This overview is intended to provide the context and scientific background to a number of key issues relating to the AIDS pandemic and in particular to the issue of treatment adherence. There are many important aspects of the pandemic which are not discussed in this overview - for example, the importance of prevention, strategies for harm reduction and the need to tackle stigma and address the needs of vulnerable populations. The UNAIDS 2006 Report on the global AIDS epidemic provides detailed information on these and other aspects of the epidemic.

Since it was first recognized in 1981, Acquired Immune Deficiency Syndrome (AIDS) has killed more than 25 million people worldwide, making it one of the most destructive pandemics in recorded history. In 2005 alone, AIDS claimed an estimated 2.8 million lives and an estimated 4.1 million people were newly infected with the virus. In 2005, an estimated 38.6 million people worldwide were living with the human immune deficiency virus (HIV), of whom 24.5 million live in sub-Saharan Africa (UNAIDS, 2006).

AIDS is caused by HIV, which can be spread through sperm, blood, breast milk and vaginal secretions. The most common route of transmission is unprotected sex. However, among particular risk groups, other methods of transmission may be dominant. For instance, among intravenous drug users the use of contaminated needles is a major cause of transmission. HIV spreads rapidly in situations where conflict, disasters, or socioeconomic pressures contribute to the displacement of people and disruption of family life. In these situations, women and children are especially vulnerable.

When a person is initially infected by HIV, they may experience a mild flu-like illness and then become asymptomatic, remaining well for an average of about eight years. However, throughout that period, they are able to infect other people. Once the infection progresses, the immune status of the individual becomes depressed and they become vulnerable to a range of different illnesses (opportunistic infections). The first of these is often tuberculosis, but many other conditions can emerge, including pneumonia, some cancers, meningitis and fungal infections. AIDS kills because the individual's immune system is not able to fight infections.

It is possible to measure the effect that HIV has on an individual by measuring the number of CD4 cells (key components of the immune system) in the blood stream and, where the required equipment is available, the actual viral load. The normal count is about 1000. Once the CD4 count falls, the individual is likely to suffer from various different infections unless they are treated with antiretroviral medicines (ARVs). When the CD4 count falls below 200, treatment with ARVs is usually started. However, if there are AIDS-defining conditions and a CD4 count below 350, treatment may also be started. People in these two categories are referred to as "people in need of treatment." Of particular concern in Africa is the impact of AIDS on the nutritional status of the individual. Wasting and weight loss are very common in AIDS patients. The causes of wasting are complex and due to a wide range of factors. As reported by Strawford and Hellerstein (1998), they include a reduction in appetite and calorific intake as well as more complex physical alterations caused by illness and stress. The three country studies which follow suggest that treatment causes an unusually large increase in appetite - especially during the first six months - which in these settings often cannot be met. The need to ensure nutritional support for AIDS patients was recognized by UN Member States in the Political Declaration issued at the UN General Assembly Special Session (UNGASS, 2006) on HIV/AIDS in New York from 31 May-2 June 2006).

Antiretroviral therapy (ART)

The introduction of ARVs in the 1990s brought new hope to people living with HIV. More recently, the increased availability of treatment has dramatically improved survival rates and lowered the incidence of opportunistic infections in people with AIDS (UNAIDS, 2005). Those who have access to ARVs and the care needed to maintain therapy can live for many years with what is now considered to be a complicated but manageable chronic disease. Although ARVs are not a cure for HIV, they are very effective in controlling the virus, and can even reduce the level of the virus to a point where it is no longer possible to detect any HIV in the blood. These medicines prevent HIV from multiplying rapidly and, at the same time, boost the body's immune system. In this way, they can increase the length and quality of life and enable people to lead full and productive lives. There are many kinds of ARVs, which attack the virus in different ways. For this reason, treatment today always involves a combination of ARVs. The most commonly used are from three classes: nucleoside reverse transcriptase inhibitors (NRTIs), non-nucleoside reverse transcriptase inhibitors (NNR-TIs) and protease inhibitors. A new class of drugs - entry inhibitors -prevent HIV from entering the T-cells, which are crucial in maintaining the individual's immune response. All ARVs have different side-effects and these can have an impact both on how the medicines will be used and how patients take them.

