UBC Faculty Research and Publications

Advanced HIV disease at presentation to care in Nairobi, Kenya: late diagnosis or delayed linkage to… van der Kop, Mia L; Thabane, Lehana; Awiti, Patricia O; Muhula, Samuel; Kyomuhangi, Lennie B; Lester, Richard T; Ekström, Anna M Apr 18, 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12879_2016_Article_1500.pdf [ 677.67kB ]
JSON: 52383-1.0366830.json
JSON-LD: 52383-1.0366830-ld.json
RDF/XML (Pretty): 52383-1.0366830-rdf.xml
RDF/JSON: 52383-1.0366830-rdf.json
Turtle: 52383-1.0366830-turtle.txt
N-Triples: 52383-1.0366830-rdf-ntriples.txt
Original Record: 52383-1.0366830-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessAdvanced HIV disease at presentation tocare in Nairobi, Kenya: late diagnosis ordelayed linkage to care?—a cross-sectionalstudyMia Liisa van der Kop1,2*, Lehana Thabane3, Patricia Opondo Awiti1, Samuel Muhula4, Lennie Bazira Kyomuhangi4,Richard Todd Lester2† and Anna Mia Ekström1,5†AbstractBackground: Presenting to care with advanced HIV is common in sub-Saharan Africa and increases the risk ofsevere disease and death; however, it remains unclear whether this is a consequence of late diagnosis or a delay inseeking care after diagnosis. The objectives of this cross-sectional study were to determine factors associated withadvanced HIV at presentation to care and whether this was due to late diagnosis or delays in accessing care.Methods: Between 2013 and 2015, adults presenting to care were recruited at two clinics in low-income areas ofNairobi, Kenya. Participants were considered to have advanced HIV if their CD4 count was below 200 cells/μL, orthey were in WHO stage 4. Information on previous HIV diagnoses was collected using interviewer-administeredquestionnaires. Logistic regression was used to determine the association between clinical and socio-demographicfactors and advanced HIV.Results: Of 753 participants presenting to HIV care, 248 (33 %) had advanced HIV. Almost 60 % (146/248) of thosepresenting with advanced HIV had been previously diagnosed, most of whom (102/145; 70 %) presented to carewithin three months of their initial diagnosis. The median time to presentation to HIV care after an initial diagnosiswas 22 days (IQR 6-147) for those with advanced HIV, compared to 19 days (IQR 4-119) for those with non-advanced HIV (p = 0.716). Clinic (adjusted odds ratio [AOR] 1.55, 95 % CI 1.09–2.20) and age (AOR 1.72 per unitincrease in age category, 95 % CI 1.45–2.03) were associated with presenting with advanced HIV.Conclusions: Presentation to care with advanced HIV was primarily due to delayed diagnosis, rather than delayedlinkage to care after diagnosis. Variation by clinic suggests that outreach and other community-based efforts maydrive earlier testing and linkage to care. Our findings highlight the ongoing importance of implementing strategiesto encourage earlier HIV diagnosis, particularly among adults 30 years and older.Keywords: HIV/AIDS, Sub-Saharan Africa, Kenya, Advanced HIV, Presentation to HIV care, Informal settlements* Correspondence: miavanderkop@gmail.com†Equal contributors1Department of Public Health Sciences/Global Health (IHCAR), KarolinskaInstitutet, Widerströmska Huset, Tomtebodavägen 18A, Stockholm 171-77,Sweden2Department of Medicine, University of British Columbia, 828 West 10thAvenue, Vancouver, BC V5Z 1M9, CanadaFull list of author information is available at the end of the article© 2016 van der Kop et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.van der Kop et al. BMC Infectious Diseases  (2016) 16:169 DOI 10.1186/s12879-016-1500-8BackgroundRationaleDespite increased HIV testing, improved global accessto antiretroviral therapy (ART), and changes in WorldHealth Organization (WHO) recommendations to ini-tiate treatment earlier, the problem of presenting tocare with low CD4 counts persists across sub-SaharanAfrica. Those who present to care with a CD4 countbelow 200 cells/μL, or an AIDS-defining event, areconsidered to have advanced HIV disease [1]. Individ-uals who initiate care with advanced HIV are morelikely to have impaired immune recovery [2] and re-duced life expectancy [3]. Presenting with advancedHIV may also result in higher direct medical and so-cietal costs [4, 5], as well as an increased risk of on-ward transmission [6]. Furthermore, high proportionsof individuals presenting with advanced HIV will im-pede the UNAIDS 90-90-90 targets to have 90 % ofthose living with HIV aware of their status; 90 % ofthose diagnosed on treatment; and 90 % of those ontreatment virologically suppressed [7]. Understandingrisk factors for presentation to care with advancedHIV is critical to developing strategies to encourageearlier linkage to care and successful therapy.While sub-Saharan Africa is disproportionately af-fected by the HIV epidemic and persons present to carewith significantly lower CD4 counts than in other set-tings [8, 9]; the majority of research on presentation tocare with advanced HIV has been conducted in higher-resource settings. Of studies in sub-Saharan Africa, sev-eral relied on routine clinical data, and although thesestudies contributed important insights into presentationto care with advanced HIV disease, they were unable tocapture data on important variables and suffered fromlarge amounts of missing data [10–12]. Other studieshave been limited by small sample size [12, 13], andthere have been contradictory findings on some ofthe important determinants of presenting with ad-vanced HIV, for example, age and alcohol use [9, 11,12, 14, 15]. Furthermore, previous research in the re-gion lacked data on timing of HIV diagnosis. Therehas been a strong call to determine whether it is adelay in diagnosis or a delay in seeking care afterdiagnosis that leads to presentation to care with ad-vanced HIV [9, 10, 16, 17]. Here, we conduct a cross-sectional study of persons presenting at two clinics inlow-income areas of Nairobi, Kenya to evaluate thepathway to presentation to care with advanced HIVdisease and its associated factors.Objectives(1)Quantify the proportion of individuals who firstpresent to care with advanced HIV.