SEARCH

SEARCH BY CITATION

Keywords:

  • defaulter;
  • antiretroviral treatment;
  • adherence;
  • loss to follow up;
  • attrition
  • défaillant;
  • traitement antirétroviral;
  • adhésion;
  • perte de suivi;
  • Attrition
  • incumplidor;
  • Tratamiento Antiretroviral;
  • Adherencia;
  • pérdida al seguimiento;
  • desgaste

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Objective  To determine the prevalence of and factors associated with defaulting from antiretroviral treatment (ART) in Jimma, Ethiopia.

Methods  Unmatched case control study: cases were individuals who had missed two or more clinical appointments (i.e. had not been seen for the last 2 months) between January 2005 and February 2007; controls were individuals who had been on ART at least for 1 year and were rated as excellent adherers by the providers. Data were collected from patient records, and by telephone call and home visit to identify the reason for defaulting.

Results  Of 1270 patients who started ART, 915 (72.0%) were active ART users and 355 (28.0%) had missed two or more clinical appointments. The latter comprised 173 (13.6%) defaulters, 101 (8.0%) who transferred out, 75 (5.9%) who died, and 6 (0.5%) who restarted ART. Reasons for defaulting were unclear in most cases. Reasons given were loss of hope in medication, lack of food, mental illness, holy water, no money for transport, and other illnesses. Tracing was not successful because of incorrect address on the register in 61.6% of the cases. Taking hard drugs (cocaine, cannabis and IV drugs), excessive alcohol consumption, being bedridden, living outside Jimma town and having an HIV negative or unknown HIV status partner were associated with defaulting ART.

Conclusion  A significant proportion of patients defaulted from ART treatment. ART clinics should ensure that patients’ addresses are correct and complete. Programmatic and counseling efforts to decrease ART defaulting should address illicit drug and excessive alcohol use, decentralise ART services, institute home-based treatment options for seriously ill and bedridden patients, and address patients concerns.

Objectif  Déterminer la prévalence et les facteurs associés aux désistements dans le traitement antirétroviral (ART).

Méthodes  Etude cas-témoins non appariée: les cas étaient des personnes ayant manqué deux ou plusieurs rendez-vous cliniques (i.e. n’ont plus été vus au cours des deux derniers mois) au cours de la période de janvier 2005 à février 2007. Les contrôles étaient des personnes ayant été sous ART depuis au moins un an et jugés comme excellents adhérents par les pourvoyeurs de soins. Les données ont été recueillies à l’aide d’un questionnaire semi structuréà partir des fiches des patients, des coups de téléphone et des visites à domicile pour identifier les raisons du désistement. Les données ont été corrigées, nettoyées, codées, saisies et analysées au moyen du logiciel SPSS version 12.0.1.

Résultats  Sur 1270 patients qui ont commencé l’ART, 915 (72,0%) étaient des utilisateurs actifs de l’ART et 355 (28,0%) avaient raté deux ou plusieurs rendez-vous cliniques. Ce dernier groupe comprenait 173 (13,6%) désistements au traitement, 101 (8,0%) qui ont été transférés ailleurs, 75 (5,9%) décès et 6 (0,5%) qui ont recommencé l’ART. Les raisons pour le désistement n’étaient pas claires dans la plupart des cas. Les raisons invoquées étaient la perte d’espoir dans le traitement, le manque de nourriture, la maladie mentale, l’eau bénite, les coûts du transport, d’autres maladies. Retrouver les malades n’était pas facile à cause d’une mauvaise adresse sur le registre dans 61,6% des cas. La consommation de drogues dures (cocaïne, cannabis et les drogues de catégorie IV), la consommation excessive d’alcool, être alité, vivre à l’extérieur de la ville de Jimma, avoir un partenaire avec un statut VIH inconnu ou négatif, étaient associés au désistement à l’ART.

Conclusion  Une proportion significative de patients ART désistent au traitement. Les cliniques ART devraient veiller à ce que les adresses des patients soient correctes et complètes. Des efforts de programmes et de conseils pour réduire le désistement au ART devraient porter sur l’usage de drogues illicites et la consommation excessive d’alcool, la décentralisation des services ART, l’instauration d’options pour des traitement à domicile pour les patients gravement malade et alités et répondre aux inquiétudes des patients.

