8589356345

C 2017 British łflV Assoobtkm

DOI: iaim/hKf.12492 HIV MtScine(20\7\. 18. 573-579

0R1G1NAL RESEARCH

Who will be lost? Identifying patients at risk ofloss to foliow-up in Malawi. The DREAM Program Experienee

SMancindli,¹ K Nidscn-Saincs/ PGcrmano.* G Guido tti,* E Buonomo.¹ P Scarcclla.' R LungW,⁴ H Sangarc,⁴ S Orbndo/ G Liotta.¹ MC Marazzi* and L Palombi¹

'Department of Biomedieine and Prerention. Un»versity ofRome Tor Vcrgata. Romę. Italy, ^Department of Pediatrie*-Infeetious Diseasc, Unitrrsity of Cdlifomia Los Angeles, Los Angeles, CA. USA, ^JDREAM Programme, Romę, Italy, ⁴DRF.A\1 Programme, Blantyrc, Molowi and *LUMSA Untocrsity, Romę. Italy

Objcctivcs

Rctcntion of subjccts in H1V treatment programmes is crucial for the succcss of treatment We cvaluatcd rctcntion/loss to follow-up (LTFU) in subjccts rccdving established carc in Malawi.

Methods

Data for HIV-positivc patients registerrd in Drug Resourcc bnhanccmcnt Against AIDS and Malnuirition centres in Malawi prior to 2014 werr rcvicwcd. Visits entailing H!V testing/ counsclling. laboratory cvaluations, nufritional rvaluation/supplrmeritation, community support peer education, and antirctroviral (AR O monitoring/pharmacy were noted. LTFU was defined as > 00 days without an cncountcr. Parameters potcntially associatcd with LTFU were cxplorc<i with univariatc/multivariatc logistic rcgrrssion analyscs bcing performed.

Re suit s

Fiftccn thousand and nincty-ninc patients registered beforr 2014; 202 (1.3Gć) were lost to follow-up (LTFU) (1.3%). Ninę (0.5Gb) of 1744 paediatric patients were LTFU i^fs. 1.4Gb (n = 103) of 13 355 adults [P < 0.001). Subjccts who were LTFU had fewer days in care than retained subjccts (1338 vs.

1544, rcspectively; P < 0.001) and a longer duration of ART (1530 vs. 1300 days, respectively;

P < 0.001). Subjccts who were LTFU had higher baseline HI V viral loads (P = 0.016) and higher body mass indcxcs [P < 0.001), were morę likcly to livc in urban settings (80^ of patients who were LTFU livcd in urban settings) with better housing [rdative risk (RR) 2.3; 05Gb eonfidcncc intcrval (Cl) 1.67-3.00; P < 0.001), and werr morę likcly to be cducatcd (RR 1.88; 05Gb Cl 1.42-2.50; P < 0.001). Distance to the centrę and cost of trans po rtation were associatcd with LTFU (RR 3.4; 05Gb CI 2.84-5.37; P < 0.001), as was absence of a maternal figuir (RR 1.57; 05Gb CI 1.17-2.00;

P < 0.001). Viral load, distance indcx, education and a maternal figurę were predictive of LTFU.

Condusions

bducated, urbanized HlV-infected adults living far from programme ccntrcs are at high risk of LTFU, particularly if there is no maternal figurę in the houschold. Tłicsc variablcs must be taken into consideration when dcvcloping rrtention strategies.

Keywords: H1V, loss to follow-up. Malawi, prcdictors, rrtention Aecepted 17 Norentber 2016