Human immunodeficiency virus (HIV) continues to be a major global public health problem, having claimed more than 25 million lives over the past three decades (World Health Organization (WHO), 2006). No region of the world is spared but, developing countries in particular, in sub-Saharan Africa and part of Asia, have much higher rates of infection (WHO, UNAIDS, UNICEF, 2011). The effect of this disease extends to social, economic and political factors threatening the stability of those countries

with high disease burden (Dixon et al., 2001). Because of this unprecedented effect, it prompted a global response which included preventive, treatments and care strategies with the birth of antiretroviral therapy (ART) (WHO, 2010). Although, there is no cure for acquired immunodeficiency syndrome (AIDS), life-prolonging drugs (ART) have become more affordable and accessible to let people living with HIV have healthy and productive lives (Russell et al., 2007).

Ethiopia is one of the few countries with the highest number of people living with HIV/AIDS globally. The estimated number of people living with HIV/AIDS in 2011 was 790,000 and the reported number of people receiving antiretroviral therapy were 222,723 and AIDS deaths were 54,000 (WHO, 2006; WHO, UNAIDS, UNICEF, 2011). Task shifting and decentralization of the service was done in Ethiopia as an initiative to expand the availability of ART (Federal Ministry of Health, 2007; WHO/UNAIDS/UNICEF, 2009; Ministry of Health, 2005, 2006). Despite this, a large number of individuals who initiate ART do not receive long-term follow-up care. Understanding of lost to follow-up (LTFU) at every level is a key to design appropriate intervention and improve HIV care after ART initiation (Rueda et al., 2006). The rate to loss to follow up varied with period of follow up in different studies, for instance. the LTFU rate was 16, 25, and 40% after 6 months, one year and two years, respectively (Rueda et al., 2006; Brinkhof et al., 2008). In South African, nearly 30% were lost to follow-up within three years (Carole Leach-Lemens, 2010). A study conducted in Gonder University hospital in Ethiopia also revealed LTFU of 46% in one year treatment (Wubshet et al., 2012).

Factors such as age, having low baseline CD4-cell count, being ambulatory and financial constraints were found to be predictors for LTFU in many studies (Rueda et al., 2006; Wubshet et al., 2012). Other reasons for LTFU were, improvement in health, adverse effects and feeling sick or being hospitalized (Braitstein et al., 2006; Ammassari et al., 2002). Stigma and social problems such as fear of disclosure and social isolation were also identified as barriers to treatment adherence (Hardon et al., 2007; Ammassari et al., 2001). Patients also reported cost of care and difficulties in taking their drugs when they were among employers or friends as barriers for ad-herence (Zachariah et al., 2006; Forster et al., 2008).

It is a well-established fact that ART reduces mortality from HIV/AIDS related causes (HAPCO, 2010). For patients on ART, retention in care is needed to prevent medication interruptions, maintain immunologic benefits, prevent drug resistance, and monitor the effects of therapy (Ministry of Health, 2005, 2006). Despite the immense benefit of ART care to the patient and entire society, the magnitude and predictors of loss to follow-up of ART clients is not well documented in North Ethiopia, Tigray regional state. Thus this  study  was  aimed  to determine the magnitude and factors associated with loss to ART follow-up.

Study area and setting

The study was conducted at Wukro primary public hospital. The hospital provides comprehensive HIV/AIDS care such as prevention services, curative and support services by a multi-disciplinary team which comprises a doctor, a nurse, a pharmacy technician, a data clerk, case manager and outreach workers. A case-control study design was employed using patients’ charts to collect data on LTFU among adult patients (>15 years old) enrolled for ART service from January 1st, 2009 to December 31^st, 2012. Data extraction format was used to select subjects and an assumption of 1:3 cases to control ratio was applied to make a sample size of 320 using a random sampling technique (Figure 1).

Measurements

The dependent variable was LTFU; ART clients from treatment to three or more months were considered as LTFU (Bekolo et al., 2013) and coded as “1” otherwise as “0”. The predictor variables include both socio-demographic and clinical factors. The tool was developed by reviewing literatures and from previously tested questionnaire modified in the context of the study.

Data analysis

Data were entered, coded, cleaned and analyzed using SPSS for windows version 16. Descriptive statistics of frequency and proportions were calculated for categorical variables and presented in the form of tables. Multivariable logistic regression model was used to identify the independent predictors of LTFU and 5% of alpha level was used to declare level of significance.

