Clinical Guidelines > Archived Documents > Adherence (2001) >

4.0 How can high adherence be promoted?

The definitive intervention to promote adherence to antiretroviral therapy has yet to be developed. To date, interventions have either been ineffective or have produced only short-term benefits. One explanation is that single interventions fail to take account of the full range of enablers of, and barriers to, high adherence. Research conducted over the last five years has improved our understanding of the psychology of adherence and identified some of these enablers and barriers. Recent research suggests that adherence interventions are likely to be more effective if they adopt a multi-factorial approach, taking account of the patient’s perceptions as well as the practicalities of taking treatment22 23. The following section will review this literature in order to provide a series of evidence-based recommendations for provision of adherence support interventions in clinical and community-based settings. These are considered in three categories: patient factors, provider factors and regimen factors. It is recognised that divisions of this kind are somewhat arbitrary, and that there is a degree of overlap between them.

4.1 Patient factors

Patient commitment: The optimal time to begin antiretroviral therapy is not known and expert opinion on the matter is divided, and subject to revision24 25. Individuals who are committed to their treatment and have been actively involved in its selection may be more likely to achieve the high levels of adherence required for successful HIV therapy. Though various constituents have a role to play in fostering and maintaining high adherence, the role of patients themselves is fundamental.

Demographics: Low adherence is common and hard to predict. Devising effective and appropriate means of identifying and predicting who will be likely to adhere poorly has become a significant challenge for health care professionals, and a substantial proportion of the literature on adherence to HIV therapies reflects this. Studies investigating the role of socio-demographic characteristics such as gender, race, age, exposure category, and educational level as predictors of adherence, have however, produced largely inconsistent results26. The desire to link low adherence to (often deprived) social groups is a well-established thread in the general literature, dating back to the 1900s when tuberculosis control occupied New York’s public health officials27. However, as later experience with antibiotics would demonstrate, low adherence is not restricted to certain social classes, but is widespread and unpredictable28. Research in the HIV field supports this perspective. Moreover, adherence rates vary not just between individuals, but within the same individual over time29. Adherence is therefore best thought of as a variable behaviour rather than as a stable characteristic of an individual – most people will exhibit low adherence some of the time30.

Cultural and socio-economic issues have great potential to impact on adherence. For example, religious beliefs around illness and medication may influence motivation and adherence; medication use may disclose HIV status; poverty may prevent individuals from following dietary advice; drug and alcohol use may impair judgement and the ability to adopt and maintain routine medication use; family responsibilities may require adults to place the health care needs of others before their own. However, these are issues to be considered and resolved on an individual basis, and there is good evidence that dedicated programmes can assist adherence in a range of populations who may be assumed to have special needs, including drug users and the homeless31.

Psychological factors: Mental health problems such as depression have been associated with low adherence in HIV-positive adults32 and adolescents33, as have other psychological variables such as perception of one’s ability to follow a medication regimen, or ‘self-efficacy’34 35.

Health beliefs: Beliefs about health and illness, in particular about the necessity of medication to ward off illness, and concerns about its potential adverse events, have been found influential in both HIV disease36 37, and in other disease areas including asthma, diabetes, kidney disease, heart disease and cancer38 39 40 41. The effect of symptom experiences on views about medication may be complex. At one level symptoms may stimulate medication use by acting as a reminder or reinforcing beliefs about its necessity. However, patients’ expectations of symptom relief are also likely to have an important effect. This could be problematic if expectations are unrealistic, or where treatment is prescribed for asymptomatic disease, as is frequently the case in HIV infection42. In addition, patients’ concerns about HAART’s potential harm may be entirely rational, and indeed Horne and colleagues have proposed that for some individuals, missed doses may be a logical attempt to moderate this risk by taking less medication43. People with HIV themselves frequently attribute low adherence to the experience of side-effects44; both transient events and longer-term problems such as lipodystrophy45.

