Prevalence and Incidence Rates of Dementia: A Nationwide Population-Based Study of Electronic Health Records in Israel

INTRODUCTION

The increase in life expectancy leads to population aging worldwide. The proportion of persons aged 65 and older in Israel has increased from 4%in 1948 to 12%at the end of 2019 [1]. Dementia is the most common neurodegenerative disorder in the older population and represents one of the most significant public health challenges [2, 3]. Early-onset dementia, defined as dementia occurring before age 65, is significantly less well studied than the late-onset despite early-onset dementia often presenting with a more aggressive disease progression [4]. In a systematic review of the global literature on the prevalence of dementia (1980–2009) and meta-analysis, the estimated age-standardized prevalence for those aged ≥60 years was 5%–7%in most world regions [5]. In a large meta-analysis of 160 studies including Europe and the US the pooled point prevalence of dementia in 23 community-setting studies among individuals 60 and over was 4.9%(CI95%: 4.2–5.6), while the pooled incidence rate of dementia was 17.2 per 1,000 person-years (CI95%: 13.9–21.2) [6]. In a large systematic review of 114 studies in the US among people aged 65 and over, the prevalence rate of dementia ranged from 6.3%in Japanese Americans to 20.9%in African Americans [7]. In this review the annual incidence rate of dementia was 1.4%to 5.5%for African American, 2.3%to 5.3%for Caribbean Hispanic populations, and 0.8%to 2.7%for non-Latino white populations [7]. The 2015 global cost of dementia was estimated to be 818 US$ billion, reflecting high medical and societal costs [8].

Data on the incidence and prevalence of dementia is essential for planning and developing appropriate services at the national level. Yet, only a few countries have established national registries (e.g., Sweden, Argentina, Austria, France, and South Korea) and a few others have regional dementia registries (e.g., United States, England, Scotland, Denmark, and Spain) [9].

In Israel, some studies have examined the prevalence of dementia among particular geographical or cultural subgroups [10–14], but national-level dementia data have not been previously studied. The National Program to Address Alzheimer’s Disease and other Dementias in Israel has been led by the Ministry of Health, Division of Geriatrics, and its partners since 2013 [15]. One of the program goals is to establish national-level data on dementia. The objectives of this study, using electronic health records (EHRs), were: 1) To estimate the national incidence and prevalence of dementia among people aged 45 and older, according to age, sex, and socioeconomic status (SES); 2) To examine dementia-related medication use among prevalent and incident people with dementia.

MATERIALS AND METHODS

RESULTS

DISCUSSION

In this Israeli national dementia dataset for 2016, based on routinely collected data in EHRs, the prevalence rates of dementia among individuals aged 45+ and 65+ in 2016 were 2.5%and 6.4%respectively, and the incidence rates were 0.49%and 1.3%, respectively, when the case definition of dementia included both ICD codes and purchasing anti-dementia drugs. The rates were paradoxically lower in lower SES compared to intermediate and higher SES statuses.

Consistent with other studies [6, 14], we found a significant increase in the prevalence and incidence of dementia with age, particularly among people aged 85 years and older. We found similar dementia rates in females and males aged 65 to 74 years; however, the rates were much higher in females than males for those aged 75 years and older. This may be explained by women’s longer life expectancy and higher level of seeking out healthcare providers [18].

Dementia prevalence and incidence rates reported for different countries show considerable variability [6, 7]. Comparisons among the different studies are difficult due to methodological differences in study design, dementia definition, and use of different standard population for calculation of standardized rates [19]. In addition, there is great variability in dementia detection rates between countries [6, 7]. There are many barriers for health care systems to identify the true prevalence and incidence rates of dementia, both from the side of the patient and his family and from side of the care providers and the healthcare system. On the one hand there may be a lack of public awareness, fear of stigmatization, misinterpretation, or denial of symptoms of dementia and of the severity of cognitive impairment. One the other hand there may be lack of professional awareness and expertise, insufficient access to specific diagnostic services, brief physician-patient encounters in the face of multiple comorbid conditions [20, 21]. Indeed, studies have demonstrated 31%to 69%of patients with dementia do not receive a documented diagnosis, especially among mild and moderate dementia cases [20–22]. Previous Israeli survey found that only one fifth of the family caregivers of those older people identified by the study as having dementia reported that they had already been told by a physician that their older relative had dementia [10]. According to the recent World Alzheimer Report, standardized to the Western European population, dementia prevalence in people aged 60 years and older has substantial regional variation from 4.7%in Central Europe to 8.7%in North Africa and the Middle East [23]. Using Western Europe population as the standard population, we estimated that dementia prevalence in our study was 5.4%in people aged 60 years and older. A cut-off of 60 years is sometimes used arbitrarily and consequently the data on prevalence and incidence of early onset of dementia are much more limited [4].