Objetivo  Determinar la prevalencia de y los factores de riesgo asociados con el incumplimiento del tratamiento antirretroviral (TAR).

Métodos  Estudio caso control no pareado: los casos fueron individuos que habían perdido dos o más consultas clínicas (es decir, que no habían sido vistos en los últimos dos meses) en el período de Enero 2005 a Febrero 2007; los controles fueron individuos que habían estado en TAR al menos durante un año y estaban considerados por los proveedores como excelentes cumplidores. Se recolectaron los datos utilizando un cuestionario semi-estructurado de las historias de los pacientes y de llamadas telefónicas y visitas al domicilio para identificar la razón por la cual han faltado. Los datos fueron editados, limpiados, codificados, entrados y analizados utilizando SPSS versión 12.0.1.

Resultados  De 1270 pacientes que comenzaron TAR, 915 (72.0%) usaban activamente el TAR y 355 (28.0%) habían perdido dos o más consultas clínicas. Estos últimos incluían 173 (13.6%) incumplidores, 101 (8.0%) que habían cambiado de domicilio, 75 (5.9%) que habían muerto, y 6 (0.5%) que habían reiniciado TAR. Las razones para incumplir no estaban claras en la mayoría de los casos. Las razones dadas eran pérdida de fe en la medicación, falta de comida, enfermedad mental, agua bendita, no poder pagar el transporte, otra enfermedad. En un 61.6% de los casos no fue posible encontrar a los pacientes debido a un error en la dirección en los registros. El tomar drogas duras (cocaina, cannabis y drogas IV), un exceso en el consumo de alcohol, el estar encamado, vivir fuera de la población de Jimma y tener una pareja VIH negativa o con un seroestatus desconocido, estaban asociados con el incumpliendo TAR.

Conclusión  Una proporción significativa de pacientes faltaron al TAR. Las clínicas de TAR deberían asegurarse de que las direcciones de los pacientes fueran las correctas y estuviesen completas. Los esfuerzos de programación y del aconsejamiento para disminuir el incumplimiento del TAR deberían tener en cuenta el problema del uso de drogas ilícitas y exceso de alcohol, descentralizar los servicios del TAR, institucionalizar las opciones de tratamiento en el hogar para pacientes seriamente enfermos o encamados, y tener en cuenta las preocupaciones de los pacientes.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Antiretroviral treatment (ART) is the single most dramatic development in the treatment of HIV (Amico et al. 2006). Defaulting from treatment may increase the risk of drug resistance and treatment failure (Adam et al. 2003; Malcolm et al. 2003). Defaulting diminishes the immunological benefit of ART and increases AIDS-related morbidity, mortality and hospitalizations (Hogg et al. 2002). In resource-constrained settings where the health care services are not well developed, poor adherence to treatment and defaulting from treatment are the two major challenges that ART programs face (Hogg et al. 2002).

The HIV/AIDS epidemic in Ethiopia continues to pose a threat to the lives of its people. Currently 1.32 million people are infected and there were 134 500 deaths in 2005 (Ministry of Health 2006). The national prevalence of HIV in 2005 was estimated to be 1.4% (Central Statistical Agency and ORC Macro Ethiopia Demographic and Health Survey 2005 2006). AIDS accounted for an estimated 34% of all young adult (15–49 years) deaths in Ethiopia and 66.3% of all young adult deaths in urban Ethiopia (Ministry of Health 2006). In response to this, the Ethiopian government introduced the ART program with the goal to prolong the lives, to restore mental and physical functions, and to improve the quality of life of PLHA (Ministry of Health 2005). ART was first offered in July 2003 through 12 government hospitals on a co-payment basis. In early 2005, 211 000 men, women and children needed ART but only 16 400 received it. In January 2005, free ART through the fund provided by PEPFAR (US President’s Emergency Plan for AIDS Relief) became available in 22 hospitals (Panos Global AIDS Program 2006).

