DOI: 10.1002/alz.076485 ISSN: 1552-5260

An analysis of psychosocial predictors of cognitive impairment in older people living in social housing ‐ data from longitudinal study: “Porto. Importa‐se”

Sónia Martins, Hélder Alves, Idalina Machado, Sara Melo, Joana Guedes
  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Geriatrics and Gerontology
  • Neurology (clinical)
  • Developmental Neuroscience
  • Health Policy
  • Epidemiology



Cognitive Impairment (CI) is a growing public health problem among the older population, with well‐established risk factors. However, the role of psychosocial predictors still needs to be clarified. This study aimed to analyze whether psychosocial factors may be predictors of CI in older persons.


In the framework of “Porto. Importa‐se” longitudinal project, this study included older persons (≥70y.o. living alone and couples ≥75y.o.), living in social housing in Porto city, Portugal. The research protocol included: Six‐item Cognitive Impairment Test (CI), Barthel and Lawton Index (Basic/Instrumental Activities of Daily Living‐BADLs/IADLs), Geriatric Depression Scale‐15 (depressive symptoms) and Lubben Social Network Scale‐6 (social isolation). Loneliness was assessed with a categorical question: Do you feel lonely?. A logistic regression analysis was conducted to identify predictors of CI. Odds Ratio (OR) and its 95% Confidence Interval (95%CI) were calculated. A p<0.05 was considered statistically significant.


Of the 1213 older persons (mean age 81±5.7) included, most were female (74%) and 36% had low education level (<4years). About 11% were severely dependent on IADLs and 27% presented CI. Around 35% of participants were at risk of social isolation, 54% reported loneliness and 48% showed depressive symptoms.

The logistic regression model was statistically significant [X2(5,N = 1213) = 148.596,p<.001], explaining 19.7%(Nagelkerke R2) of variance in CI (correctly classified 77% of cases). The adjusted model had a moderate sensitivity (63%) and specificity (78%). The oldest persons (≥80 years old) were 1.77 as likely to have CI than younger (OR = 1.77, 95%CI[1.3, 2.4]). Those with less education (<4years) were 0.26 times as likely to present CI than reference group (≥4years) (OR = 0.26, 95%CI[0.2, 0.4]) and females 1.5 times more than males (OR = 1.50, 95%CI[1.0, 2.2]). Older persons at risk of social isolation were 1.85 as likely to have CI (OR = 1.85, 95%CI[1.4, 2.5] and those with depressive symptoms were 1.39 likely to present CI (OR = 1.39, 95%CI[1.0, 1.9]).


Older age, female gender, low education, depressive symptoms and social isolation were predictors of CI. The existence of potentially modifiable factors (e.g., social isolation, depression), means that CI prevention is possible through the implementation of effective individual and community‐level interventions, aiming to protect older persons from CI.

More from our Archive