Only a few studies were previously conducted in Israel on dementia prevalence, none on a national level [10, 12–14]. Israeli studies in which the cognitive assessment was performed by trained medical workers, found a higher prevalence of dementia [10, 12], but these were small non representative samples. In a recent Israeli report based on data form one HMO, the prevalence of dementia, based on dementia diagnoses or dispensation of anti-dementia drugs more than once in EHRs, was 6.6%among people 60 and over [14].

Low SES status has been associated with cognitive decline, a negative impact on health and longevity [24] through less healthy lifestyles, poorer general health, and less access to appropriate medical care resources [25]. Unlike other studies, in our study the age-standardized rates of dementia prevalence and incidence were significantly lower for people who lived in low SES neighborhoods compared to intermediate and high SES groups. Possible explanations include decreased access to health care services or necessary medical procedures, reduced access to qualified medical personnel, or lower disease awareness [26]. All of these could lead to higher rates of undetected dementia among low SES groups. Indeed, studies conducted in Israel have shown inequalities in the use of health services between Jews and Arabs [27]. A recent meta-analysis [28] as a study from China [29] reported that the proportion of undetected dementia was high, particularly in populations with low SES status.

In Israel, medical treatment for Alzheimer’s disease is included in the basic basket of services provided by the HMOs under the provisions of the National Health Insurance Law. In our study approximately 40%of people with prevalent dementia used anti-dementia medications in the last 12 months. In a multicenter European study, the authors found that the rate of AChEIs use for patients with dementia was 10.3%, varying from 3.0%in Holland to 20.3%in France [30]. The low anti-dementia treatment rate across the countries can be explained by the limited clinical efficacy of these medications and their lack of substantial impact on dementia course [30]. Moreover, there is no consensus on how long treatment with anti-dementia medications should last. Therapy is often stopped when side effects are not tolerated or when there is no perceived clinical response after a three to six-month trial.

In the present study the overall prevalence of the use of medications to treat the behavioral and psychological symptoms of dementia was 61.2%. Antidepressants were the most commonly prescribed medication (38%) in our study antipsychotic medications were prescribed in 12.3%. Patterns of antipsychotic medication use among people with dementia are variable across care settings and among countries. The prevalence of antipsychotic use among community dwelling older people with dementia ranged from 10.1%in the United Kingdom [31] to 29%in the US [32]. Depression is a common neuropsychiatric symptom of people with dementia affecting up to 43%of patients [33]. Recently published study of community-dwelling Medicare beneficiaries aged 65 years or older with a primary or secondary diagnosis of dementia, identified that a total of 49.8%filled an antidepressant prescription, followed by anxiolytics (26.8%), antiepileptics (21.9%), and antipsychotics (21.6%) [34].

Our study has several methodological strengths. Data were obtained from a non-selective nationwide population aged 45 and above, conferring good external validity. In our study, the case definition of dementia included, in addition to ICD diagnosis, also purchase of anti-dementia drugs. Although, the case definition of dementia based on dementia diagnosis in ICD-9 and ICD-10 codes may be highly specific to identifying people who actually have dementia. In addition, the positive predictive value for a dementia diagnosis in EHRs was reported up to be 90%[35]. However, because of the high rate of dementia under recording, ICD codes alone cannot serve as a gold standard for dementia diagnoses in the electronic health data [36]. General practitioners report barriers to formally recording dementia, so some patients may be known by general practitioners to have dementia but may be missing a diagnosis in their patient record [37]. Thus, using an algorithms which searches for anti-dementia medication prescriptions will be helpful for detection of true positive cases among unlabeled patients with dementia [37]. Some methodological limitations, however, should be considered. First, unlike registry, which obtains continuous online follow up for all dementia cases [38], our data was a one-time retrospective data collection including people with dementia who were alive at the end of 2016. Second, the uses of diagnoses of dementia in the EHRs were recorded mainly by primary care physicians, which may result in under-detection. Consequently, a recent meta-analysis of 23 studies, mainly from high-income countries, found the pooled rate of undetected dementia in primary care was 61.7%[28]. Third, dementia severity could not be ascertained as this data is not recorded in designated fields in the EHR. Forth, we have no valid data on dementia subtypes because in clinical practice the diagnostic criteria and the extent of workup vary substantially [39]. Finally, data on the indication for the use of medications for behavioral and psychological symptoms of dementia is not available; some of this medication use may reflect treatment for a previous psychiatric disease.

In conclusion, this national dementia data collection aimed, for the first time in Israel, to assess national prevalence and incidence rates of dementia using EHRs. This study provides the basis for a national dementia database that will help shape policy in the future including the promotion of early dementia detection, improving caregiver support, management of co-morbid conditions and behavioral problems and improving overall dementia management.