Abstract
The aim of this study was to evaluate to what extent travel-related factors may cause adherence failure to antiretroviral therapy (ART) in otherwise adherent migrants when traveling back to Africa. HIV-infected sub-Saharian migrants living in France with a plasma HIV viral load < 200 copies/mL, with no change in ART for ≥3 months and who were about to visit their native country for between 2 weeks and 6 months were enrolled for the study. Patients completed a self-administered adherence questionnaire both at enrollment and during the week following their return to France. Adherence failure occurred in 23 (11.5%) of 200 patients. Negative perception about ART effectiveness (adjusted odds ratio = 4.3; 95% confidence interval = 1.3-13.7), unexpected traumatic events during their stay in their native country (7.8; 2.3-26.1), and a prolongation of their stay (5.2; 1.4-20.4) were independently associated with a higher likelihood of adherence failure. Owning/renting one’s house in France (0.30; 0.10-0.96), singlehood (0.23; 0.05-1.00), and HIV status disclosure (0.19; 0.05-0.76) were correlates of sustained adherence during traveling.
Introduction
In France, new HIV cases are increasingly reported among migrants from sub-Saharan Africa. They now represent the second largest HIV risk group (29%), next to men who have sex with men (37%), the largest HIV risk group. 1 In addition, as there is considerable delay in the diagnosis of HIV infection in these patients, 2 uptake of and adherence to antiretroviral therapy (ART) remain the major challenges.
Among the patients who take ART, virological efficacy mainly depends on ART adherence. 3 Moreover, in individuals with controlled viremia after the first month of treatment, the virological success is compromised primarily by severe deviations from high ART adherence during follow-up. 4 Such deviations may easily occur from a change in the sociocultural context as has already been shown, for example, in migrant populations whose ART adherence decreased during visits to their native country. 5 Correlates, such as experiencing discrimination and nondisclosure of HIV status to one’s family or household members, have been identified. 5 –8 However, it is still unknown to what extent other factors related either to the purpose of such a trip or to unexpected traumatic events during the stay—such as political troubles, road injuries, and thefts—may influence ART adherence.
As migrant patients frequently return to their country of birth, we aimed to study the determinants of decreased ART adherence in those who, before the journey, were adherent, by particularly exploring the local structural and environmental factors (ie, loss of family structure, family pressure, political troubles, losing one’s treatment, theft, duration and extension of the stay, etc), that are frequent events in the context of sub-Saharan Africa. The VIHVO study included sub-Saharan African HIV-infected migrants receiving a protease inhibitor (PI)-based or nonnucleoside reverse transcriptase inhibitor (NNRTI)-based regimen, who traveled to their native country at a time when their infection was virologically controlled. The primary objective was to describe the course of ART adherence before and during travel and to identify predictors of loss of adherence among virologically controlled HIV-infected sub-Saharan migrants traveling to their native country.
Patients and Methods
Design
Between July 2006 and June 2009, the ANRS VIHVO study enrolled 268 HIV1-infected native sub-Saharan African migrants living in France. Inclusion criteria were as follows: aged at least 18 years, having the current plasma HIV viral load (pVL) less than 200 copies/mL, and unchanged combination ART regimen for at least 3 months prior to enrollment. The ART included a minimum of 3 antiretroviral (ARV) drugs with at least 1 NNRTI or 1 PI. Patients were enrolled only if they planned to visit a sub-Saharan country for a period of between 2 weeks and 6 months, within the next 8 weeks. Medical and social data were collected at enrollment (ie, during the 8 weeks preceding the trip) and also during the week following their return to France. We selected stays longer than 2 weeks in order to exclude touristic trips that do not pose great problems for HIV physicians in terms of care. The 6-month limit was chosen for 2 reasons: first to ensure that participants would return to France before the study ends and second because it was not possible for pharmacies to dispense more than 6 months of ART on one occasion when authorized by the prescribing physician. In this way, we were sure that the participants had the possibility to leave France, taking the whole treatment stock required for the full duration of their trip. Patients were excluded if the pVL at the enrollment visit was more than 200 copies/mL. Patients were enrolled from 24 French university outpatient clinics after they signed a letter of informed consent to participate.
Medical and social data were collected at enrollment and during the week following their return to France. Demographic, social, clinical, and immunovirological data were retrieved from the medical records of the patients. Self-administered questionnaires, distributed to patients at each visit, collected information about ART adherence, factors associated with the trip, which could potentially interfere with ART adherence, disclosure of HIV status, and perceived ART efficacy and toxicity. Disclosure of HIV status was assessed using items investigating whether HIV status had been disclosed to patients’ partners, parents, other family members (brothers/sisters, children, others) and friends, both in France and in their native country.
