Article Text
Abstract
Background Cause-of-death statistics underestimate the end-of-life occurrence of many diseases, including diabetes. Our aim is to estimate the end-of-life occurrence of diabetes by combining multiple cause of death data with register data covering primary care, hospital care and medication prescriptions. We investigate the added value of each medical register and assess the extent to which reporting of diabetes as a cause of death is associated with disease severity.
Methods Our study population consisted of all deaths in the Netherlands (2015–2016) included in the Dutch primary care database (NIVEL-PCD; N=18.162). The proportion of deaths with diabetes (type I or II) within the last two years of life was calculated using cause of death and medical register data in isolation and combined. We assessed whether the proportion of diabetes reported in the causes of death registries varied according to disease severity as defined by medication prescriptions.
Results Of all deaths, 2.2% had diabetes reported as the underlying cause of death, while 7.7% of the death certificates mentioned diabetes. Primary care registration yielded the highest rate in isolation (27.1%), followed by the medication (22.4%) and the hospital-any diagnosis (17.1%) data, while hospital-main diagnosis was limited (1.1%). According to all data sources combined, 28.7% of the study population had diabetes at the end of life. Of all deaths among those who were prescribed insulin (indicating severe diabetes), 11.9% had diabetes recorded as the underlying cause of death and 35.8% of the death certificates mentioned diabetes as a cause of death. For patients using oral antidiabetic medication, these proportions were only 5.7% and 24.2%, and for patients not using antidiabetic medication 4.2% and 16.7%.
Conclusion More than one fourth of the Dutch population has diabetes at the end of life. Only a minority of this end-of-life diabetes occurrence is recorded as a cause of death, even for persons with severe diabetes. In the Netherlands, combining primary care data with multiple causes of death allows to find most cases with diabetes at the end of life.