Recommendations

  1. Engage the individual in discussion about their knowledge and understanding of the need for HAART and their concerns about taking it. Regularly evaluate their concerns about anti-HIV therapy and address specific concerns and misconceptions promptly.
  2. Evaluate the patient’s motivation and commitment to taking anti-HIV therapy prior to starting, and at regular intervals following initiation. If this raises persistent concerns, motivational interviewing techniques can be employed to help patients strengthen their intentions and adherence behaviour46 47. Peer interventions have been argued to be effective in a number of sexual behaviour settings48 49 50 51, and the use of credible peers to discuss and promote adherence could be adapted to influence group adherence motivation.
  3. Mental health factors should be assessed regularly and significant mental health problems managed through appropriate referral to clinical psychology, psychiatry or other allied health professional, prior to starting anti-HIV therapy. All HIV health care professionals should be aware of signs of depression and other forms of psychological distress.
  4. Assess the extent to which the patient has the behavioural skills to adhere to therapy. Where skills deficits are an issue, input from nursing, pharmacy, health advising and psychology can help develop skills needed to manage adherence behaviour better. Such difficulties include establishing pill-taking routines, prompts, timetabling, assertiveness, strategies for travel, managing disclosure or discovery by others, and problem solving.
  5. The manner in which environmental and social factors may influence adherence should be assessed regularly. Such factors include housing, employment, relationships, drug and alcohol use, and immigration status. Health care professionals should be familiar with social care services provided by local statutory and community-based organisations, and provide referrals as appropriate.

4.2 Provider factors

Offer support universally: Given that low adherence is widespread and that predictions about future adherence behaviour cannot be made on the basis of demographic characteristics, there can be no justification for withholding treatment solely on the grounds of assumptions about an individual’s personal circumstances. Adherence support services must be provided to all patients in recognition that any individual is capable of both low and high adherence.

Support when changing therapy: Virological response in patients who change their failing antiretroviral therapy is improved in those who report high adherence to their new regimen. This suggests that therapy change is a key time-point for the type of adherence support interventions which may be more routinely available to individuals initiating therapy. Data from the ARGENTA study suggest that the use of genotypic testing when changing a failing regimen is enhanced in patients who report high adherence compared to those reporting low adherence52. ACTG 370, which switched patients on dual nucleoside therapy to a HAART regimen, found that low adherence to prior nucleoside therapy predicted low adherence to HAART, and consequently to an increased risk of failure on HAART53. Further, analysis of both antiretroviral naïve and experienced participants in two large CPCRA studies found adherence to be poorer in those taking second-line therapy54.

Patient education: Patients who understand the rationale for anti-HIV therapy and the role of adherence in the development of drug resistance and treatment failure report higher adherence levels than those without this information55 56 57 58. Reinforcing information provided verbally with written information to take home, and by checking that information delivered has been heard correctly is likely to be beneficial as patients commonly misunderstand their health care providers’ instructions. One study found that 13% of patients prescribed HAART were not taking their medication correctly, despite believing that they were59. In addition, the ATHENA study reported that 50% of patients with low plasma nelfinavir levels who were counselled regarding nelfinavir’s dietary requirements, had drug levels above the target level when next measured60. The HIV voluntary sector is resourced to provide patient education. Partnerships between clinical settings and community-based organisations, including those who target different ethnic communities, can improve the uptake of information resources61.

Medication alerts: Numerous medication alert strategies are available, from programmable wrist watches to SMS text messaging software, and the frequency with which HIV-positive people cite forgetfulness as the cause of missed doses suggests these may play an important role in establishing a routine when beginning a new regimen62 63 64. Telephone or postcard reminders, pill diaries and charts, medication containers, and the enlistment of family and friends as reminders have all been shown to improve adherence in other disease areas65 66.