Previous studies conducted in Ethiopia addressed adherence to HAART and identified factors associated with skipping doses for a shorter period. The broader aspects of defaulter, reasons and associated factors were not identified. To reach the 95% (or more) level of antiretroviral drug dosing adherence to maintain suppression of viral replication, it is necessary to identify, describe, and deal with factors associated with long-term defaulting from treatment. We therefore set out to determine rates of and factors associated with defaulting among ART users.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Setting

This study was conducted with ART users at Jimma University Specialized Hospital (JUSH) in Jimma zone, Ethiopia. The hospital serves as a teaching and referral centre for the population of Jimma and adjacent zones and regions in Southwest Ethiopia. It has a separate unit for the ART program and launched its ART treatment service in July 2003. The hospital registers and gives free ART services for those who fulfil standard diagnostic criteria to start ART. All diagnosed HIV positive individuals from VCT, PMTCT or inpatient services were sent to the ART center and registered. If they were eligible to start ART, a unique ART number was given to each patient. Once diagnosed with HIV and eligible for ART, patients were provided with ART medication and care free of charge. Drugs for self-administered treatment were collected every month. Most AIDS cases were treated on ambulatory basis, with the exception of a few patients who required admission for medical reasons. Those living in rural areas far away from the diagnostic centre were referred to the nearest health facility for treatment initiation and follow-up.

Design and data collection

We conducted an unmatched case control study. Cases were those individuals who had missed two or more clinical appointments (i.e. had not been seen for the last 2 months) in the period of January 2005 to February 2007. Controls were individuals who had been on ART for at least 1 year and were rated as excellent adherers by the providers. For each case, a control that was enrolled in the same month was selected. If more than one control enrolled in the same month, the closest date of enrolment determined selection. Using semi-structured questionnaires (a case note extraction), socio-demographic and behavioural information, and clinical factors that could potentially be associated with defaulting (which were recorded at the time treatment started) were retrieved from patient records and cards. Data were also gathered on telephone calls and home visits to identify the reason for defaulting. The patient card and other forms included questions on socio-demographic information, past medical status, substance use, disclosure status, mental health status, and concerns for adherence, knowledge on HIV and knowledge on ART.

The retrieval was conducted by two trained statisticians and a data clerk in the hospital. Four trained peer counselors traced and interviewed defaulters regarding the main reason for defaulting from treatment and advised them to resume treatment. A default case was defined as an HIV positive patient who had been on ART treatment and missed two or more monthly clinical appointments.

Data analysis

Data were edited, cleaned, coded and entered into a computer and analysed using SPSS version 12.0.1 for Windows. First description on the data was done to determine the rate of defaulter. Bivariate analysis was done to determine presence of statistical significant association between explanatory variables and the outcome variable.

Logistic regression models were performed to identify independently associated factors. All explanatory variables that were associated with the outcome variable (defaulting) in bivariate analyses with a P-value of 0.05 or less were included in the initial logistic models. The models were evaluated using both forward and backward stepwise elimination procedures. P-value 0.05 was used to enter a variable in the model and 0.1 for removal from the model. But variables related to previous medication exposure are likely to be on the causal pathway between many of the other factors and ART defaulting. Therefore, we excluded medication-related factors from the final model since, these factors were back down the causal pathway. Most participants’ baseline CD4 was missing, so we were not able to include it in the final model. This study was approved by the Ethical Committee of Jimma University.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

During the period of register review, 1270 patients were registered as starting ART in the hospital. Of 1270 patients on ART 706 (55.6%) were female, 55 (4.3%) were younger than 15 years. Most 915 (72.0%) were active ART users and 355 (28.0%) had missed two or more clinical appointments. As depicted in Table 1, these comprised 75 (5.9%) dead; 173 (13.6%) defaulted; 101 (8.0%) transferred out; and 6 (0.5%) restarted. Based on these results for all 173 defaulters, one control was identified and analysed.

Table 1.   Pattern of ART use among HIV positive individuals in JUSH Southwest Ethiopia, July 2007
StatusNumberPercent
Active ART users91572.0
Defaulted17313.6
Transferred out1018.0
Died755.9
Restarted60.5
 1270100.0

In terms of socio-demographic characteristics, the 346 cases and controls appeared similar in all but place of residence (the cases were from out of Jimma town compared to controls (P = 0.000); Table 2). Most respondents were male, married, orthodox Christians with secondary education, not employed, aged 24–35 years and resident within the town.