In some situations, several ARVs are combined in a single tablet, a fixed-dose combination (FDC), which ensures that patients always take multiple doses together. As a first-line treatment, antiretroviral therapy (ART) can be delivered relatively cheaply. However, if drug-resistance develops and these medicines are no longer effective, second-line ARVs may be required. These second-line ARVs are much more expensive than first-line medicines (see Figure 1).

Figure 1: Current prices (US$) of first- and second-line ARVs in Africa

Price of ARVs in Africa 2005

Source: Médecins Sans Frontières (MSF) with data from Clinton Foundation HIV/AIDS Initiative and WHO Global Price Reporting Mechanism

Epidemic in sub-Saharan Africa: new initiatives and treatment roll-out

Sub-Saharan Africa is the global epicentre of the AIDS pandemic and remains the region worst affected. Although it has little over 10% of the world's population, the region is home to more than 60% (24.5 million) of the people living with HIV. In 2005, an estimated 2 million people died of AIDS-related illnesses in this region, while a further 2.7 million people became newly infected with HIV (UNAIDS, 2006). The catastrophic impact of HIV in sub-Saharan Africa is threatening development in all sectors of society. While people of all ages are affected by HIV, most of those infected are in the 25-45 age group. This group is particularly important not only in terms of economic productivity but also as carers, parents and providers. The loss of productive workers and increases in spending on health care and social services require difficult decisions about resource allocations across all government sectors (UNAIDS, 2005).

In response to the epidemic and the need to ensure the availability of ARVs in resource-poor settings - a major focus of global advocacy efforts - a number of strategic partnerships and new organizations have been established and new initiatives launched. Meanwhile, the price of most first-line ARVs has fallen substantially due to the involvement of Indian generic pharmaceutical companies which have produced ARV formulations at prices that allow patients to be treated for less than US$ 1.00 per day. The first of these initiatives to improve ART availability was the Accelerating Access Initiative, a partnership established in 2000, involving UN organizations and a number of pharmaceutical companies which offered to make their products available at reduced prices. In 2001, the Global Fund to Fight AIDS, Tuberculosis and Malaria (the Global Fund) was established. Although essentially a financing organization, the Global Fund has assisted many countries in their efforts to obtain ARVs. The Global Fund aims to provide ART to 1.8 million people over five years and to support many more people with voluntary counselling and testing for HIV, medical services and community care. In 2003, WHO launched the "3 by 5" initiative, which aimed to ensure that 3 million patients were on ART by the end of 2005. Although this milestone was not reached, many hundreds of thousands of patients who would not otherwise have been treated were initiated on therapy, particularly in sub-Saharan Africa. Also in 2003, the United States Government established the President's Emergency Plan for AIDS Relief (PEPFAR). This initiative focuses on 15 high-burden countries and provides substantial resources, not only for ARVs, but also for systems support. The target of PEPFAR is to treat 2 million patients within 5 years.

This book focuses on the early treatment experiences of three countries in sub-Saharan Africa: Botswana, the United Republic of Tanzania and Uganda. Botswana received early support from both the Gates Foundation and Merck (the Merck Company Foundation/Merck & Co., Inc.). Tanzania has received support from both the Global Fund and PEPFAR. Uganda was very active during the Accelerating Access Initiative and in the "3 by 5" activities, and also benefits from PEPFAR funding. Thus, in all three countries, multiple organizations have been providing support. Access to ARVs has improved dramatically, and many patients have been started on treatment. However, taking ARVs twice a day is a challenging task for people who may be ill and who may also face obstacles to regular adherence.

Drug-resistance

Drug-resistance is a well-recognized biological phenomenon occurring with infectious organisms including bacteria, viruses and parasites such as malaria. When an infectious organism is exposed to an antibiotic, an antiviral agent or an antiparasitic agent such as chloroquine, some of the organisms will be very sensitive to the agent while others will be partially resistant, due to genetic variation. If the course of therapy is interrupted, those organisms that were sensitive will have been killed, while the more resistant organisms will have survived and can replicate. If there are repeated interruptions in treatment, only drug-resistant organisms will survive. In the case of HIV, the problem is complicated by the very rapid replication of the virus and the fact that it can go into a non-active state where it is not affected by ARVs. When the viral load is very high, as it is at the beginning of therapy, high adherence to ARVs is required (Carrieri et al., 2003). If an individual misses a few doses at this stage, the danger of the development of drug-resistant organisms is far higher than it would be after six months of regular treatment, by which time the person is likely to have a low viral load. To prevent the development of drug-resistance, an adherence level of at least 95% is required for the duration of therapy and especially in the first six months (Paterson et al., 2000). This is very difficult to achieve.