Multi-disciplinary approach: The limited time available for doctor-patient consultation and the appropriate use of skilled professionals allied to medicine, point towards a multidisciplinary approach to HIV care; a model already in relatively common use within the UK67. A pilot study investigating the effects of a six month, nurse-led educational intervention reported improved virological responses in treatment experienced patients68, whilst pharmacist-led programmes have also been shown to improve adherence levels in people receiving HAART69.

Provision of ongoing support: Low adherence even to very short medication courses is common. A 1960s study of antibiotic use found that only 19% of patients were able to follow a ten day treatment course70. High adherence is a process, not a single event, and adherence support must therefore be integrated into clinical follow-up. Data from large studies involving both naïve and experienced patients indicate that the likelihood of maintaining 100% adherence diminishes over time71. Tuldra and colleagues, who undertook a randomised, controlled trial (RCT) of a ‘psychoeducative’ intervention to improve adherence to HAART (see table 2), reported a need for ongoing support if high adherence is to be maintained72. A number of other RCTs have demonstrated that one-off interventions have only a short-lived benefit73 74 75.

Professional development: The move from the use of the word compliance to adherence in the late 1990s was in part a recognition that effective relationships between health care professionals and patients should not be expected where there is undue emphasis on the submission of patients to their doctors’ demands. There is growing recognition that optimal use of medicines involves a partnership between clinicians, patients and others. Adherence to antiretroviral therapy is improved where patients view their relationship with their doctor positively76. Patients vary in their desire to play an active part in treatment decisions. Exploring patients’ preferences for involvement and structuring consultations accordingly may act as a driver to adherence. A validated scale has recently been developed to address these issues in HIV therapy77.

It should be recognised however, that few HIV professionals have received specific training around treatment adherence78, and there may be benefits in the establishment of accredited programmes run by professional bodies.

Provision of Post-exposure prophylaxis (PEP): Tolerability of both indinavir- and nevirapine-based HAART regimens prescribed following occupational or sexual exposure to HIV has been reported to be low, raising concerns about adherence in such circumstances (though adherence levels required for successful PEP are not known)79 80. Participants in the San Francisco PEP study received medication adherence counselling at baseline, week 1, and week 2 of a four week antiretroviral treatment course, most frequently involving dual therapy with AZT/3TC, given as Combivir81. Counselling involved a review of the treatment regimen, identification of possible dosing cues, and discussion of potential barriers to adherence. In addition, medication was dispensed on a weekly or two weekly basis. Participants self-reported their adherence once each week by recalling the number of doses taken over the prior four days. Seventy-eight percent of individuals completed the four week course. During this period, those reporting having taken all of their medication varied between 78% and 84%, despite significant adverse events.

Obstetric care: Low adherence to antiretroviral therapy has been reported to be common amongst both pregnant and postpartum women82 83.These studies involved a large cohort of women in New York state who were in receipt of Medicaid-supported care during the period between 1993-96. Adherence was measured using pharmacy reports. There seem to be no data on interventions aimed specifically at pregnant women. Pregnant women’s concerns about the effects of medication on the foetus are likely to be an additional factor in this context, and should be addressed.

Paediatric care: Though many aspects of this document are relevant to the care of HIV-positive children, a number of distinct issues should also be noted, and these are reviewed in more detail elsewhere84. In the UK, most HIV infection in children has been acquired vertically. As such, HIV in children is generally considered and represented within a family context, and managed using a family-centred, multi-disciplinary model of care. This approach can be crucial in dealing with medical, nursing, social, psychological and nutritional aspects of antiretroviral treatment. Children are not a homogenous group, however. Age-specific differences are important in terms of needs, including those which relate to adherence. For example, food and fasting restrictions can be particularly difficult to enforce in infants and young children, whilst school age children may need to avoid regimens which may disclose HIV status. Swallowing tablets or unpalatable liquids may be difficult for children of all ages, and developmental progress may bring with it behavioural changes which challenge routine medication use. Observational data from a small group of HIV-infected children treated with combination antiretroviral therapy suggest that improved formulations, counselling of families and children, and tailoring regimen selection may improve adherence in this group85. Where more than one family member is being treated, there may be benefits in aligning their regimens where appropriate.