Adherence Measures
Ten questions about adherence to ART were included in the self-administered questionnaires at each visit (enrollment visit and post-travel visit), in accordance with the methodology established by the AIDS Clinical Trial Group. 9 For each adherence assessment, a 3-level score was computed, according to a previously validated algorithm. 10,11 Patients were classified into 1 of the 3 categories at each assessment—highly adherent (100%), moderately adherent (between 99% and 80%), or nonadherent (<80%)—according to their adherence to treatment scheduling in the 4 days and 4 weeks prior to the enrollment visit and before the end of the trip. Patients were first asked to list, for each drug included in their ARV regimen, the number of pills they had actually taken on each of the 4 days prior to the visit or prior to the end of their trip. Those who reported taking all of their prescribed doses in the 4 days before the visit or the end of the travel were classified as highly (100%) adherent, unless they also reported in subsequent answers that they had skipped a dose during the past week or had “almost totally” followed their ART regimen, or had modified the prescribed scheduling several times or had taken all their medication at one time each day, in which case they were classified as moderately adherent. Patients were also classified as moderately adherent if they reported taking 80% to 99.9% of their prescribed doses of ART in the previous 4-day period. Patients who had taken less than 80% of self-reported pills in the previous 4-day period or had “partially” or “not at all” followed their ART prescriptions were considered nonadherent. Highly adherent participants were then reclassified as moderately adherent if they reported taking 80% to 99.9% of their prescribed doses in the previous 4 weeks or if they declared having frequently modified the dosage of the treatment during the same period. Participants were reclassified as nonadherent if they reported that in the previous 4 weeks they had taken less than 80% of pills or had never complied with the medical prescriptions or had stopped their treatment. This questionnaire was validated among migrant patients in France for whom there were no evidence of more biased self-reports when compared with their nonmigrant counterparts. 12 Help to complete the questionnaire was provided to all patients, whatever be their literacy level, by a nonmedical health care professional who was not previously involved in the medical care of the patients.
Statistical Analysis
In order to estimate the impact of individual characteristics and of ART adherence failure during the trip in previously adherent participants—defined as switching from adherence (either highly or moderately) to nonadherence—a logistic regression was used. Selection criteria for data analysis included the patients adherent to ART before traveling (pre-travel visit) and who also had a complete assessment of adherence at the post-travel visit. Potential correlates considered for the analysis were social status (literacy level, employment, type of housing, and marital status), perception about treatment efficacy and toxicity, HIV status disclosure and variables related to the travel: region of destination (Central Africa versus others), locations visited (urban, rural, or both), period of traveling (during the Ramadan fasting period, respecting the fast or not), use of other medicine (ie, traditional) during the stay, illness leading to a medical consultation, unexpected traumatic events during the stay including political, natural, accidental (road injury, theft), familial (bereavement, constraints, conflicts) troubles during the stay, and, finally, trip extension. Variables associated with adherence failure in the univariate model (P < .20) were considered eligible for the final model that was built using a backward procedure based on the log-likelihood ratio test. SPSS, Stata, and SAS statistical packages were used to perform the statistical analyses.
Results
Study Population Characteristics
The 268 participants enrolled in the study are described in Table 1. Participants were mainly female, with a median age of 42 years. Despite the fact that 31% of the participants had AIDS, they had good immunological levels, most having more than 350 cells/mm3. HIV RNA was less than 50 copies/mL in 94% of the participants, while the median duration on ART was 4 years. In all, 68% of the participants enrolled in the VIHVO study reported being owners or renters of their home in France, 53% reported comfortable housing conditions, 60% were employed, 49% were married or living in a stable relationship, and 31% reported that their spouse was living in France. Only 16% of the participants were of French nationality. For the vast majority (97%), the main purpose of the trip was to meet their family. Accommodation while abroad was anticipated as follows: 92% intended to stay at their family’s house or in their own personal housing (owned or long-term rented accommodation) where their family usually lives, 2% with friends, and 6% in a hotel or short-term rented accommodation. Overall, the median (interquartile range [IQR]) planned duration of time spent abroad was 5 weeks (4-9) and the median (IQR) actual duration was 6 (4-9), with 4 participants staying for more than 6 months (11, 14, 15, and 24 months).
Characteristics of the Patients Included in the VIHVO Study According to Adherence Level at Study Entry (Enrollment Visit).
