Drug Abuse Patterns, Selected Psychosocial Conditions and Associated Risky Sexual Behavior a mong Women Who Inject Drugs Living in Informal Urban Settlements in Nairobi County Kenya

Abstract

An estimated 16 million people who inject drugs worldwide are female. In Kenya it is estimated that 18,327 people inject drugs and 10% are women with a HIV prevalence of 36%. Women who inject drugs (WWIDs) experience great disparities in health outcomes relative to their counterparts in the general population, most notably in HIV. HIV clusters together with intimate partner violence (IPV), substance abuse (SA), and depression among WWIDs. This thesis applied ecological approach guided by syndemics theory to determine drug use patterns, selected psychosocial conditions and associated risky sexual behavior among women who inject drugs living in informal urban settlements in Nairobi Kenya. The objectives of the study were to determine the prevalence of substance abuse, intimate partner violence, depression and risky sexual behavior, establish patterns of drug use, determine the co-occurrence of substance abuse, intimate partner violence, depression, risky sexual behavior and investigate the socio-demographic and socio-economic variables associated with substance abuse, intimate partner violence, depression and risky sexual behavior among women who inject drugs.This study used a cross sectional study design with a mixed method approach. 306 women, ≥18 years of age, and injecting heroin in the preceding year were recruited using targeted mobliser driven sampling. Statistical analysis software STATA version 15 was used for statistical analyses. Multiple methodologies including descriptive analyses, standard logistic regression, classification trees algorithm for predictive modelling were employed. Thematic analysis was used for qualitative data. The prevalence of SA, IPV, depression and risky sexual behaviour were 88%, 84%, 77.1% and 69.3% respectively. Persons who introduced drugs used at age of initiation of substance use was associated with current poly substance use (Fisher exact P=0.0001). There was a significant association between SA and depression and with risky sexual behavior. Each additional psychosocial condition was associated with 6-fold odds of having risky sexual behaviour. (P=0.0001). Standard logistic regression analyses returned three significant variables: SA*depression interaction effect, age of delivery of the first child and income. Classification tree modelling predicted SA, depression, time lived in informal settlement, type of family women grew in and number of children to have the highest influence on risky sexual behaviour. This thesis provides new evidence on prevalence of the psychosocial conditions and drug abuse patterns among WWIDs. Further, it presents evidence on individual and cumulative effects of IPV, depression, SA on risky sexual behaviour outcome and socio-demographic and socio-economic variables associated with the IPV, depression, SA and risky sexual behaviour among WWIDs. The findings of this study have great public health significance and important implications for further research, interventions, and policy.