If an individual develops resistance to ARVs, two problems occur. Firstly, the first-line ARVs will no longer work and the individual will start to suffer from multiple opportunistic infections. Secondly, the individual may transmit the drug-resistant virus to their contacts and when those individuals go for treatment they will discover that their virus does not respond to the first-line therapy. Once a person develops resistance to first-line medicines they will need to be changed to second-line ARVs. However, at present these products are substantially more expensive than the first-line ARVs and have a different range of side-effects. Moreover, the decision to change a person from first-line to second-line therapy is a difficult one to make, especially if CD4 and viral load testing equipment is not available. At present, the cost of second-line therapy is about 10 times more than first-line therapy. The danger is that if too many patients progress to second-line therapy, the increased costs involved will limit access to treatment for many people who would benefit from first-line therapy. Therefore, every effort should be made to ensure a high level of adherence (at least 95%) to the first-line ARVs in order to delay the emergence of drug-resistance and enable individuals to be treated for many years with first-line ARVs.

Adherence

Drug-resistance is not the only cause of treatment failure. The natural history of HIV infection is very unpredictable and people respond to treatment regimes in different ways, as reported by O'Brien et al. (2000). Sub-optimal adherence itself is an important cause of failure. If people are sharing ARVs or interrupting their daily dosage regimes they simply do not get enough of the medicines for effective treatment and they will generate drug-resistance. Inappropriate use of ARVs is a multifaceted problem increasing the likelihood of drug-resistance and contributing to direct treatment failure. Policies and programmes that aim to provide increased or universal access to treatment face a key challenge: in order to succeed, these programmes need to achieve an exceptionally high level of adherence for an indefinite period of time. An extensive review of interventions for improving adherence was undertaken by Haynes et al. in 2005. The authors define adherence as the extent to which patients follow the instructions they are given for prescribed treatments. In this review, the term compliance, which is less commonly used today, is considered to be, in terms of measurement, the same as adherence. The term adherence is intended to be non-judgmental, a statement of fact rather than of blame of the patient, prescriber or treatment.

Measurement of adherence: particular requirements for ARVs

While adherence can be expressed quantitatively as a percentage of expected treatments actually taken, its accurate measurement is very difficult. There is no gold standard for adherence measurement. While adherence can be ensured by directly observed treatment, this may not be practical for twice-daily therapy which has to be taken for life. As a result, levels of adherence can only be estimated by use of indirect measures. The most commonly used methods include pill counts, pharmacy refill records, various self-reporting tools such as questionnaires and visual analogue scales, measurement of blood levels, and electronic drug monitors, which monitor the number of times the cap of a pill container is removed. A recent report from Malawi has compared three of these methods of measuring adherence. The authors compared four-day recall, visual analogue scale (using visual measurement tools to capture self-reports) and self-administered questionnaires, and found that self-reporting on the basis of four-day recall was the best predictor of viral load (Ferradini et al., 2006).

When reporting adherence to ART, two measures are frequently used. The first is the overall adherence recorded as the number of tablets taken correctly as a proportion of those which were prescribed. This measure is an important marker for the clinical evaluation of individual patients, as well as for counselling purposes. The second measure, which is more sensitive to the particular requirements of ART, reports the percentage of patients taking at least 95% of their tablets correctly, and is essential for programmatic or public health evaluation.

When reporting population adherence rates, it is critically important to include the percentage of patients who achieve the optimal adherence rate of at least 95%. Thus the average (mean) overall population adherence rate may be 85% (i.e. 85% of all pills that should have been taken were taken) but only 60% of this same population may have achieved the optimal adherence rate of at least 95%. The first measure of overall adherence is important for individual counselling, while the second, which is a population measure of adherence, is useful for evaluation and planning. Although having both measures in any study complicates analysis, both are needed, in particular the second measure. Both measures can be collected in the same way with pill counts, patient recall, visual analogue scales or electronic monitoring.