Recommendations

  1. Every HIV treatment centre should have a written adherence strategy in place which should be audited regularly in the context of the growing knowledge base.
  2. Interventions to improve adherence must be offered to all patients taking HAART.
  3. Adherence support interventions should be offered when starting therapy, changing therapy, and as part of ongoing routine follow-up.
  4. Providers must ensure that patients have sufficient understanding of HIV pathogenesis; the rationale for anti-HIV therapy; the relationship between adherence and resistance; the requirements of their regimen; and its potential side-effects (see appendix I). The support of voluntary sector resources should be enlisted, and verbal information should be supported by written information. Early follow-up, including a telephone assessment (performed by, for example, a community nurse 24-48 hours after medication has been dispensed) should evaluate whether patients have continuing information needs (though this has not been demonstrated to improve adherence to HAART).
  5. Medication alerts and containers should be provided as appropriate.
  6. Interventions offered by treatment centres should involve multi-disciplinary input.
  7. HIV professionals require continuing education in adherence issues, including professional development and skills-based training.
  8. Multiple interventions are likely to be necessary to maintain high adherence amongst clinic populations who are administered chronic therapy (see tables I and II).

4.3 Regimen factors

Lifestyle assessment: BHIVA recommend that, in the absence of definitive data regarding choice of therapy, regimens should be selected on an individual basis86. In addition to differing drug histories, patients will have varying needs and preferences in relation to toxicity, dosing, and avoidance of harmful drug interactions. A range of methods may be adopted to assist patients to consider the likely fit of any given regimen with their own lifestyle. The use of these methods prior to starting or changing treatment may help patients and health care workers foresee likely problems and devise coping strategies.

Dosing frequency: A recent systematic review indicates that medication which is dosed three or four times a day is associated with lower adherence levels than regimens dosed once or twice daily87. An open-label trial reported that virological responses to a twice daily HAART regimen (abacavir/Combivir) were at least comparable to one dosed three times daily (indinavir/Combivir) in naïve patients after 48 weeks therapy88. Adherence to therapy, measured by self-reported recall, was significantly higher in abacavir recipients (72% versus 45% reported taking more than 95% of their medication). Regimen factors which may be assumed to have influenced adherence in this study include twice daily versus three times daily dosing, differences in pill burden89, and differences in food and fluid requirements90. However, this is not an exhaustive list, and the relative weights which individuals may place on each of these factors is not reflected in a crude analysis of this kind.

Delayed dosing and pharmacokinetic issues: An analysis of adherence to PI-based HAART found no difference between twice daily and three times daily regimens with regard to the number of doses taken, but found that fewer doses were taken on time in the three times daily group91. Inter-patient and inter-regimen pharmacokinetic variability make the effects of late dosing difficult to predict, though data from a comparative study of different electronic MEMS (Medication Event Monitoring System) systems in people taking PI-containing HAART, found treatment failure was predicted by dosing late by one or by two hours in multivariate analysis92. While the longer half-lives of NNRTIs are often assumed to offer greater leeway over the timing of doses, in the absence of data regarding the effect of late dosing on treatment outcomes, it may be safer to advise all patients to space their antiretroviral doses consistently. Such regimen factors are also likely to be crucial to the outcome of treatment interruptions93.

Drug class: A longitudinal study involving over 1,100 naïve and experienced HAART recipients found higher levels of adherence in NNRTI recipients compared to PI recipients, after four and eight months of follow-up94. However, participants in these studies were taking a variety of regimens, and it is unclear therefore if this broad distinction is helpful. Moreover, a prospective, randomised comparison of antiretroviral-experienced individuals found no difference in adherence levels between those treated with an efavirenz-containing regimen versus a protease inhibitor-containing regimen95. The advent of improved formulations and pharmacokinetic enhancement means that regimens with low pill burden, and with no more than twice daily dosing, are now available in all classes.

