Factors related to full adherence to ART
In a multivariate analysis, gender remained a significant factor after adjusting for potential confounding variables. In a study in rural Uganda, female patients had a significantly higher CD4 cell count at the initiation of ART and lower mortality six months later than male patients, as female patients had more opportunities to access care and start treatment at less advanced stages of HIV, potentially through their participation in prevention of mother-to-child transmission (PMTCT) programs . In Zambia, PMTCT initiative was launched in 1999 and expanded such that an estimated 69% of pregnant women living with HIV had received antiretroviral (ARV) drugs for PMTCT by the end of 2009 . As the case of Uganda, national PMTCT services in Zambia may have contributed to earlier access to ART and supported better ART adherence for a larger group of HIV-positive female, since early ART results in less AIDS progression and death with no increase in adverse events or loss of virologic response compared to deferred ART .
Although not determined in this study, sex differences in treatment response and side effects could also have contributed to the outcomes observed in this study. Another study showed that female patients had better responses to treatment compared with that of male patients, and side effects related to ARV drugs were more frequently observed in male patients than in female patients .
Additionally, female patients may have had greater motivation to adhere to ART as suggested by other study that female patients caring for children emphasize their role as primary care to children and their children are known to be a facilitating factor in adherence among them .
However, other studies have also shown that HIV-positive females often experience gender-related barriers to accessing health services, thus affecting ART adherence [34, 35]. For example, many females have to obtain permission from a male spouse or a relative to seek HIV care, which is difficult when females have to ask for money and take time away from household chores. In addition, where costs for treatment are involved, families may prioritize paying for male’s treatment . Gender-based violence has also affected female’s access and ART adherence . Although these barriers were not found in this study, the issue demands further exploration, particularly given the different social influences on male and female .
Regarding HIV disclosures, over 80% disclosed their status to their spouse. This was positively associated with treatment adherence in bivariate analysis, although it was not seen in the final analysis. The spousal disclosure rate was relatively higher in this study than in a study conducted in another part of rural Zambia in 2005–2006 . The higher rate of disclosure in this study may be attributed to the establishment of peer counselors or treatment supporters during the past few years in Zambia, which encourages disclosure and treatment adherence in the district hospital and rural health centers. However, disclosure could still have both positive and negative consequences. Disclosure has the potential to yield much-needed social support. Alternatively, it may also lead to stigmatization, discrimination, abandonment or gender-based violence mentioned above after disclosing their HIV status to their spouse or partners. Strategies are needed to take account of HIV positive patients who want to disclose HIV status safely to their surroundings.
Having a treatment partner such as a spouse, family member, friends or peer counselor is known to be positively associated with ART adherence . In this study, having a spouse who is also on ART was found to be positively associated with ART adherence. Spouses on ART might play a role as a treatment partner more readily than spouses not on ART because they have a better understanding of treatment adherence for their partners and themselves. While it could not be determined from the results whether spouses were supportive, ART programs should consider the potential benefit of treatment support provided by persons close to patients, especially from those on ART.
A number of qualitative studies have reported that food insufficiency is an important barrier to ART adherence [15, 16, 39, 40]. In urban Peru, Franke et al.  found that individuals who reported food insufficiency in the month prior to interview were more likely to experience suboptimal adherence than those who did not. In three rural ART clinics located in another part of rural Zambia, patients who had skipped a meal because of a lack of food in the past week were more likely to have poor adherence .
However, on the contrary, this study found that the experience of food insufficiency in the previous 30 days from the baseline interview date was positively associated with treatment adherence. This may be explained by enhanced social support targeting people living in such extreme poverty that they cannot afford to buy food described as below. A pilot study in Zambia found that individuals with food insufficiency who received nutritional support demonstrated significantly better ART adherence compared with a group who did not receive this support . The World Food Program (WFP) has been implementing a program in Zambia since 1967 and is committed to providing food assistance to approximately 2.3 million people in Zambia in 2011 . Patients who were extremely poor and suffered from insufficient food at the initiation of ART might more easily receive such assistance compared with patients who were not so poor that they ever experienced food insufficiency at the time, although it is difficult to rely only on such explanation to account for everything since some donors have suspended their assistance in Mumbwa.
