Socio-demographic factors such as age, gender, socio-economic status, level of education, income and ethnicity have been used in studies to understand their influence on adherence [48,49]. It has not been possible to absolutely predict adherence based on all these demographic characteristics as no consistent correlation has been found between demographic characteristics and patient adherence levels [3,48]. Psychosocial factors like drugs and alcohol use, social stability, depression and psychiatric illness have also been used in other studies to find out if there were correlation between any of them and adherence to ART [10,49,50,51]. What studies have found are barriers to adherence such as substance abuse, unstable housing, depression, mental illness, fear of disclosure of HIV status, decreased quality of life, work and family responsibility and past history of non-adherence [7,9,52]. Although socio-demographic factors have not been consistently correlated with adherence to ART in several studies, they can be used to identify particular populations that may benefit more extensively from targeted interventions that address specific barriers to adherence [3,49]. The importance of these findings in ART initiation is that denying an individual the benefit of ART based on the assumption that the person would not adhere due to his or her demographic characteristics is a futile exercise as no data are available to substantiate such action [4].

Other socioeconomic factors implicated as part of barriers to adherence to ART in a number of studies in Africa are the cost of ART, availability and accessibility to medications [23,43,53,54,55,56]. There are also some indirect costs associated with ART which influence adherence; these are the time taken off work, the time spent in hospital and inability to fend for one’s family during bouts of opportunistic infections [43,57]. The access to drugs while away from home is another factor that has been identified which affects adherence among patients on ART [58,59]. Certain structural factors tend to influence adherence to ART, which are issues that are beyond the control of the patients [60]. Structural approach to adherence does not see adherence as a simple individual behaviour but as one that occurs within social and environmental contexts. In other words, it is difficult for some people to adhere to ART because the adherence barriers are not generated at an individual level [60]. Some of these are out-of-pocket cost of health services and medication, transportation, homelessness, lack of access to food and water and poor accessibility to health facilities. All these are structural barriers to adherence that have to be addressed at the societal level in order to improve adherence [23,43,60]. Reduction in these structural barriers to healthcare has been shown to improve patient’s adherence levels in ART in some interventional studies to enhance adherence to ART [60].

ART is a complex treatment that is characterized by pill burden, dietary and fluid restrictions and timing of medication intake. The complexity of drug regimens is one of the causes of non-adherence among patients on chronic medication [3,52]. This has also been reported among HIV patients taking ARV; daily dosing regimens of three times or more are associated with non-adherence [60,61,62]. The pill burden is a major challenge since combination ART is used and it often contains 2–20 pills that have to be taken in a day together with the other factors to be considered like timing of dosages and food requirements [9]. The numerous and potentially debilitating side-effects contribute to irregular drug use and deliberate discontinuation of medication intake by some patients [23,63]. Regimens with significant side-effect profiles are usually associated with poor adherence [3,52]. Health care providers should consider the circumstances of patients while prescribing ART [64]; a potent combination therapy may not fit into a patient’s daily schedule and may therefore affect the adherence to such medication. The rate of adherence to once-daily antiretroviral regimens has been found to be better than the rate of adherence to twice-daily regimens [65]. Regimen-based strategies that could improve long term adherence among patients include simplification of ART regimens which will assist in the long term adherence to treatment and maintaining efficacy of treatment [3,9,51,61,64]. The clinical outcomes achieved with the use of ART are linked to adherence levels [3]. Although the Swiss HIV Cohort Study revealed that long standing HIV disease was among the predictors of worsening adherence [52], prior opportunistic infections in an HIV patient before initiating ART have the potential of influencing adherence as the patient may perceive the disease to be severe enough to require good adherence to treatment in order to achieve the desired treatment outcome [3].

Providers’ characteristics and clinical settings affect patients’ adherence; overall patients’ satisfaction with the level of care has been found to correlate with increased adherence [66]. The aspects of clinical setting that could positively influence adherence are a friendly and supportive environment, non-judgmental health care providers, convenient appointment schedule and confidentiality in service provision [67,68]. Long waiting times, poor staff attitudes, intermittent drug availability and other procedural barriers decrease patients’ adherence to ART and also results in poor clinic attendance [53,54]. Continuous access to health care services and medications by patients also influences treatment adherence [53]. The patient-provider relationship is another factor that has been well researched in terms of adherence to ART; a good patient-provider relationship results in patient’s trust and confidence in the provider which in turn influences good adherence [66,67].

The burden posed by adherence is not limited to the patient alone but extends to the health care provider; therefore, the relationship between the patient and the health provider should be that of a therapeutic alliance where both parties work toward a common goal of improving the health of the patient. This is why frequent change in health care providers is associated with poor adherence [52].Discussing ART initiation with an informed patient is an effective strategy that enhances patient-provider interaction. It enables an alliance formation which allows the patient and the provider to define therapy goals, side-effects, medication management and adherence monitoring [64]. Social support and ability to disclose one’s status are important issues to be considered when assessing treatment adherence irrespective of the context [32]. A study on adherence of patients to chronic medications recommends that therapeutic alliance between patients and providers should be strengthened, time should be devoted to address adherence among patients, and patients should be well assessed to identify barriers to treatment adherence [69].