(2)Determine whether presenting to care withadvanced HIV was due to delayed diagnosis or adelay in seeking care after diagnosis.(3)Determine factors associated with first presentationto care with advanced HIV.MethodsStudy designThis cross-sectional study used baseline data collectedduring a randomized controlled trial and supplemen-tary cohort study. The trial involves evaluating the ef-fectiveness of a text-messaging intervention toimprove retention in early HIV care [18]. Patientswho did not fulfil phone-related eligibility criteria forthe trial were invited to participate in a supplemen-tary cohort study to examine patient retention duringthe first year of HIV care. Adults testing positive forHIV at two comprehensive care clinics in Nairobi,Kenya were assessed for study eligibility.Study setting and participantsBetween April 2013 and June 2015, participants were re-cruited from the Kibera Community Health Centre, anAmref Health Africa clinic located in a large informalsettlement. At this comprehensive care clinic, there areno direct patient costs for HIV care and treatment. Thepopulation the clinic serves lacks or has minimal accessto services such as education, water, sanitation, or otherpublic services. HIV prevalence among adults tested forthe first time is estimated at 13 % [19]. In March 2014,recruitment began at a second comprehensive careclinic, the Baba Dogo Health Centre, which is situated inanother large informal settlement in Nairobi’s Eastlandsarea and operated by the Kenya AIDS Control Project.At each clinic, clinical staff introduced potentialparticipants to a research nurse, who completed aneligibility assessment. Patients were eligible to partici-pate in the study if they were 18 years old or older,HIV-positive, and willing to provide informed con-sent. Patients previously assessed for ART eligibility,with prior ART exposure, or on ART were excluded.Women known to be pregnant were also excluded.Patients were screened for study participation at thetime of a positive HIV diagnosis, although potentialparticipants had one week to decide whether to enrollin the study. Screened patients were a mixture ofthose who: 1) presented to the clinic for HIV testingand counselling (HTC) services; 2) sought treatmentfor an illness and then the clinician referred them toHTC; 3) had been diagnosed with HIV elsewhere andpresented to the clinic specifically to receive HIVcare.van der Kop et al. BMC Infectious Diseases  (2016) 16:169 Page 2 of 9OutcomesPresentation with advanced HIV diseasePresentation with advanced HIV disease was defined aspresenting with a CD4 count <200 cells/μL or at WHOstage 4, regardless of CD4 count. This definition is basedon the consensus definition of advanced HIV disease,which is presenting with a CD4 count <200 cells/μL oran AIDS-defining event, regardless of CD4 count [1].WHO Stage 4 was used as a proxy for AIDS-definingevents because of the overlap between AIDS-definingconditions and the clinical events comprising WHOstage 4 [20]. Presentation to care was defined as “attend-ance at a health care facility that is able to monitor pro-gression of HIV infection and initiate appropriatemedical care, including ART, as appropriate” [1].Delay in seeking careA delay in seeking care was defined as presenting to caremore than three months after a previous HIV diagnosis.Determinants of presenting with advanced HIV disease andpotential effect modifiersVariables were selected if there was prior strong evidenceof their association with advanced HIV disease at presen-tation, e.g., sex (male or female) [9, 10, 14] and education(some secondary versus no secondary) [9, 12, 14]; or if evi-dence was conflicting, e.g., age (<30; 30–39; 40–49;≥60 years) [9, 14, 21], travel time to the clinic (<30, 30–59,≥60 min) [9, 14], and alcohol use (hazardous drinking ver-sus non- hazardous drinking, [9, 12, 15] identified by theAUDIT-C questionnaire score) [22]. We also investigateda novel individual-level variable that may be a factor inpresenting with advanced HIV, current illicit drug use(within 30 days of the baseline visit) e.g., heroin, cocaine,etc. Since clinics are operated by different organizationsand serve different populations, clinic attended was alsoconsidered (Baba Dogo v. Kibera). A priori informationwas not available on potential interaction betweenfactors associated with presentation to care with ad-vanced HIV; therefore, we explored plausible inter-action between sex and travel to time to clinic, as aninteractive effect has been found in studies investigat-ing retention in HIV care [23, 24].Data sources and measurementAt