On the other hand, one of the most cited reasons for missing doses by this study population was also ‘food insufficiency’, and the experience of food insufficiency in the previous 30 days from the follow-up interview date was also associated with poor treatment adherence by bivariate analysis (data not shown). This paradoxical finding could be understood by considering the timing of the interviews. The question about reasons for missing doses was asked six weeks after starting ART, while the positive association was found between treatment adherence and the experience of food insufficiency in the previous 30 days from the baseline interview date. Therefore, patients who experienced food insufficiency in the 30 days previous to ART initiation might have received food supplementation or counseling afterwards, which might have supported their adherence positively as mentioned above. However, patients who were not so poor thus they could not receive food assistance at the initiation of ART might have needed financial assistance after starting the treatment because of transportation fee and loosing wages due to long waiting times for a clinic visit. They might eventually have fallen into poverty and food insecurity after starting ART, and could not adhere to treatment, even if the ART services are provided free of charge in Zambia.
Moreover, some patients may have been taught and believed that ARV drugs always need to be taken with food and some might have missed the medication as they missed their meal due to the food insufficiency. Thus, individuals who missed ARV drugs might consider food insufficiency the reason. However, more studies are needed to better understand the association between ART adherence and food issues, because a single question was used to assess food insufficiency in this study, which is only one aspect of the issues.
In addition to ‘food insufficiency’, the frequently cited reasons for missing doses in this study were ‘long distance to health facilities’ and ‘being busy with other things like work’. This observation is consistent with findings in other studies [16, 18, 27, 39, 43–54].
Although time required for transportation to the health facilities and transportation fees were not significantly associated with treatment adherence, accessing to treatment facilities can be a problem for many patients living in Mumbwa. This is supported by the result that it took over one hour to reach health facilities in half of patients on ART in this study. Because of this long distance to access to health facilities in Mumbwa, patients who missed doses might report that the distance from home to the district hospital or rural health centers caused the disruption in ART adherence.
In addition, patients whose occupation was agriculture were more likely to have poor adherence. It is possible that patients who worked in agriculture had difficulty coming for health facilities because of their work’s seasonal nature. This is probably supported by our finding that ‘being busy with other things like work’ was the major reason for missing doses.
To enhance understanding of self-stigma and depressive symptoms among patients on ART in Zambia, the associations between these psychological factors and treatment adherence were assessed. Although these factors have been identified as factors associated with poor adherence in multiple studies, no associations were found in this study. The limited numbers of subjects with poor adherence may have prevented the identification of potential associations. More work is needed to investigate patients’ self-stigma and depressive symptoms with having to adhere to lifelong regimens.
This study has several limitations. First, assessment of treatment adherence based on a self-report may be subject to recall and social desirability bias that may result in under-reporting of missed pill intakes. Thus, an over-estimation of adherence is possible. However, there is evidence that a simple self-report adherence questionnaire provides a sensitive measure of non-adherence that predicts viral rebound and is almost always reliable [55–57]. It is also an inexpensive and quick method to use in a field research and resource poor setting.
Second, we could only include those individuals who initiated ART at the target health facilities, and returned six weeks later. Hence, we may also have slightly over-estimated the actual adherence levels of this population, and our sample may not be enough to detect significant associations between patients who were adherent to ART six weeks after starting the treatment and who were not. However, this influence is likely to be limited because there were no significant differences in basic sociodemographic or health characteristics between patients who were included or excluded in this study except for the required time to access health facilities (Table 1).
Third, in relation to recent changes in Zambia, this study was conducted in a rural area where ART services have been initiated. Therefore, it is difficult to generalize the study findings to the population in areas where ART services are not yet available, although there were no differences in patients’ characteristics between those who visited the district hospital and rural health centers.