Abbreviations: ARV therapy, antiretroviral therapy; ART, antiretroviral therapy; IQR, interquartile range; pVL, plasma HIV viral load; PI, protease inhibitor; NNRTI, nonnucleoside reverse transcriptase inhibitor.
a Level of adherence: patients were classified as ART nonadherent when the adherence was <80%; patients were classified as adherent when the adherence was 80% to 99% (moderately) or 100% (highly adherent).
Among the 268 included patients, 251 had fully completed the adherence questionnaire before traveling, among whom 26 (10%) were classified as nonadherent and the other 225 (90%) as adherent, according to our classification criteria. Of these 225 patients, 200 also completed the adherence questionnaire at the end of the trip and were included in the analysis of risk factors for adherence failure. These 200 patients had demographic, social, and medical characteristics similar to the whole study population and are shown in Table 1. In this group, HIV infection seemed to be stabilized, as the median duration on ART was 3 years, the CD4 count level was between 350 and 500 cells/mm3 for 64 participants (32%) and was ≥500 cells/mm3 for 72 participants (36%), the pVL was below 200 copies/mL since a median time of 2 years and the pVL at the inclusion visit was less than 50 copies/mL for 95% of the patients. The ART mainly included a PI for 128 (64%) of these 200, ritonavir boosted for 119 of them. Disclosure of HIV status, that is, disclosure to patients’ partners, parents, other family members, and/or friends, in France and/or in their native country, was 87%.
Correlates of Adherence Failure
Among the 200 patients, 177 (89%) remained adherent, while 23 (11%) experienced adherence failure during the period abroad. All patients traveled to their country of origin: 98 (49%) travelled to Central Africa (54 to Cameroon, 19 to Congo, 13 to the Democratic Republic of Congo, 7 to Central African Republic, 4 to Gabon, and 1 to Chad), 89 (45%) traveled to West Africa (28 to the Ivory Coast, 27 to Mali, 8 to Senegal, 6 to Guinee, 5 to Ghana, 4 to Burkina Faso, 4 to Mauritania, 4 to Togo, 1 to Benin, 1 to Nigeria, and 1 to Cape Verde), and 13 (6%) to East Africa (6 to Ethiopia, 3 to Burundi, 1 to Uganda, 1 to Kenya, 1 to Madagascar, and 1 to Rwanda). Of the 200 patients, 12 also travelled to another country nearby the first destination of their trip. The pLV was detectable (≥50 copies/mL) at the post-travel visit in 13 patients (7.3%) among the 177 patients who were consistently adherent during the trip and in 10 patients (43.5%) among the 23 patients who experienced adherence failure during the trip.
Results of the univariate and multivariate analyses of potential correlates of adherence failure are shown in Table 2. Multivariate analysis showed that not living in one’s own house (adjusted odds ratio [aOR] 3.30; 95% confidence interval [CI] 1.04-10.5 compared with owning or renting one’s house), undisclosed HIV status (aOR = 5.21; CI = 1.31-20.74), perception of poor or moderate ART effectiveness (aOR = 4.30; 95% CI = 1.35-13.68), unexpected extension of the stay (aOR = 5.27; 95% CI = 1.36-20.36), and unexpected traumatic events during the stay (aOR = 7.78; 95% CI = 2.32-26.12) were all associated with a higher risk of adherence failure, while being single (aOR = 0.23; 95% CI = 0.05-1.00 compared with living as a couple/being divorced) was associated with a lower likelihood of adherence failure. Of note, 16 out of 23 patients with adherence failure extended their stay, of whom 7 did not have enough treatment for the whole duration of the trip, 6 of the 7 failing to obtain supplies of drugs during the stay abroad. Among the 177 patients without adherence failure, 6 of the 93 patients who extended their stay did not have enough treatment for the whole duration of the trip and failed to obtain supplies of drugs and 1 of the 84 patients who did not extend their stay left France without enough treatment for the whole duration of the trip.
Factors Associated with Adherence Failure during a Trip Back to Sub-Saharan Africa in 200 HIV1-Infected Migrants Adherent to ART and Living in France: Univariate and Multivariate Logistic Regression Analyses (VIHVO Cohort).a
Abbreviations: ART antiretroviral therapy; IQR, interquartile range; OR, odds ratio; CI, confidence interval.
a Adherence failure: decrease in adherence rate from being adherent, moderately (adherence 80%-99%) or highly (100% adherent), to being nonadherent (adherence < 80%).
b OR per 1 year increment.
c Troubles included political, natural, accidental (road injury, theft) and familial (bereavement, constraints, conflicts).