the baseline visit, the research nurse administered aquestionnaire in the participant’s language of choice,English or Kiswahili. Prior to starting recruitment, thequestionnaire was translated from English to Kiswahili,back-translated, and pre-tested with clinic patients (n =10). The questionnaire collected information on demo-graphic characteristics, HIV testing history, and sub-stance use. Blood was drawn at the baseline visit forlaboratory CD4 testing. HIV and CD4 testing wereconsistent with routine clinical practice. Data were en-tered in Microsoft Access on a weekly basis. Verificationprocedures included cross-checking data files with ori-ginal forms and clinical records, as well as range andconsistency checks.Study sizeA conservative rule is that logistic regression modelsshould have 10 outcome events per predictor variable tobuild stable models [25]. A preliminary descriptive ana-lysis indicated that there were 152 events of presentationto care with advanced HIV in this cohort, [26] whichwas adequate to build stable models with the six selectedfactors.Statistical methodsDescriptive analyses of the study population, includingthe proportion of patients presenting with advancedHIV disease, were conducted in SPSS v14. To com-pare the time to presentation to care between ad-vanced HIV and non-advanced HIV groups (for thosewith a previous diagnosis), a Mann–Whitney U testwas used. Analyses were restricted to individuals withcomplete data.Logistic regression was used to determine factors asso-ciated with advanced HIV at presentation to care. First,univariable analyses were performed to assess thestrength of the association between each factor and theoutcome. Variables were then included in an initial mul-tivariable model if they had a univariable p-value of≤0.25 or were considered important based on prior evi-dence (i.e., sex). In the final adjusted models, variableswere selected based on a significance threshold of p <0.05. Nested models were compared using likelihood ra-tio tests to examine interaction between sex and traveltime, and to determine whether to include a linear effector indicator variables for ordered categorical variables.The fit of the final model was tested with the Hosmer-Lemeshow goodness-of-fit test [27]. Results are pre-sented as estimated odds ratios (OR) and adjusted ORs(AOR) with corresponding 95 % confidence intervals(CI) and p-values. All p-values are two-sided and re-ported to three decimal places with those less than 0.001reported as p < 0.001. Analyses were performed usingStata version 12 (Statacorp, College Station, TX).EthicsThe study protocol was approved by the University ofBritish Columbia’s Clinical Research Ethics Board(H12-00563) and Amref Health Africa’s Ethics andScientific Review Committee (P40/12).van der Kop et al. BMC Infectious Diseases  (2016) 16:169 Page 3 of 9ResultsStudy populationBetween April 2013 and June 2015, 1068 HIV-positiveindividuals presenting to the Baba Dogo and KiberaHealth Centres were screened for study participation,and 775 were recruited (Fig. 1). The most frequent rea-sons patients were ineligible to participate were previousenrolment in HIV care (n = 160/262; 61 %) and preg-nancy (n = 88/262; 34 %). Less than 3 % of screened par-ticipants (n = 31/1068) declined participation. Of the 775participants recruited, baseline CD4 data were availablefor 97.2 % (n = 753/775) of the cohort.The mean age of participants was 34 years (standarddeviation 9.82), and males comprised 40 % of the cohort.Additional demographic and clinical characteristics aresummarized in Table 1. The median baseline CD4 countwas 302 cells/μL (IQR 148-463); 60.7 % (457/753) had aCD4 count lower than 350 cells/μL. Approximately 1/3(n = 248/753; 32.9 %) of the cohort presented to carewith advanced HIV (CD4 count <200 cells/μL or WHOstage 4).Late diagnosis versus delayed presentation to careOf those who presented to care with advanced HIV, 146(59 %) had been previously diagnosed with HIV. Thiswas similar to the proportion of those with a previousdiagnosis in the non-advanced HIV group (n = 306/505;61 %; chi-square p-value 0.650). Most participants withadvanced HIV presented to care within three months oftheir initial diagnosis (102/145; 70 %), including 44 indi-viduals who presented within one week. Data on thedate of first HIV diagnosis was missing for one partici-pant. The median time to presentation to HIV care afteran initial diagnosis was 22 days (IQR 6-147) for thosewith advanced HIV, compared to 19 days (IQR 4-119)for those with non-advanced HIV (p = 0.716).Factors associated with presentation to care withadvanced HIVTable 2 shows the association between clinical andsociodemographic characteristics and presenting to carewith advanced HIV. In both univariable and multivari-able analyses, age was linearly associated with presentingto care with advanced HIV, with a final AOR of 1.72(95 % CI 1.45 to 2.03) per unit increase in age category,compared to the reference category of <30 years. Indi-viduals presenting to the Baba Dogo clinic were morelikely to present with advanced HIV (AOR 1.55; 95 % CI1.09–2.20) than those at the Kibera clinic. Those withsome secondary education were less likely to presentwith advanced HIV; however this association was of bor-derline significance in the final model (AOR 0.73; 95 %CI 0.53–1.03). In the univariable analysis, male