Table 2.   Socio-demographic characteristics of defaulters, JUSH Southwest Ethiopia, July2007
VariablesCases, n (%)Controls, n (%)P-value
Sex0.589
 Male92 (53.2)97 (56.1)
 Female 81 (46.8)76 (43.9)
Marital status0.476
 Never married40 (23.5)29 (16.8)
 Married81 (47.6)92 (53.2)
 Separated9 (5.3) 6 (3.5)
 Divorced18 (10.6)22 (12.7)
 Widowed22 (12.9)24 (13.9)
Educational statues0.171
 No education25 (14.7)18 (10.4)
 Primary57 (33.5)58 (33.5)
 Secondary68 (40.0)85 (49.1)
 Tertiary 20 (11.8)12 (6.9)
Religion0.578
 Muslim43 (25.3)48 (27.7)
 Christian 127 (74.7)120 (71.4)
Place of residence0.000
 Jimma town95 (55.9)128 (74.0)
 Out of Jimma town75 (44.1)45 (26.0)
Employment status0.828
 Employed75 (43.4)73 (42.2)
 Not employed98 (56.6)100 (57.8)
Indication for ART0.256
 CD4 < 200122 (70.9)133 (76.9)
 WHO stage IV, III+TLL < 30050 (29.1)40 (23.1)
Age0.625
 15–2415 (8.7)17 (9.8)
 25–3495 (54.9)86 (49.7)
 ≥3563 (36.4)70 (40.5)

Of the total defaulters 108 (62.4%) were traced. Reasons given for defaulting were loss of hope in medication (21.3%); self-referral to other ART centre (17.6%); lack of food (17.6%); mental illness (7.4%); holly water (7.4%); other illness (7.4%); ‘I will come when I finished anti TB treatment’ (7.4%); could not afford transportation costs (7.4%); imprisoned (4.6%); and ‘unknown’ in (1.6%) of the cases.

We also identified the reasons for unsuccessful tracing of 65 (37.6%) cases: their register showed an incorrect address (61.5%); they lived far from the hospital with no telephone access (21.5%) – because of financial constraints we could not try to trace these patients; they had moved (9.2%); their address was not on record (4.6%); or they could not be found during repeated visits (3.1%). Other concerns related to adherence were also identified at enrolment: stigma for 64.1% of cases and for 56.6% of controls; and efficacy of medication for 10.6% of cases compared to 16.2% of controls and others.

In bivariate analysis many factors were associated with ART defaulting (Table 4). In the logistic-regression model, 5 factors (Table 5) were found to be independently associated with ART defaulting: taking hard drugs (cocaine, cannabis and IV drugs); drinking alcohol most of the time; being bedridden; living outside Jimma town; and having an HIV negative partner (or of unknown HIV status).

