The NaSDEK calculated national standardized dementia prevalence by adjusting crude prevalence according to age, sex, education, and region-applied sample weights, based on the national population census survey data obtained by Statistics Korea during the corresponding time period. It reported the estimated dementia population, standardized dementia, mild cognitive impairment (MCI) prevalence, and standardized dementia prevalence according to the Clinical Dementia Rating (CDR) stage. The NaSDEK also reported the estimated dementia prevalence at a specific point in time using standardized dementia prevalence at the time of publication (Table 1). The NaSDEK was compiled using data from older Korean adults who were randomly sampled from 15 administrative districts across the country [7]. The population was selected by region-based, multi-staged sample extraction, which was performed primarily in hospital settings rather than in nursing homes or nursing hospitals; therefore, the reported prevalence is likely an underestimation. Nonetheless, the methodology used in the NaSDEK was substantial enough to deem the prevalence data representative.

The NaSDEK was conducted in 2008 [8], 2012 [9], and 2016 [10], and reported the point prevalence rate of dementia, which is the prevalence of dementia at a particular point in time. While both the 2008 NaSDEK and the 2012 NaSDEK were conducted on older adults aged 65 years or older, the 2016 NaSDEK surveyed populations were aged 60 years and above. This broadening of the age criterion was done to standardize the age threshold of government-led dementia management projects. The latest survey was conducted using 5,056 participants with a two-stage design, in which the 1st stage screened for dementia using the Mini-Mental State Examination for Dementia Screening (MMSE-DS) criteria [11], and the 2nd stage diagnosed dementia using clinical criteria laid out by the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD-K) [12] and CDR [13]. A diagnosis of dementia was made according to the Diagnostic and Statistical Manual of Mental Disorders, fourth edition [14]. For the populations diagnosed with dementia, subtyping of dementia into Alzheimer dementia (AD) and vascular dementia (VD) was performed according to the National Institute of Neurological and Communicative Disorders and Stroke, the Alzheimer’s Disease and Related Disorders Association criteria [15], and the Association Internationale pour la Recherche et l’Enseignement en Neurosciences criteria [16]. MCI was defined according to the revised consensus criteria of the International Working Group for MCI [17].

The estimated number of patients with dementia was 703,968, which accounted for 9.95% of the cases of dementia among older adults aged 65 years and older reported in the 2016 NaSDEK. Raw data at each NaSDEK demonstrated a continuous increase in dementia prevalence. However, a recent report directly compared the 2008 and 2016 NaSDEKs using identical populations, as well as structure-based age- and sex-standardizations (based on the 2015 population structure obtained from the National Statistical Office in Korea), and concluded that the prevalence of dementia had decreased, although not significantly [7].

In 2016, the prevalence rate of dementia by subtype was 7.07% for AD, 0.83% for VD, and 1.60% for other types of dementia. From 2008 to 2016, VD prevalence demonstrated a continuous decrease, while the prevalence of AD increased; as a result, the 2016 NaSDEK reported that the AD/VD ratio measured approximately 8.5:1 among older adults. This dissociating trend in prevalence between VD and AD has been reported in previous studies. Prevalence in the Japanese population demonstrated an increase in AD and no change in VD [18]. A systematic review of dementia prevalence in Korea showed that the AD/VD ratio changed from 1.96 in the early 1990s to 4.13 in the late 2010s [5]. The review also suggested that the prevalence of severe dementia has decreased, and that of milder forms of dementia has increased.

The distribution of the severity of dementia in 2016 was as follows: 48.4% had questionable dementia (CDR 0.5), 27.6% had mild dementia (CDR 1), 21.9% had moderate dementia (CDR 2), and 2.1% had severe dementia (CDR 3) [7]. Increasing proportions of mild dementia accounted for the majority of cases, suggesting that early screening for dementia and appropriate medical care are of increasing importance. On the other hand, MCI showed inconsistent trends, which might be due to more rigorous diagnostic methods for subjective cognitive decline.

Each NaSDEK reported an estimated future prevalence prediction for the reporting year. The 2008 NaSDEK predicted a dementia prevalence of 9.08%, which was close to the actual dementia prevalence of 9.18% in 2012. However, the predicted estimated prevalence increased significantly with time, with prediction values increasing to 13.17% in 2008, 15.06% in 2012, and 16.09% in 2017. This is thought to be due to an increase in the older adult population secondary to increasing life expectancy. Considering the socioeconomic costs of dementia, effective strategies and policies for diagnosis and treatment are required at the national level.

