Table 1: Description of each study in Japan.

Study name/site (study years)Sampling frameAssessment protocolPrevalence of all-cause dementia among those aged 65 and older:
(1) Prevalence in percentage (95% CI)
(2) Number of cases (all subtypes combined/AD/VaD)
(3) AD percentage out of all dementia
(4) AD/VaD ratio

Hisayama
(1985, 1992, 1998, and 2005)
Total residents aged ≥ 65 in the area:
Year 1985:
Year 1992:
Year 1998:
Year 2005:
Phase 1
Suspected dementia cases selected using Hasegawa dementia scale, neuropsychological tests, and ADL
Year 1985
  (1) 6.7 (5.0–8.3)
  (2) 59/12/21
  (3) 20.3%
  (4) 0.57
Study participants at Phase 1 (participation rates)
 Year 1985: (94.6%)
 Year 1992: (96.6%)
 Year 1998: (99.7%)
 Year 2005: (91.5%) 
Sampling weights used to reflect prevalence of population at large?
(Yes/No): No
Phase 2
DSM-III-R, Karasawa’s criteria (supplementation), and Hachinski’s evaluation scale for cerebral ischemia in case vascular effect on dementia were considered to be too ambiguous to diagnose as VaD using the following guidelines:
 Hachinski’s ischemic score
 <4.0—AD
 ≥8—VaD
 5, 6, 7—mixed/other
Year 1992
  (1) 5.6 (4.4–7.1)
  (2) 68/21/22
  (3) 30.8%
  (4) 0.95
Year 1998
 (1) 7.1 (5.7–8.5)
 (2) 102/49/25
 (3) 48.0%
 (4) 1.96
Year 2005
  (1) 12.5 (10.7–14.2)
  (2) 195/96/51
  (3) 49.2%
  (4) 1.89

Okinawa (1991-1992) Randomly sampled one city from the urban districts and one town/village from the rural districts from each of 5 regions covering the entire Okinawa prefecture. Randomly selected approximately 17% of the residents from the selected cities and towns/villages in each region.Phase 1
Pilot study was conducted ( ) to examine the best MMSE cut-point for screening. MMSE score of 16 had the maximum combination of sensitivity and specificity to identify the demented in the community
Selected those with MMSE ≤16 ( )
(1) 6.7 (3.6–7.8)
(2) 170/80/53
(3) 47.1%
(4) 1.51
Study participants at Phase 1:
  
Sampling weights used to reflect prevalence of population at large?
(Yes/No): Yes
Phase 2 ( )
DSM-III-R for dementia, NINCDS-ADRDA for AD, and Hachinski’s ischemic score as a guideline for VaD

Hiroshima (Radiation effect research foundation Adult Health Survey ((RERF-AHS)) (1992–1996)Residents in Hiroshima among the Original AHA cohort (atomicbomb survivors in Hiroshima and Nagasaki and their controls followed since 1958) evaluated by biennial physical exams between 1992 and 1996. Targeted Phase 1
Subjects with CASI ≤ 75 ( ) and controls with CASI > 75 ( ) were selected
(1) 8.5 (7.2–9.8)
(2) 156/74/40
(3) 47.4%
(4) 1.85
Study participants at Phase 1:
  
Sampling weights used to reflect prevalence of population at large?
(Yes/No): No
Phase 2 ( )
DSM-III-R for dementia, NINCDS-ADRDA for AD, ADDTC for ischemic vascular dementia, and DSM-III-R for VaD

Tajiri Project (1998) All residents in Tajiri town aged ≥ 65, targeted All ( ) were evaluated by CDR & DSM-IV (not multistage sampling) (1) 8.5 (7.2–9.9)
Based on subtype analysis with MRI
(I) Using NINCDS-ADRDA, NINDS-AIREN
 (2) 32/20/6
 (3) 62.5%
 (4) 3.33
Study participants Subsample selected for dementia subtype identification study, targeted   
Sampling weights used to reflect prevalence of population at large?
(Yes/No): No
MRI study: 564 selected randomly from 1654 above, of whom 497 participated in MRI study (dementia subtype identification study). Comparisons of prevalence of VaD using 3 different criteria: (1) DSM-IV for AD and VaD; (2) NINCDS-ADRDA for probable AD, NINDS-AIREN for possible AD with CVD and probable VaD; (3) ADDTC for probable ischemic vascular dementia(II) Using DSM-IV for AD and VaD
  (2a) 32/13/13 (2b) 32/18/8
  (3a) 40.6% (3b) 56.2%
  (4a) 1.00 (4b) 2.25
The difference between (a) and (b) above is due to diagnostic differences between assessors

Ama-cho (2008) All residents in Ama-cho aged ≥ 65, targeted Phase 1 ( )
Suspected dementia cases selected through interviews with subjects and their informants, which assessed cognitive changes, psychiatric symptoms, personality changes, problem behaviors, activities of daily living, psychological and medical symptoms, and through assessments of the subjects’ medical history offered by the home doctors of the subjects
(1) 11.3 (9.1–13.2)
(2) 104/66/16
(3) 63.5%
(4) 4.12
Study participants (after excluding 23 subjects out of 943 subjects who resided out of town at the time of phase 1 interview) 
Sampling weights used to reflect prevalence of population at large?
(Yes/No): No
Phase 2 ( )
DSM-IV for dementia, NINCDS-ADRDA for AD, and NINDS-AIREN for VaD

CASI: The Cognitive Abilities Screening Instrument [19].