sex ap-peared to be associated with presenting with advancedHIV; however, this effect diminished in the multivariableanalysis and did not remain in the final model.DiscussionKey resultsIn this cohort of individuals presenting to HIV care attwo clinics in Nairobi, Kenya, approximately one-thirdpresented to care with advanced HIV, suggesting import-ant opportunities still exist to encourage earlier diagno-sis and treatment. We found that delayed diagnosis wasmore common than delayed linkage to care in explainingpresentation to care with advanced HIV. Although 59 %of those presenting with advanced HIV had been previ-ously diagnosed with HIV, almost ¾ of these individualspresented to care within three months of their initialdiagnosis. Given the average rate of decline of CD4 Tlymphocytes [28, 29], it is unlikely that many of the indi-viduals who presented to care with advanced HIV withinthree months of their previous diagnosis would have hadnon-advanced HIV when they were initially diagnosed.Overall, there was a strong linear increase in the likeli-hood of presenting with advanced HIV among agegroups older than 30. The proportion of those present-ing with advanced HIV also varied by clinic, but thismay be expected as larger structural and contextual cor-relates are likely to vary in different care settings. In thisinstance, we noted that Baba Dogo lacks the same levelFig. 1 Participant recruitment flow diagramvan der Kop et al. BMC Infectious Diseases  (2016) 16:169 Page 4 of 9of community outreach programs present in Kibera.Community outreach programs in Kibera include home-based HIV testing and counselling (HBTC) by variousorganizations, including Amref Health Africa, LiverpoolVCT, and the Centers for Disease Control and Preven-tion (CDC). Kibera also benefits from numerous clinicsat which HIV can be tested, including a Médicins SansFrontières (MSF) clinic that opened during the time ofrecruitment for this study. While HBTC efforts exist inBaba Dogo, they are less prevalent. These community-based efforts may drive earlier testing and linkage tocare, and may have led to comparatively fewer clientspresenting with advanced HIV in Kibera.Comparability with other studiesVarious definitions of ‘advanced HIV disease’ have beenused in the literature, and the term has frequently beenused interchangeably with the term ‘late presentation’.Definitions of ‘advanced HIV disease’ from studies insub-Saharan Africa have included: CD4 count <100cells/μL or WHO stage 4 [10]; WHO stage 3 or 4 [9];and CD4 < 100 cells/μL [14]. The wide array of defini-tions used makes it difficult to compare the proportionsof individuals presenting to care with advanced HIVacross studies. For instance, in a large, multi-countrystudy by Lahuerta et al., 19 % of those enrolling in carewere classified as having advanced HIV [10], comparedto 33 % in our study; however, Lahuerta et al. used alower CD4 threshold of 100 cells/μL, so more individ-uals might have been classified as having advanced HIVthan if a higher threshold of 200 cells/μL had been used.The recent development of consensus definitions of ‘latepresentation’ (CD4 below 350 cells/μL or presentingwith an AIDS-defining event), ‘presentation with ad-vanced HIV’, and even ‘presentation for care’ [1, 17], andtheir use going forward, will facilitate comparison be-tween studies in the future.During the course of this study, the clinics transitionedfrom initiating treatment at CD4 counts of 350 cells/μLor less to the 2013 WHO’s recommendations to initiatetreatment at 500 cells/μL or lower [30]. The benefits ofwhich include improved survival, immune recovery, anda decreased risk of transmission [30]. With approxi-mately one-third of patients presenting to care with ad-vanced HIV, and over half of the population presentingwith a CD4 count lower than 350 cells/μL, the majorityof patients at these clinics will not be affected by thechange in treatment guidelines, or more recent recom-mendations to initiate treatment upon diagnosis, regard-less of CD4 count. Over the past decade, CD4 count atpresentation has not markedly increased in sub-SaharanAfrica [7], and while it is too early to tell whether imple-mentation of the new guidelines will promote earlierpresentation to care, our study emphasizes that it isTable 1 Demographic and clinical characteristics of participants.Values are numbers (percentages)Variable Non-advanced HIVat presentation tocare (n = 505)Advanced HIV atpresentation tocare (n = 248)SexMale 183 (61.4) 115 (38.6)Female 322 (70.8) 133 (29.2)Age (years)Mean (SD) 32 (9.22) 37 (10.26)<30 227 (79.9) 57 (20.1)30–39 178 (62.9) 105 (37.1)40–49 71 (57.7) 52 (42.3)≥50 29 (46.0) 34 (54.0)EducationNo secondary school 327 (64.9) 177 (35.1)Some secondary school 178 (71.5) 71 (28.5)ClinicKibera 364 (68.4) 168 (31.6)Baba Dogo 141 (63.8) 80 (36.2)CD4Median (IQR) (cells/μL) 389 (298–545) 90 (42–147)≤350 210 (46.0) 247 (54.0)>350 295 (99.7) 1 (0.3)WHO Stage1 356 (78.1) 100 (21.9)2 72 (63.7) 41 (36.3)3 57 (39.9) 86 (60.1)4 0 (0.0) 8 (100.0)Missing 20 (60.6) 13 (39.4)Previous HIV diagnosisNo 199 (66.1) 102 (33.9)Yes 306 (67.7) 146 (32.3)Travel time to clinic<30 min 236 (63.6) 118 (36.4)30–59 min 219 (70.9) 90 (29.1)≥60 min 77 (67.5) 37 (32.5)Missing 3 (50.0) 3 (50.0)Alcohol useNon-heavy/hazardous drinking 356 (67.3) 173 (32.7)Heavy/hazardous drinking 149 (66.5) 75 (33.5)Illicit drug useNot a current drug user 469 (66.4) 237 (33.6)Current drug user 36 (76.6) 11 (23.4)Abbreviations: SD standard deviation, IQR interquartile rangevan der Kop et al. BMC Infectious Diseases  (2016) 16:169 Page 5 of 9critical to develop and implement strategies that en-courage earlier diagnosis. Without this, stated targetsof expansion of therapy to those who are eligible andthe intended individual- and population-level effectsof the new WHO recommendations will not be fullyrealised.The importance of earlier diagnosis is further sup-ported by our findings that presentation with advancedHIV was largely due to delayed diagnosis, rather than adelay in seeking care after diagnosis. Prior studies onpresentation with advanced HIV in the region did notexamine prior diagnoses [13], considered new diagnosesonly [14], or were based on clinical records [9, 10],thereby restricting their ability to investigate the pathwayto presentation with advanced HIV. In addition to ourfinding that approximately ¾ of those with advancedHIV who had been previously diagnosed presented tocare within three months, the proportion of individualswho had had a previous diagnosis was similar betweenthose with or without advanced HIV, and the mediantime to first presentation to HIV care did not signifi-cantly differ between the two groups. This supports ourconclusion that advanced HIV at first presentation wasprimarily due to delayed diagnosis. This is not to under-estimate the importance of promoting timely linkage tocare; however, as almost ¼ of those with advanced HIV(who had been previously diagnosed) took longer thanthree months to present to care, and many individualswho test positive do not link to care [31, 32].Similar to Kigozi et al. in their Ugandan study [9], wefound that older age was associated with advanced HIVat presentation. This may be due to simply having livedlong enough for the disease to progress to an advancedstage, or age-associated differences in HIV awareness,knowledge or stigma that may affect testing and othercare-seeking behaviours [33]. Reducing the barriers toand encouraging earlier diagnosis among older adults isparticularly important because of the smaller gains madein CD4 response and increased risk of mortalitycompared to younger age groups once ART is initiated[33-35]. A study from South Africa found no associationwith age and presentation with advanced HIV [15]; how-ever, in the South African study, age was dichotomizedat 40 years, which may have underestimated the vari-ation in risk according to age [36].Although not statistically significant, relatively moremen than women presented to care with advanced HIV.This is in contrast to other reports [9, 10, 14], whichfound strong associations between male gender andpresentation with advanced disease. This may have beendue to the exclusion of pregnant women from this study.In studies on CD4 at presentation to care, those with afocus on prevention of mother-to-child transmission(PMTCT) reported a higher mean CD4 count at presen-tation (395 cells/μL) [7] than non-PMTCT-focussedstudies, and those enrolling in PMTCT services havebeen found to have a lower likelihood of presenting withadvanced HIV disease than others [10]. By excludingpregnant women in this study, the difference betweenthe genders in the risk of presenting with advanced HIVmay have been attenuated.Other factors of interest, such as educational level [9, 12],hazardous drinking [9, 12, 15], and travel time toclinic [9], have been found to be associated with pres-entation to care with advanced HIV in previous stud-ies but were not in our cohort. There are severalpossible explanations for this beyond the differentpopulations under study. First, the use of varying def-initions of advanced HIV may underlie the differingresults: factors that are associated with advanced HIVwhen a lower CD4 threshold is used may not besimilarly associated with advanced HIV when the con-sensus definition is applied. Other possible explana-tions include high levels of missing data in someTable 2 Univariable and multivariable analysis of variables associated with presentation to care with advanced HIV diseaseCrude ORs Adjusted ORs Final adjusted ORsVariable OR 95 % CI p-value OR 95 % CI p-value OR 95 % CI p-valueAgea 1.66 1.41–1.96 <0.001 1.65 1.39–1.97 <0.001 1.72 1.45–2.03 <0.001Presenting at the Baba Dogo clinic 1.23 0.88–1.71 0.220 1.53 1.08–2.17 0.018 1.55 1.09–2.20 0.014Secondary education 0.74 0.53–1.03 0.070 0.69 0.49–0.98 0.040 0.73 0.52–1.03 0.073Male gender 1.52 1.12–2.07 0.008 1.30 0.93–1.82 0.128Illicit drug use 0.60 0.30–1.20 0.155 0.53 0.26–1.09 0.084Hazardous drinking 1.04 0.74–1.44 0.835Travel timeb 0.98 0.84–1.34 0.782Previous HIV diagnosis 0.93 0.68–1.27 0.650Abbreviations: OR odds ratio, CI confidence intervalaOR corresponds to an increase in the odds ratio per unit increase in age category (<30 years, 30–39 years, 40–49 years, ≥50 years)bOR corresponds to an increase in the odds ratio per unit increase in travel time category (<30 min, 30–59 min, ≥60 min)Hosmer-Lemeshow goodness-of-fit p = 0.199van der Kop et al. BMC Infectious Diseases  (2016) 16:169 Page 6 of 9previous studies, which may have impacted findings;investigation of a large number of variables, increas-ing the likelihood of chance findings; and a largersample size in some of the previous studies, increas-ing