Determinants of semantic and episodic memory decline among older adults in Ghana: Evidence from the WHO study on global AGEing and adult health Ghana wave 2

Objective: Determined factors associated with semantic (SM) and episodic memory (EM) among older adults aged 50 years and older in Ghana. Methods: Data from WHO Study on Global AGEing and Adult Health (SAGE) Ghana Wave 2 was used for this study. Semantic memory (SM) and Episodic memory (EM) were the main study outcomes separately. The study employed Nested Ordinary Least Square regression analysis by sequentially adding 6 blocks of variables and comparison tests between the nested models. Results: The study involved 3575 adult Ghanaians aged 50 years and older with a mean ± standard deviation of 62.6 ± 18.4 years. The overall mean ± SD of EM and SM were 5.86 ± 2.51 and 11.69 ± 8.59 respectively. Overall, analysis from block 6 showed a significant variation in SM by approximately 16.9%(ΔR² = 1.17%) where increasing age, never married (β = −1.55; 95% CI = −2.41–0.69), being resident in Greater Accra (regional disparity) (β = −3.45; 95% CI = −4.73–2.20), underweight (β = −0.81;95% CI = −1.34–0.27), and moderate self-rated health (SRH) (β = −0.98; 95% CI = −1.52–0.45) significantly decreased SM. Similarly, increasing age, separated/divorced (β = −0.22; 95% CI = −0.35–0.87), being resident in Greater Accra (β = −0.53; 95% CI = −0.80–0.26), and moderate SRH (β = −0.20; 95% CI = −0.36–−0.04) significantly decrease EM with an overall significant variation of approximately 22.9%(ΔR² = 2.7%). Conclusions: Increasing age, sex, marital status, regional disparity, and poor SRH significantly decreased both Semantic memory and Episodic memory. Higher educational attainment and life satisfaction significantly influenced SM and EM. These provide pointers to important socio-demographic determinants of SM and EM with implications for the implementation of the Ghana national ageing policy 2010, ‘ageing with security and dignity’, and as a key consideration for healthy ageing towards 2030.