The KDO report has been published annually from 2017 to 2020 by the Ministry of Health and Welfare and the NID [19-22]. Every KDO presents two different numbers of dementia patients: the estimated number of dementia patients, and the actual number of patients diagnosed with dementia. First, the estimated number of patients with dementia is calculated in a similar way to that in the latest NaSDEK. In other words, the KDO of a specific year uses the NaSDEK 2016 data and population census values for a specific year. Thus, the 2017 KDO values are similar to those of the 2016 NaSDEK. Second, the number of patients with a diagnosis of dementia is obtained from patients whose hospital medical records were confirmed by the NHIS to contain one of six disease codes (F00, F01, F02, F03, G30, G31) based on the WHO’s Global Dementia Observatory. In addition, the number of patients with MCI was obtained from patients whose medical records were confirmed by the NHIS to contain the disease code F067. Both the estimated number of dementia cases and the actual number of diagnosed patients with dementia increased continuously between the 2017 KDO and the 2020 KDO. Similarly, the prevalence of MCI has displayed a consistent increase (Table 2).

There are large differences between the 2017 KDO and the remaining KDOs as a result of differences in data sources; that is, the 2017 KDO used the severity ratio from the 2016 NaSDEK, while the other KDOs used the ratio from the 2012 NaSDEK. Thus, caution is needed when comparing trends in dementia severity from one KDO to another.

“Dementia Today” reports prevalence using data obtained from the NID website (available at https://www.nid.or.kr/info/today_list.aspx). It periodically updates the number of diagnosed patients with dementia, similar to the KDO. Therefore the latest prevalence data and the number of patients with a diagnosis of dementia are reported and are representative of the population of a given year, using the standardized dementia prevalence parameters established in the 2016 NaSDEK.

We have summarized the estimated number of dementia cases and prevalences from the NaSDEK and NID. The projected values of future figures are estimated for each report (Figs. 1, 2).

The NHIS is a universal, mandatory, single-payer, nationwide health insurance system that covers approximately 100% of the Korean population [23]. As previously mentioned, collected data represents patients with a dementia diagnosis code who received inpatient or outpatient treatment. Variations with regards to diagnosis code definitions, primary, and secondary diagnoses are present.

In contrast to the data reported in the KDO and “Dementia Today,” the NHIS report [24] calculated the number of dementia patients based on the number who received a KCD diagnosis code for at least one of a multitude of dementia diagnoses: dementia in Alzheimer disease (F00, F000, F001, F002, F009), Alzheimer disease (G30, G300, G301, G308, G309), vascular dementia (F01, F010, F011, F012, F013, F018, F019), alcoholic dementia (F107), frontotemporal dementia and Pick’s disease (G3100), semantic variant primary progressive aphasia (G3101), dementia in other diseases classified elsewhere (F02), dementia in Pick’s disease (F020), dementia in Creutzfeldt-Jacob disease (F021), dementia in Huntington’s disease (F022), dementia in Parkinsonism (F023), dementia in human immunodeficiency virus disease (F024), dementia with Lewy bodies (G3182, F028), dementia in other specified disease classified elsewhere (F028), unspecified dementia (F03), and delirium superimposed on dementia (F051). It also includes codes applicable to extended health insurance claims for severe dementia: nonfluent primary progressive aphasia (G3102), logopenic primary progressive aphasia (G3103), NEC and progressive isolated aphasia (G3104).

Table 3 shows the number of patients with dementia as both the primary and secondary diagnosis. For the purpose of comparison, Table 3 contains values from other data sources in the year. The number of people diagnosed with dementia is rapidly increasing, which is thought to be due to the increasing interest in dementia by both the medical field and the general public. In comparison to the NaSDEK and KDO, the number of patients with a primary diagnosis of dementia was lower across all years. On the other hand, the number of patients who were diagnosed with dementia as the primary or secondary diagnosis was lower in 2008 and increased significantly after 2012 compared to that from the NaSEDEK and KDO data. This suggests that the actual number of patients with dementia may be much higher than previously believed, requiring a more proactive approach for diagnosing patients with subclinical or less apparent dementia.

Another method for calculating the dementia population uses injury codes applied to special cases for the Copayment Decreasing Policy: injury codes V800 (ICD codes F001, F002, F010, F011, F012, F013, G301) and V810 (F000, F020, G300, G3100, G3101, G3102, G3103, G3014, G3182) [25]. This method produced similar results to the estimated numbers obtained using cases with a primary diagnosis code of dementia (Table 3).