Clinical epidemiology · 2024
Cardiovascular Data Quality in the Danish National Patient Registry (1977-2024): A Systematic Review.
СтатьяОбзорНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Lund Katrine Hjuler, Fuglsang Cecilia Hvitfeldt, Schmidt Sigrun Alba Johannesdottir, Schmidt Morten
Аннотация
BACKGROUND: The increasing use of routinely collected health data for research puts great demands on data quality. The Danish National Patient Registry (DNPR) is renowned for its longitudinal data registration since 1977 and is a commonly used data source for cardiovascular epidemiology.
OBJECTIVE: To provide an overview and examine determinants of the cardiovascular data quality in the DNPR.
METHODS: We performed a systematic literature search of MEDLINE (PubMed) and the Danish Medical Journal, and identified papers validating cardiovascular variables in the DNPR during 1977-2024. We also included papers from reference lists, citations, journal e-mail notifications, and colleagues. Measures of data quality included the positive predictive value (PPV), negative predictive value, sensitivity, and specificity.
RESULTS: We screened 2,049 papers to identify 63 relevant papers, including a total of 229 cardiovascular variables. Of these, 200 variables assessed diagnoses, 24 assessed treatments (10 surgeries and 14 other treatments), and 5 assessed examinations. The data quality varied substantially between variables. Overall, the PPV was ≥90% for 36% of variables, 80-89% for 26%, 70-79% for 16%, 60-69% for 7%, 50-59% for 4%, and <50% for 11% of variables. The predictive value was generally higher for treatments (PPV≥95% for 92%) and examinations (PPV≥95% for 100%) than for diagnoses (PPV≥80% for 71%). Moreover, the PPV varied for individual diagnoses depending on the algorithm used to identify them. Key determinants for validity were patient contact type (inpatient vs outpatient), diagnosis type (primary vs secondary), setting (university vs regional hospitals), and calendar year.
CONCLUSION: The validity of cardiovascular variables in the DNPR is high for treatments and examinations but varies considerably between individual diagnoses depending on the algorithm used to define them.
OBJECTIVE: To provide an overview and examine determinants of the cardiovascular data quality in the DNPR.
METHODS: We performed a systematic literature search of MEDLINE (PubMed) and the Danish Medical Journal, and identified papers validating cardiovascular variables in the DNPR during 1977-2024. We also included papers from reference lists, citations, journal e-mail notifications, and colleagues. Measures of data quality included the positive predictive value (PPV), negative predictive value, sensitivity, and specificity.
RESULTS: We screened 2,049 papers to identify 63 relevant papers, including a total of 229 cardiovascular variables. Of these, 200 variables assessed diagnoses, 24 assessed treatments (10 surgeries and 14 other treatments), and 5 assessed examinations. The data quality varied substantially between variables. Overall, the PPV was ≥90% for 36% of variables, 80-89% for 26%, 70-79% for 16%, 60-69% for 7%, 50-59% for 4%, and <50% for 11% of variables. The predictive value was generally higher for treatments (PPV≥95% for 92%) and examinations (PPV≥95% for 100%) than for diagnoses (PPV≥80% for 71%). Moreover, the PPV varied for individual diagnoses depending on the algorithm used to identify them. Key determinants for validity were patient contact type (inpatient vs outpatient), diagnosis type (primary vs secondary), setting (university vs regional hospitals), and calendar year.
CONCLUSION: The validity of cardiovascular variables in the DNPR is high for treatments and examinations but varies considerably between individual diagnoses depending on the algorithm used to define them.