Ethical consideration

Ethical clearance was secured from the ethical committee of Mekelle University, College of Health Science. All data accessed were kept confidential and patient identifiers were excluded from the data base. Those who already knew the details of the patients and were responsible for their ART treatment were used in the data collection process. Hospital manager and ART team were communicated on how to restart ART for those LTFU.

A total of 727 patients aged 15 years and above were on ART during the initiation of the study. Of those ever started, 80 (11%) were found to be LTFU (cases) with response rate of 100% and 240 were controls. Of the final study sample, (n = 320), majority (44.7%) were aged 25 to 34 years at the time of initiation of ART. The median age of patients was 34 years. Women contribute nearly 69% of the patients (Table 1)

Clinical characteristics

Concerning clinical factors, the mean baseline weight for study participants was 45.4 ± 7.5 kg. At the time of ART initiation; 76.2% of patients were in World Health Organization (WHO) stage III or IV which represent advanced stage of the disease. This was also evidenced by the fact that 79.4% had CD4 count less than 200 cells/ul at the time of initiation of ART (Table 2).

Lost to follow-up status

The proportion of lost to follow up in this study was 11 and 100% of the lost clients who were reached by the outreach workers and case managers by using telephones and making home visits. Among the LTFU clients, 44% were in the age group 15 to 34, nearly 69% were females, 46% were rural residents and 56% had no education (Table 1).

Factors associated with LTFU

Socio demographic variables were analyzed but were not significant, except grief or bereavement concern and clinical factors such as INH prophylaxis, presence of side effect and duration of follow up were found significantly associated with loss to follow up to ART treatment in the final model. Those who did not feel grief at initiation of ART were less likely to LTFU compared to those who felt grief (AOR = 0.12; 95% CI: 0.046, 0.30). Moreover, those who did not take INH prophylaxis were three times more likely to LTFU (AOR = 3.04; 95% CI: 1.3, 7.3) (Table 3).

Reasons for LTFU among patients on ART

The tracing result revealed that 76 of them were found to be alive and four were dead. Of those found alive 40 (52.6%) of them restarted treatment in the hospital. The commonest cause (28%) of LTFU was preference of traditional healer (Figure 2).

INH which is recommend by the National Treatment Guideline may have an indirect effect on patient retention since it decreases the occurrence of tuberculosis disease which is a common cause of morbidity and mortality for HIV patients on ART (Federal Ministry of Health, 2007). Existing interventions such as management of opportunistic infections could encourage patients to engage in care and could need greater effort to retain patients from the beginning of HIV care and after ART initiation.

Presence of ART drugs side effects was also one of the factors significantly associated with LTFU (AOR = 12.34: 95% CI: 4.86, 31.35) which is supported by a finding from South Africa (Ive et al., 2007). This implies the need for health professionals to identify and give special attention and closer follow-up for patients who have treatment side effects. Duration of ART treatment was also found to be a significant determinant of loss to follow up. Patients who had duration of follow-up three years and below had significantly more chance of getting lost from follow up (AOR = 23.54; 95% CI:8.87, 62.45). This is scientifically agreeable in that patients on ART get stabilized on the ART care as the duration of follow up gets longer (Sieleunou et al., 2009).

Understanding of the reasons for not returning to care is important to look at alternative strategies on ART programs. This study found that preference for traditional healers is commonest reason for LTFU. Similarly, other studies reported self-transfer to other facility, cost of care, access to ART, acceptance to ART as a reason for not returning to the health care (Hardon et al., 2007; Zachariah et al., 2006; Forster et al., 2008). In this regard, efforts to decentralization of services, task shifting to lay care providers, longer drug refill periods for stable patients, as well as provision of transport need to be addressed to tackle the problems.

The level of lost to follow up was 11%. Most of the clinical factors were found major predictors to LTFU. Though some socio-demographic factors were associated with LTFU in other studies, none was found significant in this study except being grieved. This relatively low rate of LTFU may imply the improvement of quality care and decentralization of services as well as the increasing community awareness. Since the clinical factors were found significant, care providers working in ART and counseling service should work on the risk factors to increase retention of ART patients.

The authors would like to acknowledge Mekelle University, College of Health Sciences for facilitating the study. Also, we would like to thank the staffs of Wukro Hospital and data reviewers for their substantial contribution to the success of the research. Last but not least, we would like to appreciate colleagues who edited the manuscript.

The authors declare that they have no competing interests.