Table 4.   Patient characteristics at ART enrollment, JUSH Southwest Ethiopia, July, 2007
VariablesCases, n (%)Controls, n (%)Crude OR (95% CI)P-value
Mental status
 At ease149 (86.1)164 (94.8)1.0 
 Not at ease24 (13.9)9 (5.2)2.9 (1.3–6.5)0.008
Strength
 Normal101 (58.4)108 (62.4)1.0 
 Weak56 (32.4)62 (35.8)1.0 (0.6–1.5)0.880
 Bed ridden16 (9.2)3 (1.7)5.7 (1.6–20.2)0.007
Tobacco
 Never127 (74.7)159 (91.9)1.0 
 Sometimes 11 (6.5)8 (4.6)1.7 (0.7–4.4)0.257
 Most of the time 32 (18.8)6 (3.5)6.7 (2.7–16.5)0.000
Alcohol
 Never 87 (51.2)135 (78.0)1.0 
 Sometimes 21 (12.4)22 (12.7)1.5 (0.8–2.9)0.240
 Most of the time 62 (36.5)16 (9.2)6.0 (3.3–11.1)0.000
Hard drugs
 Yes35 (20.6)1 (0.6)1.0 
 No135 (79.4)172 (99.4)0.02 (0.003–0.17)0.000
Place of residence
 Jimma town95 (55.9)128 (74.0)1.0 
 Out of Jimma town75 (44.1)45 (26.0)2.2 (1.4–3.5)0.000
Took Past OI treatment
 Yes162 (94.2)173 (100)X2 = 8.4, df = 10.004
 No10 (5.8%)0 (%)
Anti TB treatment completed
 Yes57 (34.1)81 (46.8)1.0 
 No110 (65.9)92 (53.2)1.7 (1.1–2.6)0.018
Was on ART treatment before
 Yes7 (4.1)25 (14.5)1.0 
 No165 (95.9)148 (85.5)4.0 (1.7–9.5)0.002
Taken Cotrimoxazole before
 Yes108 (63.2)166 (96.0)1.0 
 No63 (36.8)7 (4.0)13.8 (6.1–31.3)0.000
Baseline CD4 count
 ≤5033 (31.1)19 (15.3)2.5 (1.3–4.7) 
 51–20073 (68.9)105 (84.7)1.00.005
Partner HIV status
 HIV positive29 (16.8)46 (26.6)1.0 
 HIV negative 11 (6.4)5 (2.9)3.5 (1.1–11.1)0.034
 Unknown133 (76.9)122 (70.5)1.7 (1.02–2.9)0.041
Table 5.   Non-treatment-related factors significantly linked to defaulting in a multivariate model
VariablesCases, n (%)Controls, n (%)Crude odds ratio (95% CI)Adjusted OR (95% CI)P-value
  1. In the model mental status, strength, tobacco, alcohol, hard drugs, place of residence and partner HIV status were included.

  2. Because of missing values the model was built on 343 patients.

Hard drugs
 Yes35 (20.6)1 (0.6)1.01.0 
 No135 (79.4)172 (99.4)0.02 (0.003–0.17)0.041 (0.005–0.324)0.002
Strength
 Normal101 (58.4)108 (62.4)1.01.0 
 Weak56 (32.4)62 (35.8)1.0 (0.6–1.5)1.06 (0.63–1.76)0.826
 Bed ridden16 (9.2)3 (1.7)5.7 (1.6–20.2)7.44 (1.93–28.60)0.004
Place of residence
 Jimma town95 (55.9)128 (74.0)1.01.0 
 Out of Jimma town75 (44.1)45 (26.0)2.2 (1.4–3.5)2.71 (1.63–4.52)0.000
Alcohol
 Never 87 (51.2)135 (78.0)1.01.0 
 Sometimes 21 (12.4)22 (12.7)1.5 (0.8–2.9)1.02 (0.49–2.12)0.966
 Most of the time 62 (36.5)16 (9.2)6.0 (3.3–11.1)3.57 (1.78–7.14)0.000
Partner HIV status
 HIV positive29 (16.8)46 (26.6)1.01.0 
 HIV negative 11 (6.4)5 (2.9)3.5 (1.1–11.1)3.71 (1.04–13.24)0.044
 Unknown133 (76.9)122 (70.5)1.7 (1.02–2.9)1.91 (1.05–3.49)0.034

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Use of Highly Active Antiretroviral Therapy (HAART) in clinical practice has led to a significant decline of morbidity and mortality among subjects infected with the human immunodeficiency virus (HIV) (Palella et al. 1998). But some HIV-infected individuals refuse treatment and spontaneously interrupt the prescribed drugs for short or long periods of time. Patients discontinuing HAART usually develop rapid viral rebound and loss of CD4 T lymphocytes (Papasavvas et al. 2000). Therefore, the rate of defaulting from treatment needs to be minimal or non-existent. This study showed that the defaulter rate from ART was 13.6%. This figure is much higher than reported in Malawi (5%) (Kwong-Leung Yu et al. 2007) and comparable with that reported in Trinidad (11.3%) (Quava-Jones & Barthlomew 2004).

