Objectives: To compare inpatient burden (i.e. likelihood of hospitalization, number of admissions and length of stay) in persons with newly diagnosed dementia to the general population without dementia. Additionally, to evaluate whether inpatient burden is increased during the years prior to and post dementia diagnosis, and to identify factors associated with increased inpatient burden. Method:The Israeli National Dementia Dataset (2016) was cross-linked with the National Hospital Discharge Database of the Israeli Ministry of Health (2014)(2015)(2016)(2017)(2018). Dementia definition was based on documented dementia diagnoses and/or the purchase of medications during 2016. Mixed-effects models were applied to identify demographic and health characteristics associated with inpatient burden in the one and 2 years prior to and after dementia diagnosis. Results:The dataset included 11,625 individuals aged ≥65 years, identified as incident dementia cases. Compared to the general population of older-adults without dementia, those with newly diagnosed dementia had a higher agestandardized proportion of hospitalizations (26.4% vs. 40%). The odds for hospitalization were highest during the year preceding dementia diagnosis (OR = 3.19, 95% CI 2.51-4.06) compared to 2 years prior to diagnosis, and remained high (although slightly decreased) after dementia diagnosis. Older age was associated with inpatient burden after, but not prior to dementia diagnosis. Conclusions:Older persons with dementia are a vulnerable population group with increased utilization of inpatient burden compared to those without dementia, particularly in the years surrounding dementia diagnosis. Sociodemographic risk factors may differ with respect to the time surrounding dementia diagnosis.

Background Studies show that incident dementia is associated with an increased risk of hospitalization. Recent evidence also suggests that hospitalization may be a risk factor for dementia. We aimed to capture the changes in hospitalization rates, annual number of admissions and length of stay (LOS) during 4 years surrounding dementia diagnosis in a national population‐based sample. In addition, we examined the association of sociodemographic and clinical factors with hospitalization measures in the immediate years before and after dementia diagnosis. Method The sample for this retrospective cohort study included 11,625 individuals aged 65 years and older, who were identified as incident cases of dementia in the Israel dementia national registry during 2016 (mean age= 81.21±7.19y; 61% women). Hospitalization data were obtained from National Hospital Discharge Database (NHDD) of the Israeli Ministry of Health, and clinical and sociodemographic data were extracted from the national registry database. Mixed regression models were used to assess within‐subject changes in the likelihood of hospitalization, the number of hospitalizations and LOS among survivors of the entire follow‐up period (N=8,000). We also tested for possible interactions of time respective to diagnosis with age, sex, ethnicity, socioeconomic status, BMI and comorbidity with regard to hospitalization indices. Result The risk for hospitalization as well as the number of visits and LOS were highest during the year preceding dementia diagnosis (OR =3.19, 95%CI 2.51‐4.06; OR= 3.69, 95%CI 2.86‐4.76; OR=3.53, 95%CI 2.75‐4.54, respectively, compared to the reference period of 2 years prior dementia diagnosis). These hospitalization measures slightly decreased with time after dementia diagnosis, but remained high, compared to the reference period (figure 1). The associations between age, sex, comorbidity and BMI with odds of hospitalization differed significantly depending on the time respective to dementia diagnosis (figure 2‐5). Similar associations were found for annual number of hospitalizations and LOS. Conclusion This study points to an increased burden due to hospitalizations during the year prior to dementia diagnosis and the years proceeding it. Our findings also highlight the need to consider the time surrounding dementia diagnosis in the prediction of hospital burden based on factors such as age and comorbidities.