethnicity and region (see Figure to trends in diseases, procedures and drug use (see Figure analyses are by the study period since 2006 to 10 outcome measures for each disease, procedure and medical treatment (see Figure required access from online data on all heart diseases and procedures at all Danish hospitals during online data on use of and drugs in Denmark since 2006 to and in study the Danish population has a and the uniform health small hospital from of specific hospitals, health or age and virtually complete follow-up of all patients no is the Danish Civil Registration System records vital status and on a The have been found high for all procedures, drug and most diagnoses some of will almost always be and should be into when interpreting the As an example, the codes for atrial fibrillation also the codes for atrial the However, cardiovascular events or treated only by general practitioners are not based on disease, procedure and drug data, the data are only available in Data is therefore not available on age, sex, education, ethnicity and region (see Figure to trends in diseases, procedures and drug use (see Figure analyses are by the study period since 2006 to 10 outcome measures for each disease, procedure and medical treatment (see Figure required access from and of Danish Heart Statistics as a research data source online data on all heart diseases and procedures at all Danish hospitals during online data on use of and drugs in Denmark since 2006 to and in study the Danish population has a and the uniform health small hospital from of specific hospitals, health or age and virtually complete follow-up of all patients no is the Danish Civil Registration System records vital status and on a The have been found high for all procedures, drug and most diagnoses some of will almost always be and should be into when interpreting the As an example, the codes for atrial fibrillation also the codes for atrial the However, cardiovascular events or treated only by general practitioners are not based on disease, procedure and drug data, the data are only available in Data is therefore not available on age, sex, education, ethnicity and region (see Figure to trends in diseases, procedures and drug use (see Figure analyses are by the study period since 2006 to 10 outcome measures for each disease, procedure and medical treatment (see Figure required access from online data on all heart diseases and procedures at all Danish hospitals during online data on use of and drugs in Denmark since 2006 to and in study the Danish population has a and the uniform health small hospital from of specific hospitals, health or age and virtually complete follow-up of all patients no is the Danish Civil Registration System records vital status and on a The have been found high for all procedures, drug and most diagnoses some of will almost always be and should be into when interpreting the As an example, the codes for atrial fibrillation also the codes for atrial the However, cardiovascular events or treated only by general practitioners are not based on disease, procedure and drug data, the data are only available in Data is therefore not available on age, sex, education, ethnicity and region (see Figure to trends in diseases, procedures and drug use (see Figure analyses are by the study period since 2006 to 10 outcome measures for each disease, procedure and medical treatment (see Figure required access from data output from Danish Heart Statistics, by rates of first-time hospitalization for myocardial infarction among men (M) and women (W) in stratified according to age, education, ethnicity and region to in (European Standard Population) incidence rate 100 000 of first-time hospitalization for myocardial infarction among men and Danish Heart of standardized incidence prevalence and mortality rates of cardiovascular disease in Denmark by municipality level The Danish Heart Statistics provides researchers, and with access to on cardiovascular disease incidence and mortality over with for the figures between and feature of Danish Heart Statistics is of data for the and of new The information on disease and drug prevalence in Danish Heart Statistics of and of a study is Moreover, such data and measures can be incorporated into and study to the As an example, a study the to extent of or was with long-term cardiovascular on the the 5-year rates of myocardial infarction and heart failure based on data from Danish Heart Statistics, provided the study to approval from the Scientific The Danish Heart Statistics is the most online and freely available nationwide resource on cardiovascular disease and Whereas data on incidence and mortality rates are available online in some countries (e.g. the Nordic and the other for cardiovascular diseases are and have less or than the Danish Heart Statistics. In the the Heart Foundation provides on trends in cardiovascular however, data be to the of the individual The for and of in Heart disease to in several cardiovascular and as such the data of the Danish Heart is a registry with all data online by the Similar to the Heart Foundation Statistics, provides users with online the of and outcomes and over time and between However, in to the freely available data in the Danish Heart Statistics, data in are only available by application to a use as an data source for to use the Danish Heart Statistics only a reference to the Danish Heart Statistics Danish Heart with the date and the based on data from Danish Heart are information on the Danish Heart Statistics or for new can be to at the Danish Heart Foundation The Danish Heart Statistics provides free online nationwide statistics on incidence, prevalence and prognosis of cardiovascular diseases in Denmark since 2006, as well as data on medical, interventional and surgical treatment performed for these to other nationwide registries within the same search complete follow-up of death and subgroup analyses according to demography, education, ethnic origin and region. The validity of cardiovascular procedures and mortality data is The Danish Heart Statistics is a unique resource for of study and trends in cardiovascular disease incidence and and therefore has profound value for researchers, and to use Danish Heart Statistics only reference to the Danish Heart is by research from the no part in the the study and the and and the Danish Heart Statistics. All to of the and to be for all of the Institute of Public University of Southern Denmark, is acknowledged for to the of the Danish Heart Statistics. of