The commonest (21%) reason for defaulting was loss of hope in medication. At enrollment, 10.6% of cases and 16.2% of controls articulated concerns about medication efficacy (Table 3). Obviously ART does not completely cure from the virus, but it can improve quality of life and decrease hospitalisation. Therefore, comprehensive counseling sessions (sharing experiences between those on treatment and new comers) may help reduce the number of defaulters.

Table 3.   Adherence concerns among cases and controls at enrollment, JUSH Southwest Ethiopia, July, 2007
Adherence concernCases, n (%)Controls, n (%)
Stigma109 (64.1)98 (56.6)
Medication may not work18 (10.6)28 (16.2)
Side effect13 (7.6)26 (15.0)
Depressed6 (3.5)3 (1.7)
Forgetting5 (2.9)4 (2.3)
Others19 (11.2)14 (8.1)

Similar to a recent study (Kwong-Leung Yu et al. 2007), the commonest reason for unsuccessful tracing was an incorrect address on the register. We could not determine whether this was due to patients giving a wrong address or to clinic staff failing to take comprehensive registration data. In any case, it is important to improve prompt registration of address because this would be helpful for successful tracing in case of a patient is lost to follow-up. Since people living with the virus usually live in rented houses, they are prone to moving when they are identified as being HIV positive. Therefore, monthly updating of patient addresses would help.

Concerns about side-effects and medication not working, which were more prevalent among controls, resolved when ART worked well and without side-effects. This contrasts with concerns over stigma and depression, which were more frequent in cases, and less tractable.

In bivariate analysis, many factors were associated with defaulting. Four were related to past history of drug intake. Those who did not take Cotrimoxazole prophylaxis were defaulters, as opposed to their controls. This implies that those individuals who are prone to default will terminate while they are on chronic care (Cotrimoxazole) before starting ART. Previous non-adherence is a barrier to current adherence (Spire et al. 2002; Bangsberg et al. 2006). Those who did not complete TB treatment were also prone to be defaulters compared to those who completed. This may be due to the fact that they already adhered to the long-term TB treatment, or because they believed that both treatments cannot be taken together.

Five factors were found to be independently associated with defaulting. Having an HIV-negative partner or a partner of unknown HIV status was one. These may be HIV-infected individuals who had an HIV-positive partner who provided support and encouragement compared to those who had HIV-negative or unknown HIV status partner. Patients who have experienced AIDS-related symptoms perceive the disease as serious and are usually more adherent than patients who never had symptoms, or who consider their symptoms unimportant (Steele et al. 2001; Malta et al. 2005). Bedridden individuals were more likely to be defaulters. Brown et al. (2006) suggested that patients with compromised immunologic status are at greater risk of being lost to follow-up, perhaps because repeated illness interferes with clinic appointments and leads to defaulting. Repeated illness was mentioned as a barrier to adherence to ART in the study of Mengesha (2005). This result highlights the importance of home-based ART refill for seriously ill patients.

Alcohol or drug abuse was associated with defaulting, confirming the findings of other studies (Herrera et al. 1992; Gordillo et al. 1999). Drug abuse may impair judgment and the ability to adopt and maintain routine medication use. Therefore, screening for drug abuse and excessive alcohol use, and supportive counseling and treatment for drug abuse might help in promoting long term adherence to ART. Patients living outside Jimma town were more likely to be defaulters than those living in town, which suggests the need to decentralise ART services to local health institutions or develop outreach ART services for remote rural communities. Another way to help these patients could be facilitating monthly transport support when appropriate and feasible.

