Early detection of dementia enables more effective planning and can enable access to treatment and support. The Mini-Cog is a widely used screening instrument in Indonesia; however, this instrument has never undergone a translation and cultural adaptation process. Currently, there is no data on how accurate the tool is against diagnostic criteria, particularly in low-education.
Embedded within the community-based dementia prevalence study was the Strengthening Responses to Dementia in Developing Countries (STRiDE) project; older adults (aged ≥ 65 years) were randomly recruited from sites in Jakarta and North Sumatra, Indonesia. All participants were asked to complete the Mini-Cog and the 10/66 short dementia diagnostic schedule. The accuracy of three Mini-Cog algorithms (Mini-Cog1, Mini-Cog2, and Mini-Cog3) were compared against and the 10/66 short dementia diagnostic schedule. Additional analysis explored its performance accuracy at different educational levels.
The Mini-Cog test performance assessment was conducted on 2098 older adults
The area under the curve (AUC) of Mini-Cog1, Mini-Cog2, and Mini-Cog3 receiver operator characteristic (ROC) curves were 0.66, 0.62, and 0.64, respectively. All algorithms demonstrated high sensitivity (Sv) but low specificity (Sp). (Mini-Cog1: Sv 83.2%; Sp 49.2%, Mini-Cog2: Sv 87.1%; Sp 37.8% and Mini-Cog3: Sv 72.5%; Sp 56%). All algorithms showed no affected by education. Only 59.1% of people without dementia could do the CDT.
The high sensitivity of the Mini-Cog1 algorithm lends itself to screening purposes. Given that the specificity is still low, and less than 60% of patients without dementia can complete the CDT. Further research is needed, as is the development of screening instruments with high accuracy values in low- and middle-income countries, particularly in Indonesia.