|dc.description.abstract||Introduction: The global burden of insomnia is high in the general population but much higher in those affected by HIV. Untreated insomnia is prognostic of HIV outcomes and increases the risk for psychological comorbidity and cardiac events. In Uganda the burden and associated factors of insomnia among HIV positive patients in a primary care setting where the bulk of
consultations occur is not known. In addition, the tools used to assess insomnia have not been validated in our setting. Therefore this study aimed at determining the prevalence and associated factors of sleep disorders among the HIV population in Entebbe General Hospital, a primary care setting. It also validated the tools used to measure insomnia in the same setting.
Methods: A sample of 174 adult HIV positive patients attending the Entebbe General Hospital HIV clinic during February to April 2017 were assessed for insomnia sleep disorder using DSM5 criteria; poor sleep quality using the Pittsburg Sleep Quality Index (PSQI>5); severity of insomnia using the Insomnia Severity Index (ISI – sub-threshold – 8-14; moderate/severe>14). Strength of association for selected bio-psycho-social variables biological [Low CD4, viral load, type of medication, adherence; medical comorbidities e.g. hypertension, diabetes], psychological [stress, anxiety, mental illness e.g. depression; consumption of stimulants e.g. caffeine, alcohol, drugs] and social factors [sleep hygiene; quality of significant relationships; marital status; level of education; occupation] with each sleep measure was assessed using logistic regression. Cronbach’s alpha for each sleep
measure and Spearman Rank Correlation between each of the sleep measures were calculated. The diagnostic sensitivity for PSQI and ISI against DSM5 criteria was assessed using Receiver Operator Curves.
Results: 13% of the study population had insomnia sleep disorder. 46% experienced poor sleep quality - 23% categorized as sub threshold clinical insomnia and 7.2% as moderate/severe type. Internal reliability for each measure was high DSM5 (0.96); ISI (0.99), PSQI (0.88). Spearman Rank Correlation between DSM5 criteria with either PSQI (0.261 p=0.001) or ISI (0.239 p=0.003) was low. The PSQI had a fair ability to discriminate patients with and without ISD (AUC 0.71). The cut-off point of 5 correctly identified 57% of patients with ISD (Sensitivity 75% and Specificity 54.0%). The Insomnia Severity Index had a fair ability to discriminate patients with and without ISD (AUC 0.75). The cut off-point for sub-threshold insomnia (8) correctly identified 68% of patients with ISD (Sensitivity 60% and Specificity 69%); the cut-off point moderate/severe (15) insomnia correctly identified 80% of patients with ISD (Sensitivity 25% and Specificity 90%). Insomnia sleep disorder was not associated with any of the factors that were studied. Poor sleep quality was associated with generalized anxiety disorder (OR 3.02 CI 1.39-6.554 p <0.05) and use of Efavirenz (OR 0.44 CI 0.20-0.96 p <0.05). Moderate to severe insomnia was also associated with generalized anxiety (OR 3.82 CI1.80-8.11 p<0.05). The other factors that were studied – age, gender, marital status, income levels, living with a partner, CD4 levels, viral load, being on antiretroviral therapy (HAART), random blood sugar, blood pressure, depression, alcohol/substance use – were not associated with poor sleep quality or the severity of insomnia.
Conclusion: Given the high prevalence of insomnia sleep disorder, poor sleep quality with moderate/severe symptoms there is a need to incorporate screening and diagnostic questions on sleep in the routine consultation at primary care centres managing HIV. Screening should also include assessment for generalized anxiety disorder, which was the main explanatory factor. Because the DSM5, PSQI and ISI are poorly correlated, it may be necessary to conduct all three tests to come to a definitive
diagnosis about need for therapy.||en_US