Our results should be interpreted cautiously. Cases were individuals who had missed two or more clinical appointments. In the case of untraceable patients, individuals may have self-referred to other ART clinics or they might have already died, introducing misclassification. With this limitation in mind, we conclude that a substantial proportion of PLHA default from ART in Jimma. The major reason for defaulting is loss of hope in medication, hence, providers should focus on dealing with concerns of PLHA before and during enrollment in ART services. Tracing patients proved difficult because of incorrect address, therefore, ART clinics need to ensure that patient records are accurate and comprehensive. Programmatic and counseling efforts to decrease ART defaulting should address illicit drug and excessive alcohol use; decentralize ART services; and provide home-based treatment for the seriously ill.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  • Adam BD, Maticka-Tyndale E & Cohen JJ (2003) Adherence practices among people living with HIV. AIDS Care 15, 263.
  • Amico KR, Harman JJ & Johnson BT (2006) Efficacy of antiretroviral therapy adherence interventions. A research synthesis of trials, 1996 to 2004. Journal of Acquired Immune Deficiency Syndromes, 41, 285297.
  • Bangsberg DR, Acosta EP, Gupta R et al. (2006) Adherence–resistance relationships for protease and non-nucleoside reverse transcriptase inhibitors explained by virological fitness. AIDS, 20, 223231.
  • Brown DM, Thorne JE, Foster GL et al. (2006) Factors affecting attrition in a longitudinal study of patients with AIDS. AIDS Care, 18, 821829.
  • Central Statistical Agency and ORC Macro (2006) Ethiopia Demographic and Health Survey 2005, Addis Ababa Ethiopia and Calverton, Maryland, USA.
  • Gordillo V, Del Amo J, Soreno V et al. (1999) Socio demographic and psychological variables influencing adherence to antiretroviral therapy. AIDS 13, 17631769.
  • Herrera B, Gato A & Gaspar G. (1992) [Follow-up of the patient with HIV infection: a hospital task?] Atencion Primaria, 9, 251254.
  • Hogg RS, Heath K, Bangsberg D et al. (2002) Intermittent use of triple-combination therapy is predictive of mortality at baseline and after 1 year of follow-up. AIDS, 16, 10511058.
  • Kwong-Leung Yu J, Chih-Cheng SC, Wang KY et al. (2007) True outcomes for patients on antiretroviral therapy who are “lost to follow-up” in Malawi. Bulletin of the World Health Organization, 85, 550554.
  • Malcolm SE, Ng JJ, Rosen RK & Stone VE (2003) An examination of HIV/AIDS patients who have excellent adherence to HAART. AIDS Care 15, 251.
  • Malta M, Petersen ML, Clair S, Freitas F & Bastos FI (2005) Adherence to antiretroviral therapy: a qualitative study with physicians from Rio de Janeiro, Brazil. Cad. Saude Publica, 21, 14241432.
  • Mengesha A. et al. , (2005). Assessment of Adherence to Antiretroviral Therapy Among HIV-Infected Persons in the Ministry of National Defense Force Hospitals, Addis Ababa and Debreziet. MPH thesis, Addis Ababa University.
  • Ministry of Health (2005) Guideline for Implementation of Antiretroviral Therapy in Ethiopia, Addis Ababa Ethiopia, Federal HIV/AIDS Prevention and Control Office, Federal Ministry of Health.
  • Ministry of Health. (2006). AIDS in Ethiopia, 6th Edition, Disease Prevention and Control Department, Ministry of Health, Addis Ababa.
  • Palella FJ, Delaney K, Moorman A et al. (1998) Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. New England Journal of Medicine, 338, 853860.
  • Panos Global AIDS Program, (2006) Antiretroviral Drugs for all? Obstacles to Access to HIV/AIDS: Treatment Lessons From Ethiopia, Haiti, India, Nepal and Zambia. Available at http://www.etharc.org/arvinfo/arvsforall.pdf. Panos Global AIDS Programme, Lusaka.
  • Papasavvas E, Ortiz GM, Gross R et al. (2000) Enhancement of human immunodeficiency virus type-1 specific CD4 and CD8 T cells responses in chronically infected persons after temporary treatment interruption. Journal of Infectious Diseases, 182, 766775.
  • Quava-Jones A & Barthlomew C. (2004) Challenges faced with adherence and defaulters at and HIV treatment clinic, Port of Spain, Trinidad. International Conference on AIDS. 15, 1116 abstract no. WePeB5805.
  • Spire B, Duran S, Souville M, Leport C, Raffi F & Moattia J-P (2002) Adherence to highly active antiretroviral therapies (HAART) in HIV-infected patients: from a predictive to a dynamic approach. Social Science & Medicine 54, 14811496.
  • Steele RG, Anderson B, Rindel B et al. (2001) Adherence to antiretroviral therapy among HIV-positive children: examination of the role of caregiver health beliefs. AIDS Care, 13, 617629.