the power to detect effects.Strengths and limitations of the studyOne of this study’s principal strengths is our collectionof information on the date of participants’ first HIVdiagnosis, helping us to illuminate the pathway to pres-entation to care with advanced HIV. A second majorstrength of this study is the completeness of the data:CD4 data was available for 97 % of the cohort, and infor-mation on additional variables, such as alcohol or illicitdrug use, was available for all participants. Additionalstrengths include a high participation rate, which mini-mized the possibility of non-participation bias. This,combined with the inclusion of two sites in the study,improves this study’s generalizability.We used a consensus definition of advanced HIV, partof which is based upon presenting with an AIDS-defining event regardless of CD4 count. WHO stage 4was used as a proxy for AIDS-defining events. Althoughmost AIDS-defining conditions are included in WHOstage 4, WHO staging data was only available for 85 %of the cohort, which may have led to the misclassifica-tion of some participants as non-advanced HIV. Further-more, CD4 was measured at only one point in time.Given laboratory variability in CD4 measurements, andthe possibility that other factors may temporarily influ-ence CD4 counts [1], it would have been preferable tohave a confirmatory CD4 count.Other limitations include the self-reported nature ofdata on the occurrence and timing of prior HIV diagnoses.There is a lack of data from Kenya on the validity of self-reported HIV testing data; however, studies from otherparts of sub-Saharan Africa suggest that HIV-positive in-dividuals may underreport past testing [37, 38]. While thereasons for underreporting a previous diagnosis are notwell understood, one possibility is that individuals fear be-ing turned away from the clinic if they indicate they areaware of their status. At the study sites, care and treat-ment guidelines were developed to protect clients frombeing denied care. It is standard clinical practice to test allpatients who come to the clinics for HIV care, regardlessof whether they have been tested or diagnosed with HIVbefore. These guidelines may encourage more honestreporting on previous diagnoses than might otherwiseoccur. To improve the quality of self-reported data, partic-ipants completed the questionnaire after being assured ofconfidentiality. In addition, the questionnaire was admin-istered by an experienced HIV research nurse in a privateroom. Finally, data from this section of the questionnairewere cross-checked with a later section to assessconsistency. A high degree of consistency was found, andany discrepancies were investigated further and resolved.Despite the limitations inherent in self-reported data, thestudy is unlikely to suffer from recall bias. There is nostrong reason to believe that the advanced HIV versusnon-advanced HIV groups would differentially reportevents; however, data on the dates of previous HIV diag-noses may be less valid than if we had had access to clin-ical records.ConclusionsPresentation to care with advanced HIV continues toburden global HIV programs. In our study, this appearedto be largely due to delayed diagnosis, rather than delaysin seeking care after diagnosis. The benefits of early HIVcare and treatment for both individual health reasons,and population benefits through treatment as preven-tion, are now widely accepted. Efforts are needed tomaximize earlier diagnosis and entry into care at thefront end of the HIV care continuum to fulfil new globaltargets. Otherwise, changing guidelines to recommendtreatment earlier in the course of HIV infection will notachieve their intended outcomes.Ethics approval and consent to participateThe study protocol was approved by the University ofBritish Columbia’s Clinical Research Ethics Board (H12-00563) and Amref Health Africa’s Ethics and ScientificReview Committee (P40/12). Individuals provided writ-ten informed consent to participate.Consent for publicationNot applicable.Availability of data and materialsData supporting our findings will be shared upon request.AbbreviationsAIDS: Acquired Immune Deficieny Syndrome; AOR: adjusted odds ratio;ART: antiretroviral therapy; AUDIT-C: Alcohol Use Disorders Identification Test;CD4: cluster of differentiation 4; CI: confidence interval; HIV: HumanImmunodeficiency Virus; HTC: HIV testing and counselling; IQR: interquartilerange; OR: odds ratio; SPSS: Statistical Package for the Social Sciences;TX: Texas; WHO: World Health Organization.Competing interestsThe trial in which many participants were enrolled uses a technologyplatform (WelTel/SMS) that has been developed by a non-profit organizationand a private company. The primary investigator of this study, Dr. RichardLester, has financial as well as professional interests in both organizations.For more information, please contact Dr. Lester at richard.lester@ubc.ca.None of the other authors declared conflicts of interest.Authors’ contributionsMVDK conceived the study. RTL and AME contributed to its design. LTprovided guidance on the statistical analyses. MVDK performed the statisticalanalyses and took primary responsibility for writing the manuscript. SM, LB,PA, LT, AME, and RTL contributed to the interpretation of data analyses andcritiqued the manuscript. All authors read and approved the final