Discussion
Among migrant patients originating from sub-Saharan Africa with both stable virological response and adherence while on ART before journeying to their native country, 11.5% reported adherence failure during their stay abroad. Some structural factors, such as unexpected traumatic events during the stay or extending the stay, were significantly associated with adherence failure. These results remained valid after controlling for individual cofactors like owning or renting one’s home, living alone, HIV status disclosure, and perception of high ART effectiveness, all of which were correlates of maintaining adherence.
Only 2 previous studies assessed ART adherence in HIV-infected immigrants living in France. 5 –12 The first one studied association between self-reported adherence level and undetectability in migrants and nonmigrants receiving ART out of 2932 HIV-infected patients living in France and participating in the national cross-sectional multicentre VESPA survey. 12 The proportion of patients reporting high adherence level was significantly lower in migrants (99 of 206, 48%) than in nonmigrants (318 of 508, 63%), χ 2 test, P = .0004, and high adherence level was significantly associated with undetectability in both groups. The second study showed that patients native to sub-Saharan Africa and living in France were less adherent to ART during a visit back to Africa, compared to their adherence level in France. 5 Factors associated with a reduced adherence level were a lower education level, being employed, living alone and duration of ART when living in France, and duration of visit, no knowledge of HIV status and/or insufficient support from the family, no knowledge of the HIV status by household members, and lower level of adherence in Africa. The main reported reasons for missing a dose were being busy or having forgotten in both France and Africa and stigma in Africa. In our study, as we did not have a comparator group of migrant patients living in France and not travelling, it is not possible to conclude that the travel in itself leads to ART adherence failure in 11.5% of the migrant patients previously adherent in France, during the occasion of a travel back to their country of origin. However, the main objective of the study was to describe the course of ART adherence during the occasion of a travel and to assess the factors related to the travel, which could interfere with adherence during the travel.
One other limitation of the study is that it was mainly designed to explore factors associated with adherence levels during patients’ stay abroad and journey back to France and did not detail the impacts of detailed social characteristics of migrants living in France, though some information such as employment and financial status, reading and writing capacities, school qualification, housing, on their whole, could be used as proxy of acculturation. This choice was made because the questionnaire was built to contain a fixed number of items to be sure that migrant patients, sometimes illiterate, would answer questions about their adherence and experience with treatment (the core of the study) as complete as possible. However, in this study, the target population concerned migrant patients having the administrative and financial possibility to travel, that is, those who had lived for at least for 3 to 4 years in France in order to obtain administrative papers, then employment, and finally the financial autonomy to be able to travel. One of the interesting features of the study is that we selected only migrants who were adherent to ART and who had good virological outcomes in France. It suggests that even in this “stabilized population” the environmental characteristics of their original country may have compromised their ART continuity and adherence when they traveled there. It also suggests that in general, patients receiving ART in sub-Saharan Africa potentially encounter additional barriers to adherence, which are more related to social and environmental contexts.
Indeed, in clinical practice, migrants frequently travel without any prior consultation with or counseling from their medical doctors, sometimes leaving in emergency situations because of a death of a family member in their native country. They also frequently extend their trip without any care or drug access in the country where they are staying. Furthermore, the individual’s spouse and family in the country of origin (if they are still living there) rarely know the HIV status of his or her visiting spouse or family member. For all these reasons, individuals are at high risk of interrupting their treatment during their trip. Knowing the factors associated with adherence loss will help medical doctors to encourage participants to come for consultation in order to together prepare for the trip before the departure date. It will make the doctors aware of the necessity to provide more treatment than expected for the anticipated duration of the trip, so that individuals can prolong their stay abroad without failing ART adherence. Maintaining adherence is crucial for positive prevention as it is one of the basic conditions for reducing the risk of sexually transmitted HIV.
These results also provide information about the additional contextual barriers to adherence in individuals receiving HIV care in sub-Saharan Africa. Indeed, unexpected traumatic events such as political, natural or accidental (road injury, theft) and familial (bereavement, constraints, conflicts) troubles during the stay were both associated with ART nonadherence. It is still unknown to what extent access to ART in sub-Saharan Africa is possible and feasible for individuals coming from abroad who do not have enough drugs. It is possible that due to fear of stigma and structural barriers to ART, obtaining the extra doses needed for a extended stay may become problematic. The role of structural barriers related to the socioeconomic and family context of the visited country is exemplified by the association between traumatic events and ART nonadherence, which remained the most significant risk factor after multiple adjustments.