Self-report remains the main measure of adherence globally although other techniques have been used widely; it involves asking patients to report their adherence periodically [71]. Studies have found correlation with actual medication intake in research done on self-report as a method of adherence monitoring [44,71]. In the Ugandan study that used three-day self-report adherence, median adherence correlated with other measures of adherence [44]. The major drawback of this method is overestimation of adherence by patients due to desirability bias [72]. Adherence guidelines stipulate that accuracy of adherence self-report could be achieved if patients are approached in a non-judgmental way during assessment [70].

Pill count as a measure of adherence involves calculating the percentage of the number of pills prescribed and dispensed for the period between hospital appointments with the number of pills returned at the following appointment [42]. It is done during patients’ visits to health facilities or unannounced at patients’ home. This technique is cheap and correlates well with adherence measured using viral load [42], but its major shortcoming is manipulation of pills by the patient. Pill dumping prior to hospital visits has been documented which will result in overestimation of adherence for the patient. Unannounced pill count is more reliable but it has the tendency to affect the trust between the patient and health care provider which may eventually hinder adherence [22]. Though they may overestimate adherence; pill counts and self-report are still the mainstay of measuring adherence in resource-limited settings [73].

This involves an electronic chip that is implanted on the bottle lid that records the opening and closing of the bottle which contained the prescribed medicines, a computer program is later used to extract information from the lid and the data is then analyzed. This technique assumes that opening of the bottle by the patient coincides with actual intake of the drug. Although this method has shown some correlation with actual drug intake in some studies [16,51], others have described it as an expensive and non-reliable device [73]. A study in Malawi that compared methods of adherence monitoring found no correlation between electronic device monitoring, self-reporting and pill counts [73]. Electronic devices may underestimate or overestimate adherence since it is difficult to know if a patient opening a drug bottle is actually taking the tablets or just opening the bottle, therefore, actual medication intake by patients is not measured by EMD [74]. A study of electronic monitoring devices used among HIV patients on ART revealed that patients who rely on pill boxes are not likely to use EMD because pill boxes are the preferred tools for patients taking numerous medications everyday [74].

Pharmacy record is a simple and effective tool for monitoring adherence in ART and has been proved to be useful in adherence monitoring in a resource-limited setting [32,48]. This method employs pharmacy records to monitor adherence among patients when collecting their medication at the pharmacy. Patients collecting their medication regularly are said to be adherent by the pharmacy. Pharmacy records in resource limited settings are useful tools for measuring adherence; adherence measured by pharmacy records have been shown to correlate with CD4 count and weight gain in patients [72]. Adequate record keeping is necessary for pharmacy refill tracking information to be useful; the major setback is that the method only assumes that patients are adherent based on their empty pill boxes and the regular collection of their medication, it does not measure actual medication intake [32,48,72]. This method also requires patients to be collecting their medication at the same pharmacy for all refills which is not always possible [32,48,72].

Viral load level and CD4 counts are used as indicators for treatment outcomes and success. Viral load is regarded as the main indicator of the risk of therapeutic failure and can also be used in measuring adherence [42,75]. Low viral loads and increases in CD4 count are suggestive of good adherence although some patients may have a high viral load despite taking the ART regularly. CD4 is a good measure of adherence and correlated with weight gain, self-report and pharmacy records [72]. Viral load was also found to correlate with pill counts and good clinical outcomes in patients [42]. Pill counts may be a reliable and economical tool for monitoring adherence in resource-limited settings although viral load monitoring has been described as the preferred method [42]. The major drawback is cost and availability in resource limited settings [70].

In the provider estimation method of measuring adherence, the health care providers estimate a patient’s adherence based on factors such as socio-demographic and economic factors. In this method, there is no correlation with actual medication intake; studies have shown that health care providers overestimate adherence to HIV treatment when factors such as demographic characteristics are used to predict adherence because it is very difficult to predict adherence based on demography [16]. Furthermore, consistent clinic attendance does not usually correlate with other measures of adherence among patients on ART [72].

Therapeutic drug monitoring (TDM) involves measuring the levels of the drug in the blood stream of the HIV patient. This method is not used routinely as most ARVs have short circulating times in the body, coupled with the fact that it is very expensive [71]. Currently, therapeutic drug monitoring is restricted for research purposes. A study that combined TDM and viral resistance monitoring to find causes of virologic failure in patients on ART revealed that TDM is necessary in patients who developed adverse reactions while on a ritonavir boosting regimen because of the risk of drug toxicity in these patients [29]. TDM assists clinicians in choosing the best regimens for their patients and could be used routinely in clinical practice [29]. A prospective cohort study in Cameroon to compare adherence of fixed-dose combination of nevirapine, stavudine and lamivudine using nevirapine plasma level monitoring and self-report by patients found that self-reported adherence was significantly higher than adherence measures by nevirapine level monitoring [71]. The authors concluded that nevirapine plasma concentration monitoring provides an accurate measurement of adherence compared to self-report, but cautioned that it is not feasible in most clinical settings especially in resource-limited areas due to cost [71].