version ofthe manuscript.van der Kop et al. BMC Infectious Diseases  (2016) 16:169 Page 7 of 9AcknowledgementsWe would like to thank Patrick Nagide, Richard Gichuki and Bonface Abunahfor their dedicated work on the study; and all of the clinical staff, researchteam members, and patients who participated in the study. We are alsograteful to Joshua Kimani, Lawrence Gelmon and the Kenya AIDS ControlProject for supporting the study at the Baba Dogo Clinic. Finally, we wouldlike to acknowledge Koki Kinagwi in her contributions as one of AmrefHealth Africa’s investigators.FundingResearch reported in this publication was supported by the National Instituteof Mental Health of the National Institutes of Health under award numberR01MH097558, awarded to the University of British Columbia. The content issolely the responsibility of the authors and does not necessarily representthe official views of the National Institutes of Health. This study wasadditionally supported by the CIHR Canadian HIV Trials Network (CTNS284).MLVDK is supported by a Canadian Institutes of Health Research (CIHR)Doctoral Award – Doctoral Foreign Study Award (October 2012), offered inpartnership with the CIHR Strategy for Patient-Oriented Research and theCIHR HIV/AIDS Research Initiative.Author details1Department of Public Health Sciences/Global Health (IHCAR), KarolinskaInstitutet, Widerströmska Huset, Tomtebodavägen 18A, Stockholm 171-77,Sweden. 2Department of Medicine, University of British Columbia, 828 West10th Avenue, Vancouver, BC V5Z 1M9, Canada. 3Department of ClinicalEpidemiology and Biostatistics, McMaster University, 50 Charlton AvenueEast, Hamilton, ON L8N 4A6, Canada. 4Amref Health Africa, Langata Road,Nairobi, Kenya. 5Department of Infectious Diseases, I73, Karolinska UniversityHospital, 141 86 Stockholm, Sweden.Received: 26 January 2016 Accepted: 8 April 2016References1. Antinori A, Coenen T, Costagiola D, Dedes N, Ellefson M, Gatell J, et al. Latepresentation of HIV infection: a consensus definition. HIV Med. 2011;12(1):61–4.2. The Opportunistic Infections Project Team of the Collaboration ofObservational HIV Epidemiological Research in Europe (COHERE) inEuroCoord. CD4 cell count and the risk of AIDS or death in HIV-infectedadults on combination antiretroviral therapy with a suppressed viral load: alongitudinal cohort study from COHERE. PLoS Med. 2012;9(3):e1001194.3. Johnson LF, Mossong J, Dorrington RE, Schomaker M, Hoffmann CJ, KeiserO, et al. Life Expectancies of South African adults starting antiretroviraltreatment: collaborative analysis of cohort studies. PLoS Med. 2013;10(4).Apr [cited 2015 Jan 21]. Available from: http://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1001418.4. Krentz HB, Gill MJ. The direct medical costs of late presentation (<350/mm3)of HIV infection over a 15-year period. AIDS Res Treat. 2012;2012. [cited2015 Jan 21]; Available from: www.ncbi.nlm.nih.gov/pmc/articles/PMC3166713/.5. Krentz HB, Gill J. Despite CD4 cell count rebound the higher initial costs ofmedical care for HIV-infected patients persist 5 years after presentation withCD4 cell counts less than 350 μl. AIDS. 2010;24(17):2750–3.6. Girardi E, Sabin CA, Monforte AD. Late diagnosis of HIV infection:epidemiological features, consequences and strategies to encourage earliertesting. J Acquir Immune Defic Syndr. 2007;46 Suppl 1:S3–8.7. 90–90–90 - An ambitious treatment target to help end the AIDS epidemic[Internet]. [cited 2015 Nov 4]. Available from: www.unaids.org/en/resources/documents/2014/90-90-90.8. Siedner MJ, Ng CK, Bassett IV, Katz IT, Bangsberg DR, Tsai AC. Trends in CD4Count at Presentation to Care and Treatment Initiation in Sub-SaharanAfrica, 2002–2013: A Meta-analysis. Clin Infect Dis. 2015;60(7):1120–7.9. Lesko CR, Cole SR, Zinski A, Poole C, Mugavero MJ. A systematic review andmeta-regression of temporal trends in adult CD4(+) cell count atpresentation to HIV care, 1992–2011. Clin Infect Dis. 2013;57(7):1027–37.10. Kigozi IM, Dobkin LM, Martin JN, Geng EH, Muyindike W, Emenyonu NI,et al. Late-disease stage at presentation to an HIV clinic in the era of freeantiretroviral therapy in Sub-Saharan Africa. J Acquir Immune Defic Syndr1999. 2009;52(2):280–9.11. Lahuerta M, Wu Y, Hoffman S, Elul B, Kulkarni SG, Remien RH, et al.Advanced HIV disease at entry into HIV care and initiation of antiretroviraltherapy during 2006–2011: findings from four sub-saharan African countries.Clin Infect Dis. 2014;58(3):432–41.12. Kwobah CM, Braitstein P, Koech JK, Simiyu G, Mwangi AW, Wools-KaloustianK, et al. Factors associated with late engagement to HIV care in WesternKenya a cross-sectional study. J Int Assoc Provid AIDS Care. 2015. doi: 10.1177/232595741456768213. Gesesew HA, Tesfamichael FA, Adamu BT. Factors affecting late presentationfor HIV/AIDS care in southwest Ethiopia: a case control study. Public HealthRes. 2013; 3(4):98–107.14. Daniyam CA, Iroezindu MO, Shehu N, Essien M, Sati AK, Agaba EI.Characteristics of HIV/AIDS patients presenting late at a teaching hospital inNigeria. J Med Trop. 2011;13(2):68–71.15. Drain PK, Losina E, Parker G, Giddy J, Ross D, Katz JN, et al. Risk factors forlate-stage HIV disease presentation at initial HIV diagnosis in Durban, SouthAfrica. PLoS ONE. 2013;8(1), e55305.16. Abaynew Y, Deribew A, Deribe K. Factors associated with late presentationto HIV/AIDS care in South Wollo ZoneEthiopia: a case-control study. AIDSRes Ther. 2011;8:8.17. Mukolo A, Villegas R, Aliyu M, Wallston KA. Predictors of late presentationfor HIV diagnosis: a literature review and suggested way forward. AIDSBehav. 