Adherence to treatment for chronic diseases

Considerable experience exists in dealing with the challenges of long-term adherence to medication for chronic diseases other than AIDS. In 2004, DiMatteo reported on 50 years' research on adherence to medical recommendations. Within the more than 500 studies, adherence ranged from as low as 4.6% to 100% with an average of about 75% adherence for the different conditions. Only eight studies of the treatment of AIDS were included and the mean adherence in these studies was 88%. Vermeire et al., (2001) reviewed patients' adherence to treatment over the past 30 years. In addition to identifying the difficulties in measurement, they discuss the causes of poor adherence, pointing out that these are not well understood. They discuss aspects such as features of the disease, the referral process, the clinical setting and the therapeutic regimen, noting that these do not appear to affect adherence. Demographic variables are also poor indicators of adherence. Factors related to poor adherence include psychiatric disorders, treatment factors such as the duration of treatment, the number of medicines prescribed, the cost, and the frequency of dosages. Factors which can increase adherence include increased supervision, and parenteral administration. Surprisingly, side-effects are rarely seen as a reason for sub-optimal adherence. The patient's own beliefs and lifestyle are found to be important in affecting adherence. Social factors such as having positive attitudes in the community are also seen to improve adherence. The quality of the doctor-patient relationship is regarded as a very important factor.

However, the amount of research which has measured this is limited. The Uganda country report in the present publication provides an extensive review of the existing literature.

Many reasons exist for non-adherence to medical regimens, including problems with the regimen such as: adverse effects, poor instructions, poor provider-patient relationship, poor memory, patients' disagreement with the need for treatment, problems in accessing the facilities providing treatment or inability to pay for it (Haynes et al., 2005).

ARVs: a special case

While experience of adherence to treatment for chronic diseases has great relevance, ART is exceptional in requiring adherence rates of at least 95% to prevent treatment failure or drug-resistance. In 2000, Chesney reviewed the factors affecting adherence to ART, based on the US experience. She reported a strong correlation between the level of medication adherence and the percentage of patients without detectable level of the virus. When patients achieved less than 80% adherence, 87% of them had detectable virus. When adherence was between 80% and 90%, treatment failure occurred in 47% of the patients; and when adherence was more than 95%, detectable virus occurred in only 10% of patients. Although this occurred in a small number of patients, it does underline the importance of high levels of adherence. There were many factors which were reported to negatively affect adherence. These were grouped under patient factors, medication factors and the system of care factors. The reasons for sub-optimal adherence included the patients forgetting or being too busy, being away from home, a change in daily routine, side-effects, depression or illness and a lack of interest or desire to take a "drug holiday". A number of strategies were suggested to help improve adherence. These included directly observed treatment, providing clear instructions, using personal treatment plans, tailoring the drug regimen to suit the patient's lifestyle, showing patients how to keep a diary, and making the clinic appointments both convenient and pleasant. The evidence as to which intervention was most effective was not presented.

In an opinion piece published in 2005, Gill et al. suggested that there was no room for complacency about adherence to ARVs in sub-Saharan Africa. They pointed out that after initial optimism about adherence in Africa, recent reports reflected medium-to-poor adherence in the few studies that reported longitudinal data. In particular, they quoted the experience from Senegal, where 95% of patients had adherence levels exceeding 80% after one month of therapy, but after 18 months only 80% remained above that level. The 80% level of adherence would not be high enough to prevent treatment failure or the development of drug-resistance. Meanwhile, the percentage of patients with undetectable viral loads fell from 80% to 59% over time. In Cameroon, Akam was quoted as reporting that mean self-report adherence was initially under 68% and declined further over time. One of the key conclusions of the Gill et al. paper was that external multinational funds should be allocated to supporting and studying adherence. The authors also suggested that qualitative research into the behavioural reasons for patient non-adherence is urgently needed. The studies presented in this publication are intended to help fill that gap.