THE VALUE OF HOUSEHOLD CONTACT INVESTIGATION IN CONTROL OF DRUG-SUSCEPTIBLE TUBERCULOSIS AMONG CHILDREN AGED 0-5 YEARS IN KISUMU COUNTY KENYA

ABSTRACT

Ninety percent (90%) of TB infections are self-limiting; 10% will progress into active or latent TB. Persons with latent TB cannot spread disease but can go on to become infectious cases later. The risk of TB infection and progression from latent infection to active infection is higher in children than in adults. A person with active TB can infect up to 15 persons yearly through close contact. Contact investigation is therefore recommended for close contacts of TB patients to identify undiagnosed cases of active and latent TB to initiate them on curative and preventive therapy respectively. The value of TB contact investigation in childhood TB control is unknown. This study determined the value of Contact Investigation in childhood TB control (persons aged< 5 years) in Kisumu County between 2014 and 2015 a period prior to the implementation of standardized contact investigation using the following methods used to address the specific objectives: To describe risk factors for TB infection among child contacts, a cross sectional survey was conducted on 183 TB index and 257 child contact using a symptom screen, Tuberculin Skin tests (TST) and Chest X rays. Cut off for TST positivity was identified using a bimodal peak for TB prevalence surveys. Hierarchical level modeling analyses, with children nested within households, was used to describe contact and index characteristics associated with TB infection. To describe Isoniazid Preventative Therapy (IPT) uptake among eligible child contacts, in a retrospective cross-sectional survey, data belonging to 337 child contacts was linked to TB program registers and logistic regression analyses used to describe individual characteristics associated with IPT initiation. To compare the value of contact investigation (intervention arm) to contact invitation (control arm) in contributing to childhood TB control, a cluster randomized trial was employed to compare TB cases diagnosed and children receiving IPT in the pre- intervention (2012-2013) and intervention (2014-2015) years, and the intervention years using a minimum sample size of 15 per arm. A convenience sampling technique was employed for the first 3 objectives. To describe lessons learnt implementing standardized contact investigation in a setting with routine contact invitation; a mixed method design was employed. The study recruited 345 child contacts linked to 243 TB index cases. Of these 39.3% child contacts with completed TST (n=257), had TB infection based on a TST cut off of 5mm for positivity. The risk of TB infection was higher among child contacts of TB index cases aged 15-45 years (OR =2.59, 95% CI 1.1-7.2) who had a cough (OR =3.76, 95% CI 1.51- 9.35) after fixing other contact and index characteristics. Eight (8; 2.3%) children had TB. The remainder (n=337) were IPT-eligible and the 15.1% that initiated IPT were more likely to be 1st degree relatives of the index case (OR =2.57, 95% CI 1.19-5.52) and to reside in a rural area (OR=2.65, 95% CI 1.37-5.09). Index cases with child contacts initiated on IPT (19.6%) were more likely have smear positive TB (OR =4.62, 95% CI 1.02-20.88) and to resides in rural area (OR 3.06, 95% CI 1.45-6.48). In the pre- and post- intervention years, TB cases increased by 20 (75% of them were from intervention arm). During the intervention years TB cases decreased by 17 (71% were from the control arm); the intervention arm contributed to 100% and 75% of the children put on IPT in 2014 and 2015. There was inadequate diagnostics, treatment and data support for contact investigation. Contact investigation enhanced childhood TB control in comparison to routine approaches and identified factors to prioritize contacts for screening, but low IPT initiation rates negated its intended benefits. Adequate investments need to be made in infrastructure to support contact investigation. Household contact investigation should target populations at high risk of infection; IPT implementation be supported using linked index- contact registers and TB contact investigation be ingrained within community health systems