2013;17(1):5–30.18. MacCarthy S, Bangsberg DR, Fink G, Reich M, Gruskin S. Late presentation toHIV/AIDS testing, treatment or continued care: clarifying the use of CD4evaluation in the consensus definition. HIV Med. 2014;15(3):130–4.19. van der Kop ML, Ojakaa DI, Patel A, Thabane L, Kinagwi K, Ekstrom AM, et al.The effect of weekly short message service communication on patientretention in care in the first year after HIV diagnosis: study protocol for arandomised controlled trial (WelTel Retain). BMJ Open. 2013;3(6). Availablefrom: bmjopen.bmj.com/content/3/6/e003155.abstract.20. Dalal W, Feikin DR, Amolloh M, Ransom R, Burke H, Lugalia F, et al. Home-based HIV testing and counseling in rural and urban Kenyan communities. JAcquir Immune Defic Syndr 1999. 2013;62(2):e47–54.21. World Health Organization. Interim WHO clinical staging of HIV/AIDS andHIV/AIDS case defintions for surveillance [Internet]. World HealthOrganization; [cited 2015 Jan 22]. Available from: www.who.int/hiv/pub/guidelines/clinicalstaging.pdf.22. Agaba PA, Meloni ST, Sule HM, Agbaji OO, Ekeh PN, Job GC, et al. Patientswho present late to HIV care and associated risk factors in Nigeria. HIV Med.2014;15(7):396–405.23. Bush K, Kivlahan DR, McDonell MB, Fihn SD, Bradley KA, for the AmbulatoryCare Quality Improvement Project (ACQUIP). The audit alcohol consumptionquestions (audit-c): an effective brief screening test for problem drinking.Arch Intern Med. 1998;158(16):1789–95.24. Ochieng-Ooko V, Ochieng D, Sidle JE, Holdsworth M, Wools-Kaloustian K, SiikaAM, et al. Influence of gender on loss to follow-up in a large HIV treatmentprogramme in western Kenya. Bull World Health Organ. 2010;88(9):681–8.25. van der Kop ML. Factors Associated with Attrition from HIV Care during theFirst Year after Antiretroviral Therapy Initiation in Kenya. J AIDS Clin Res. 2014;05(10). [cited 2015 Jan 29]; Available from: www.omicsonline.org/open-access/factors-associated-with-attrition-from-hiv-care-during-the-first-year-after-antiretroviral-therapy-initiation-in-kenya-2155-6113.1000354.php?aid=32189.26. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation studyof the number of events per variable in logistic regression analysis. J ClinEpidemiol. 1996;49(12):1373–9.27. van der Kop M, Gichuki R, Kinagwi K, Muhula S, Abunah B, Kimani J,et al. Advanced HIV disease at first presentation to HIV care: cross-sectional analysis of baseline data from the WelTel Retain study inNairobi, Kenya. [Cited 2016 Jan 29]. Available from: www.ias2015.org/WebContent/File/IAS_2015__MED2.pdf28. Hosmer DW. Applied logistic regression. 2013.29. Urassa W, Bakari M, Sandström E, Swai A, Pallangyo K, Mbena E, et al. Rateof decline of absolute number and percentage of CD4 T lymphocytesamong HIV-1-infected adults in Dar es Salaam, Tanzania. AIDS. 2004;18(3).Available from: journals.lww.com/aidsonline/Fulltext/2004/02200/Rate_of_decline_of_absolute_number_and_percentage.9.aspx.30. Kiwanuka N, Robb M, Laeyendecker O, Kigozi G, Wabwire-Mangen F,Makumbi FE, et al. HIV-1 viral subtype differences in the rate of CD4+ T-celldecline among HIV seroincident antiretroviral naive persons in Rakai district,Uganda. J Acquir Immune Defic Syndr 1999. 2010;54(2):180–4.van der Kop et al. BMC Infectious Diseases  (2016) 16:169 Page 8 of 931. World Health Organization. Consolidated guidelines on the use ofantiretroviral drugs for treating and preventing HIV infection [Internet].[cited 2015 Apr 28]. Available from: www.who.int/hiv/pub/guidelines/arv2013/en/.32. Genberg BL, Naanyu V, Wachira J, Hogan JW, Sang E, Nyambura M, et al.Linkage to and engagement in HIV care in western Kenya: an observationalstudy using population-based estimates from home-based counselling andtesting. Lancet HIV. 2015;2(1):e20–6.33. Negin J, Nemser B, Cumming R, Lelerai E, Ben Amor Y, Pronyk P. HIV attitudes,awareness and testing among older adults in Africa. AIDS Behav. 2012;16(1):63–8.34. Greig J, Casas EC, O’Brien DP, Mills EJ, Ford N. Association between olderage and adverse outcomes on antiretroviral therapy: a cohort analysis ofprogramme data from nine countries. AIDS. 2012;26:31–7.35. Eduardo E, Lamb MR, Kandula S, Howard A, Mugisha V, Kimanga D, et al.Characteristics and outcomes among older HIV-positive adults enrolled in HIVprograms in four sub-Saharan African countries. PLoS One. 2014;9(7), e103864.36. Royston P, Altman DG, Sauerbrei W. Dichotomizing continuous predictors inmultiple regression: a bad idea. Stat Med. 2006;25(1):127–41.37. Johnson L. HIV testing in South Africa: successes and challenges [Internet].[cited 2016 Jan 18]. Available from: sacemaquarterly.com/hiv-prevention/hiv-testing-in-south-africa-successes-and-challenges.html.38. Fishel JD, Barrère B, Kishor S. Validity of data on self-reported HIV status inMalawi and Uganda and implications for measurement of ARV coverage[Internet]. Rockville: ICF International; 2014. [cited 2016 Feb 10]. Availablefrom: dhsprogram.com/pubs/pdf/MR10/MR10.pdf.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:van der Kop et al. BMC Infectious Diseases  (2016) 16:169 Page 9 of 9


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items