Late ART Initiation among adult HIV patients at university of Gondar Hospital, NorthWest Ethiopia

Introduction Late initiation of anti-retroviral therapy (ART) is associated with low immunologic response, increase morbidity, mortality and hospitalization. Therefore, this study aimed to assess the prevalence and factors associated with late ART initiation among adult HIV patients in NorthWest Ethiopia. Methods Retrospective cross-sectional study was conducted among 412 HIV patients who started ART between January/2009 and December/2014. Simple random sampling technique was used to select patient records. Data were collected by using pretested and structured extraction tool. Binary logistic regression model was fitted to identify factors associated with late ART initiation. Result A total of 410 participants were included for analysis after excluding 2 participants with incomplete data. The prevalence of late ART initiation was 67.3%. Age between 35–44 years(AOR=3.85; 95%CI:1.68–8.82), duration between testing and enrollment to care<1year (AOR=2.19;95%CI:1.30–3.69), secondary education (AOR=2.59; 95%CI 1.36–4.94), teritary education(AOR=3.28; 95%CI 1.25–8.64), being unmarried(AOR=1.88; 95%CI 1.13–3.03), bedridden and ambulatory patients (AOR=4.68 95%CI:1.49–14.68), other medication use before ART initiation(AOR=2.18; 95%CI 1.07–4.44), starting ART between 2009–2010 (AOR=5.94; 95%CI 2.74–12.87) and 2011–2012(AOR=2.80; 95%CI 1.31–5.96) were significantly associated with late ART initation at p-value <0.05. Conclusion The prevalence of late ART initiation was high. Strengthening the mechanisms of early HIV testing and linkage to care are recommended to initiate treatment earlier.


Introduction
Antiretroviral therapy (ART) has significantly improve the quality of life of the patient by reducing morbidity and mortality 1,2 . Early diagnosis of HIV and initiation of ART in low resource and high HIV prevalent countries minimize the cost of hospitalization and maximize the benefit of ART 3,4 . However, the high magnitude of late ART initiation causes low immunologic response, increased HIV related morbidity, mortality and hospitalization of the patient 1,5,6 . There were frequent changes in ART guideline to expand ART eligible patients in resource limited settings, including Sub Saharan Africa. Despite this change most HIV patients still starting treatment at a very low CD4 cell count and advanced disease 5,7,8 . The median CD4 cell count at initiation of ART was 150 cell/μl in Asia 9 , 122 cells/μl in sub-Saharan region 5,9 , and 132 cell/μl in Uganda 10 . It is still a major challenge for the success of treatment. The prevalence of late ART initiation in Latin America and Caribbean was found to be 76 % 11 , Canada 48% 2 , Australia 20 % 6 , Nigeria 63 % 12 , Lesotho 45.7 % 1 , Mozambique 38 % 13 , Uganda 37.7% 10 , Cameron 35.5 % 14 , and Ethiopia 31.2% 15 .
In Ethiopia, the median CD4 count at enrollment to care is increased from 158 cell/μl to 208 cell/ μl but the median CD4 cell count at ART initiation remained stable and pre-ART attrition contributed a lot 16 . In response to this, since 2010, the Ethiopian government updated ART initiation criteria three times. In 2010 the initiation criteria were CD4<350 cell/µl and/WHO Stage III/IV, whereas in 2014 the initiation criteria was CD4 count<500 cell/ µl 17 . Since 2015, to scale up the ART program, Ethiopia has implemented the HIV/AIDS prevention care and treatment strategic plan in an investment case approach with United Nations Programme on HIV/AIDS (UN-AIDS) 90-90-90 strategies. One of the regional as well as national strategies for achievement is the test and treat initiatives 18 .
Late initiation causes a serious public health problems which results inadequate immune response, premature mortality of patients and increased HIV transmission 19 . Additionally, high levels of virological failure are common among those individual who start ART at low CD4 count 20 . To prevent the problem it is important to identify factors contributing for late initiation. Studies in sub-Saharan Africa reported many contributing factors for late ART initiation 8 , which includes; late HIV diagnosis 7,21 , late enrollment to care, years of ART initiation 9,12,13 , low CD4 count at enrollment to care, female sex , younger and older age 2,14 , unmarried and low level of education 10,14 . Therefore, the aim of this study was to determine the prevalence and associated factors of late ART initiation among adult HIV infected patients who started highly active Anti retro viral therapy (HAART) at University of Gondar Hospital from January 2009 to December 2014. The result of this study will help to evaluate the HIV treatment program over time and provide evidence for future intervention in prevention of late ART initiation.

Study design and period
An institution based retrospective cross-sectional study was conducted at the university of Gondar hospital ART clinic from January 2009 to December 2014.Gondar is about 750 km far from the NorthWest of Addis Ababa, which is the capital city of Ethiopia. Gondar hospital is a teaching Hospital, which serves more than five million people. The hospital started ART service in 2005. More than 7500 adults and 700 pediatric patients were enrolled in the HIV care from the time in which the hospital started HIV care. Currently more than 4900 adults are actively receiving care in this hospital. For this particular study participants were adult HIV patients who were started HAART at University of Gondar Hospital from January 2009 to December 2014.

Inclusion criteria
All patients aged 15 and above who were started HAART at university of Gondar Hospital, from January 2009 to December 2014. Within this period, a total of 2,520 patients started ART and fulfilled the inclusion criteria. Records with no CD4 measurement and/WHO staging one months before or after ART initiation were excluded.

