Prognosis of dementia

Abstract

Background: In this thesis, we focused on the prognosis of patients with dementia who visited a hospital (inpatient or day clinic care) in the Netherlands. So far, absolute mortality risks for dementia were lacking in the Netherlands, whereas these risks have been available for years for cancer or cardiovascular disease. Therefore, we aimed to calculate nationwide estimates of absolute mortality. Moreover, we assessed the impact of several risk factors on mortality and morbidity in dementia. Insight into prognosis is helpful for patients and will support management in daily practice. Methods: For the studies described in this thesis, a hospital-based cohort was constructed (n=59,201) by linkage of national registers, including the hospital discharge register, the population register and the national cause of death register. Results: First of all, we showed that the validity of the hospital discharge register to identify patients with dementia was high. We also showed that absolute mortality risks, stratified by age, sex and setting of care (inpatient versus day clinic care) were high. Mortality risks were significantly higher in men than in women and in admitted patients than in those visiting a day clinic. Moreover, mortality was two times higher in patients visiting a day clinic compared with mortality in the general population and was even higher at younger ages. Mortality risk among inpatients was comparable or even exceeded the risk following a hospitalization for other severe diseases, including acute myocardial infarction, heart failure or stroke, particularly at younger ages. Furthermore, a history of cardiovascular disease (including coronary heart disease, heart failure, atrial fibrillation, stroke and other cardiovascular disease) and a low socioeconomic status were negatively associated with mortality and readmission, particularly in patients visiting a day clinic. There were no differences in prognosis between ethnic subgroups. We found a decline in mortality across the years 2000-2008. And whereas the risk of readmission decreased among patients visiting a day clinic in this period, the risk increased among patients admitted to the hospital. We also studied underlying causes of death in dementia. Leading underlying causes of death were dementia itself and CVD. Compared to the general population CVD was a less common cause of death in demented patients. The same holds true for cancer. These differences in causes of death were most pronounced in the relatively young olds (60-69 years) than in those aged ≥ 90 years. Finally, we developed two models to predict one and three year mortality. A great advantage of the models is that they are easy to apply in daily practice as information on the included variables can be easily and quickly obtained during consultation hours. The performance of the models was acceptable. Before the models can serve daily practice, they should be externally validated. Conclusion: Given the poor prognosis of dementia shown in this thesis, it is essential that clinicians involved in the care for patients with dementia take the findings of this thesis into account in daily practice, especially with respect to timely and targeted advance care planning.