Introduction The increase in life expectancy due to epidemiologic transition has brought about increased numbers of certain mental illnesses, namely depression. There is scarcity of analytical base epidemiological studies on this issue. This study was conducted to identify factors associated with depression among elderly (>60) in Karachi, Pakistan.

Method A multicentre matched case control study was conducted. A total of 234 depressed cases and 468 non-depressed controls (1case:2controls) were selected from four tertiary care hospitals in Karachi. Cases were recruited from psychiatric OPD whereas; controls were recruited from other OPD's (except psychiatric) and were individually matched on age and gender with cases. Ascertainment of cases and controls was done by psychiatrist and Geriatric Depression Scale. There was substantial agreement between the psychiatrists for diagnosing depression (κ: 0.66). A pilot tested structured questionnaire was administered and analysis was performed through conditional logistic regression using SPSS 11.5.

Results Risk factors for depression were, living in nuclear family system, (mOR: 4, CI 2.1 to 4.9), self reported difficult situation during past 1 year (mOR: 9, CI 6.0 to 19.7), hearing and/or visual impairment (mOR: 5, CI 2.0 to 9.8), presence of >2 NCD (mOR: 2, CI 1.0 to 3.4). However, interacting with friends or relatives had a protective effect against depression (mOR adj: 0.3, 95% CI 0.3 to 0.7).

Conclusion Interventions should be taken to reduce modifiable factors and non-modifiable factors should be used to identify high risk group for primary prevention. However factors identified through this study may vary in its association with depression across different communities.