In our study, literacy and employment levels were not associated with adherence failure. However, owning or renting one’s house in France, compared with living with others or in a hotel is a proxy of high social status, which was consistently found to be associated with a higher level of adherence. 5 Unlike in other studies, being single was associated with a lower likelihood of adherence failure. 13,14 We have no clear explanation for this, but as our study included many women, it is possible that those patients with children and living in a couple had a lower level of adherence in our study, as personal health could have been sacrificed for other family necessities. 15 Another possible explanation is that, in this population, some of the patients live alone in France but have a stable partner in their country of origin and our questionnaire did not accurately reflect these relationships. Patients who moved to France for financial reasons may be more adherent in order to be able to send financial support to their spouse and/or children in their native country.
In the present study, disclosure of HIV was associated with less adherence failure. Disclosure, either to one’s partner or to at least one family member, was high, reaching 87%. This value is a little higher than that previously shown in non-French citizens living in the French Antilles and French Guiana, where the levels of immigration are high. 14 Disclosure levels were also higher than those found in pregnant women in Tanzania, where 40% of the women disclosed their diagnosis to their spouse after nearly 4 years of living with HIV. 13 It is possible that the high rate of disclosure in our study population can be explained by the higher social status levels and more stabilized HIV infection of our participants in comparison to those usually described in migrant individuals from sub-Saharan Africa. Indeed, it has been shown that a longer time since diagnosis, poorer health status, lower age, and higher educational level are all correlated with disclosure. 14,16 Disclosure of HIV diagnosis should, therefore, be encouraged as it is associated with higher levels of social support and infrequent discrimination. 13,14
In conclusion, the study found that among sub-Saharan HIV-infected migrants, traveling to the native country led to ART adherence failure in 11.5% of the patients, suggesting that these patients should receive reinforced and tailored counseling for adherence and HIV status disclosure. The social–political context and possible structural barriers to adherence within a patient’s native country need to be considered and whenever possible, counseling and tools to face them should be provided to the patient before traveling. To avoid adherence failure and HIV resistance mutations, patients’ journeys should be carefully planned with health care providers and the doctor managing their HIV care, ensuring that they travel with enough medication, such that any extension of their stay can be adequately managed.
Footnotes
Acknowledgments
We gratefully acknowledge the contributions of all clinical investigators of the VIHVO study: Michele Bentata, Fatima Touam, Sophie Abgrall, Nathalie Lele, Olivier Bouchaud, Chakib Alloui (CHU Avicenne, Bobigny, France); Agnès Uludag, Agnès Lefort, Bruno Fantin (CHU Beaujon, Clichy, France); Cécile Goujard (CHU Bicêtre, Le Kremlin-Bicêtre, France); Sophie Matheron, Pascal Ralaimazava, Florence Damond, Gilles Peytavin (CHU Bichat, Paris, France); Odile Launay (CHU Cochin, Paris, France); Philippe Cheval, Jacques Reynes, Vincent Le Moing (CHU deMontpellier, Montpellier, France); Olivier Lortholary, Gaelle Obenga (CHU Necker, Paris, France); Paul-Henri Consigny (CHU Pasteur, Paris, France); Matthieu Saada, Hughes Aumaitre, Salim Gaba (Hôpital de Perpignan, Perpignan, France); Christine Katlama, Ana Canestri, Nadia Valin, Anne Simon (CHU Pitié-Salpétrière, Paris, France); Pierre Tattevin, Cédric Arvieux (CHU de Rennes, Rennes, France); Pauline Campa, Laurence Morand-Joubert, Pierre-Marie Girard (CHU Saint-Antoine, Paris, France); Marie-Aude Khuong, Chantal Chaplain, Isabelle Gros (Hôpital De Lafontaine, Saint-Denis, France); Nathalie Colin de Verdière, Jean-Michel Molina, Pierre Palmer, Caroline Lascoux-Combe (CHU Saint-Louis, Paris, France); David Rey, Jean-Marie Lang, Murielle Rondeau (Hôpital Civil, Strasbourg, France); Laurence Slama, Gilles Pialoux (CHU Tenon, Paris, France); Carine Couzigou, Daniel Vittecoq (CHU Paul Brousse, Villejuif, France); Olivier Patey (Hôpital de Villeneuve Saint Georges, Villeneuve saint Georges, France); Denis Malvy, Thierry Pistone, Philippe Morlat (CHU Bordeaux, Bordeaux, France); Michel Duong, Pascal Chavanet (CHU Dijon, Dijon, France).
Authors’ Note
The authors thank Jude Sweeney for the revising and editing the English version of this manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was supported by the Agence Nationale de la Recherche sur le SIDA (ANRS).