This entails simplifying regimen characteristics; simplification of dosing schedules, reduction of pill burden, and adjusting dietary restrictions to match patients' daily activities. In addition, identifying previous ART use and pre-existing medical conditions that could affect ART use and ensuring continuous provision of ART have also been reported to enhance adherence to ART [76]. Studies have demonstrated improved adherence with once or twice daily dosing of ART [77]. Many fixed-dose combinations (FDC) of ARVs are now available and have been shown to improve adherence in patients [65,77,78]. The buddy system has been widely used in resource-limited settings where relatives or friends agree to assist the patient in adhering to the medication [76]. Buddies remind patients to take their medication, encourage them and assist in keeping hospital appointments. This approach has led to improved ART adherence among some HIV-infected persons [76]. Social support not only assists the patient in adhering to ART but can also provide psychological support to the patient which helps them to cope with the disease [79].

This is usually the mainstay of ART programmes in any setting; Education and counsellingempowers the patient to be part of the treatment process. Knowledge about the disease, its symptoms, treatment and side-effects of the medications are crucial information that has to be passed on to the patient [80]. During counselling, potential barriers to adherence are identified and addressed. Counselling assists the patient in developing positive beliefs and perception towards the disease [81]. It also helps in setting goals and increases the self-efficacy of the patient [82]. A randomized controlled trial in the USA to compare the effect of person-to-person contact and support adherence with medication alarm techniques revealed improved responses to therapy and good adherence among the group that received interpersonal adherence support [83]. This finding is consistent with the literature where repeated supportive adherence has been described as the most effective intervention because it provides human contact and support [68].

These tools are often combined with other behavioural interventions [84]. Pill boxes are containers used for storing the ARV for regular use as prescribed. They enable the patient to take the medication correctly. Electronic versions of pill boxes with reminders to the patient are also available. The major setback of this tool is the lack of confidentiality and privacy associated with it, since some patients do not want people around them to see their medications. Another setback is the task of filling out the boxes; uneducated patients might not be able to do this correctly. Pill charts involve visual display of the pills in terms of their colour, shape, name and dosage of the medication during counselling. This is very useful especially among the uneducated patients [84]. Electronic devices such as beepers, alarms and watches that remind patients to take their medication according to the prescribed schedule are also used to enhance adherence [70]. In addition, telephone calls and mobile-phone text messages have been shown to improve medication adherence in HIV infected individuals [85,86,87,88]. Electronic pagers linked to the internet may also be used to send reminders to patients to take their drugs. Although reminder tools have been reported by patients as one of the facilitators of adherence [85,86,87,88], the major disadvantage of this strategy is the lack of privacy associated with it. Another shortcoming is the cost required to set up this kind of service and the challenges associated with the patient having their mobile phones with them all the time [88]. Medication diaries are very useful in understanding the patterns of drug use by the patient and the reasons for not taking the medication regularly. Diaries are used by the patients to document the time and date of taking the medications and missed doses and the reasons for it. This tool may also be used to identify side-effects or other problems that the patient may encounter in the course of taking ART. Research has shown that some tools appear to be of more benefit to patients when they are combined with patient education or counselling. A large multicenter randomized trial in the USA revealed that reminder devices alone do not enhance adherence and suggests that they should be combined with counselling as part of comprehensive support for patients on ART [83].

This technique has been used successfully in TB by asking patients to take medication under the supervision of adherence counsellors, trusted family members or community lay workers [89,90,91]. Directly administered ART can be clinic or home-based, once-daily, twice-weekly, or once-weekly [92]. The challenge in ART is different from TB as timing of medication intake varies across ARVs and the fact that the treatment is for life [93]. Although this approach has shown some benefit in improving adherence, research does not support routine use of directly administered ART [92,93,94]; Modified directly observed therapy (mDOT) has been suggested for use in ART and a home-based mDOT strategy has been used to improve adherence rates in resource-limited settings [92]. However, adherence to ART was only seen in the early weeks of a (mDOT) study in Kenya [94]. Despite its limited use, the benefits of mDOT in ART seen in early stages of treatment helped patients to develop good understanding of the treatment and develop good treatment taking behavior. It also assisted the patients in developing trusting relationships with the health care provider which subsequently improved their adherence to ART [92].

These include interventions that target barriers to ART adherence at the level of health care delivery of HIV services such as food supplements, transportation to health facilities, staffing and integration of services. Research has demonstrated that ART adherence and retention in care are associated with interventions that offer food supplements, address transportation issues and integrated services. Family nutritional support for HIV positive patients on ART improves adherence leading to good clinical outcomes, and this could be integrated into ART adherence interventions as an effective and comprehensive community-based primary care [95]. In a South African study, nurse-initiated ART services have been shown to produce the same outcomes as ART services run by doctors, which indicates that where human resources are limited, nurses can provide ART care [96]. In addition, addressing the issue of transportation to the health facilities among HIV patients and integration of home-based care into ART services led to improved attendance in clinic appointments [97]. ART regimens need to be chosen taking into consideration the patient’s working and family life, especially the cost implication of the drugs to ensure continuous supply so that adherence can be optimized [98].