Sample size and sampling technique
The sample size was determined by using single population proportion formula through EPI INFO StatCalc program with the assumption of 95% level of confidence, 5% of marginal error, and by taking 50% prevalence for late ART initiation. With these assumptions, the calculated sample size was 384. Considering 10% expected incomplete record the final sample size was 412. The patients' chart number was taken from the electronic data base of the Hospital and 412(out of 2520) charts were selected through a computer generated random number.

Operational definition
Late ART initiation was defined as having a CD4 count of below or through 200 cell/µl and/AIDS defining illness 17 . CD4 count or WHO stage at ART initiation included those measurement taken with in one months before and after initiation of ART. Those participants who were taking any of the three commonly used psychoactive substances: alcohol, cigarette and/or Khat in the past 30 days/span> were considered as substance user 22 .
African Health Sciences Vol 19 Issue 3, September, 2019 Data collection tools and procedures Data were collected by using data extraction tool. The tool was developed from different literature 2,13 , patient's national standard follow up chart which was adapted from WHO guidelines 23 and patients medical records. The data were collected by four nurses who took an ART training. A data clerk supported data collectors in identifying the charts. Charts were retrieved using the patient's registration number which was found in the electronic database established by I-TECH Ethiopia.

Data quality control
To maintain quality of data the recruited data collectors were trained for ART training. Prior to data collection, One day training was given to data collectors about the aim of the study and how to retrieve data. Pretest was done on 20 patients' records to check the consistency and completeness of the data item. The data collection process was monitored by the investigator throughout the data collection period. Completed extraction tool were checked regularly for completeness of information.

Data processing and analysis
Before the analysis, the collected data was checked for its completeness before and after entry. For those having incomplete or missing data, the collected data from the extraction tool and the original patient records were reviewed again by using the assigned code. Data were checked, coded and entered to EPI INFO version 7 then it was exported to SPSS version 20 for analysis. Both descriptive and analytical statistical procedures were utilized. Descriptive statistics like percentage, mean, median standard deviation and Interquartile range (IQR) were used for the presentation of socio-demographic data and prevalence of late ART initiation. Tables were also used for data presentation. Binary logistic regression was used to identify factors associated with late ART initiation among adult HIV patients. Multivariable logistic regression model was fitted to control the possible effect of confounders and finally the variables which had independent association with late ART initiation were identified on the basis of OR (95% CI) and p-value (less than 0.05). The variables were entered in the multivariable model using the Backward Stepwise regression method. Model fitness was checked using Hosmer and Lemeshow goodness of fit test (P= 0.51).

Ethical considerations
The study was approved by College of Medicine and Health Sciences Research and Ethical committee and Institutional Review Board of University of Gondar. A permission and supportive letter was obtained from clinical directors of the hospital. To retrieve the data from the patients' record permission was obtained from ART clinic head. To ensure confidentiality, personal identifiers like name were not registered in the extraction tool.

Baseline socio-demographic, clinical and immunological status of the respondents
A total of 412 records were reviewed, of which two were excluded because of data incompleteness. Finally, 410 participants were included in the analysis (Table 1). More than half 265 (64.6%) of respondent were females. The mean (SD) age of the patients at ART initiation was 33.3 + 8.7 years. The majority of respondents, 337 (82.2%) were urban residents. Regarding the level of education; 123 (30.00%) of the respondents had completed their secondary education. One hundred seventy two (42.0%) and 150 (36.60%) of the respondent starts HAART at WHO stage III and between CD4 count of 100 and 199 cell/ µl respectively. A total of 372 (90.7%) patients had disclosed their HIV status to either their family or other relatives. Three hundred fifty nine (87.6%) of the respondents did not use any type substance.

Factors associated with late ART initiation
In multivariable logistic regression analysis, age, marital status, level of education, duration between HIV testing and enrollment to care, medication before ART initiation years of ART initiation, and baseline functional status were significantly associated with late ART initiation (Table 2). 200 cell/µl at any time prior to initiation of ART, but in this study, we defined it as CD4 count less than 200 cell/ µl within one months before and after initiation of ART. The time variation in the data collection period could also have contributed for this difference (from 1999-2010 for Caribbean study). Since, early in the era of HIV only limited number of patients were eligible for ART.
In this study, median CD4 count at initiation of ART over time ( to 52.94% within the study period. This finding is consistent with other similar studies 8,9,13 . This indicates the great improvement on HIV prevention and treatment that has been made in Ethiopia for the last few years. ten at diagnosis making it challenging to directly identify correlates of late ART initiation from factors related to late enrollment or diagnosis. Second, because of the retrospective nature of the study some important variables like psychosocial and income related factors were missing in the patient's record which could affect the results.

Conclusion
The prevalence of late ART initiation was found to be high. The median CD4 count is increased over time. Accordingly, the proportion of late ART initiation decreased over time inconsistently. However, significant proportion of HIV patients started ART late. Patients who had finished secondary education and above, were aged between 35-44, were unmarried, were taking medication before ART initiation, years of ART initiation between 2009-2012, duration between HIV testing and enrollment to care less than one year, ambulatory and bedridden functional status were more likely to initiate ART late. This study highlights the need to strengthen the mechanisms of early HIV testing as well as early enrolment/linkage to care and early ART initiation among adults in the community by intensively involving the health professionals and other stakeholders by advocating the current novel strategies (90-90 strategy). It is also better to enhance an intensive counseling and education for unmarried, ambulatory/bedridden patient and patients on other medications. Further research to explore the socio-economical, institutional and contextual factors with appropriate study design is recommended.

Declaration
Authors' contributions DZ conceived the idea and wrote the proposal, participated in the data collection process, data analysis and drafted the paper. BA and BT approved the proposal with some revisions, participated in data analysis and reviewed the manuscript. All authors approved the final draft of the manuscript.