Pill burden: Meta-analysis of eighteen randomised trials of nineteen HAART regimens, in treatment naïve individuals, found that lower pill burden was associated with the likelihood of having viral load below 50 copies after 48 weeks96. The extent to which pill burden may be a marker for other ‘adherence factors’ is unclear, however. The longitudinal study noted above found adherence was not associated with daily pill burden97.

Reliance on compact regimens: Though the term ‘treatment simplification’ has become commonplace in HIV therapy, whether any antiretroviral regimen prescribed for chronic use should truly be considered ‘simple’ is a moot point. An open-label trial found that 72% of abacavir/Combivir recipients reported adherence levels above 95%, and the remainder below this98. These data suggest that compact regimens (in this case, two pills, dosed twice daily, without regard to food) are not a panacea. Moreover, ‘treatment simplification’ is also regimen change and as such is a critical point for adherence review and support.

Drug interactions and side-effects: Additional adherence factors relating to specific regimens include the potential for harmful drug interactions (which should be expected to be particularly important in people with co-morbid conditions), and side-effects, which are frequently cited by patients as a common cause of missed doses99. The influence of side-effects may be both cognitive and practical, for example where doses are missed due to nausea, vomiting, diarrhoea, or where fatigue causes individuals to sleep through doses.

Recommendations

  1. Evaluate the extent to which lifestyle factors such as eating, sleeping and working patterns may impede adherence to a proposed regimen.
  2. Individual preferences regarding regimen characteristics, including formulations and pill size, should be assessed rather than assumed; what is simple for one person may not be for another100. Regimens which are dosed once or twice daily, which do not require fasting or dietary modification, and which have lower pill burdens, are likely to enable higher adherence and should be prescribed where appropriate, according to patient preference. Patients should always see the pills they may be prescribed prior to regimen selection.
  3. Undergoing a ‘dry run’, where antiretrovirals are substituted by placebos, may be useful in identifying possible adherence barriers101.
  4. Potential for drug interactions should be assessed prior to starting therapy. Ongoing assessment should be performed by physician and/or pharmacist whenever drugs are dispensed.
  5. Side-effects require prompt palliation. Patients can facilitate this process where they have a clear understanding of the likely toxicities associated with their regimen, appropriate self-management, and processes for accessing clinical support both in and out of hours.

4.1 Patient factors

Patient commitment: The optimal time to begin antiretroviral therapy is not known and expert opinion on the matter is divided, and subject to revision24 25. Individuals who are committed to their treatment and have been actively involved in its selection may be more likely to achieve the high levels of adherence required for successful HIV therapy. Though various constituents have a role to play in fostering and maintaining high adherence, the role of patients themselves is fundamental.

Demographics: Low adherence is common and hard to predict. Devising effective and appropriate means of identifying and predicting who will be likely to adhere poorly has become a significant challenge for health care professionals, and a substantial proportion of the literature on adherence to HIV therapies reflects this. Studies investigating the role of socio-demographic characteristics such as gender, race, age, exposure category, and educational level as predictors of adherence, have however, produced largely inconsistent results26. The desire to link low adherence to (often deprived) social groups is a well-established thread in the general literature, dating back to the 1900s when tuberculosis control occupied New York’s public health officials27. However, as later experience with antibiotics would demonstrate, low adherence is not restricted to certain social classes, but is widespread and unpredictable28. Research in the HIV field supports this perspective. Moreover, adherence rates vary not just between individuals, but within the same individual over time29. Adherence is therefore best thought of as a variable behaviour rather than as a stable characteristic of an individual – most people will exhibit low adherence some of the time30.

Cultural and socio-economic issues have great potential to impact on adherence. For example, religious beliefs around illness and medication may influence motivation and adherence; medication use may disclose HIV status; poverty may prevent individuals from following dietary advice; drug and alcohol use may impair judgement and the ability to adopt and maintain routine medication use; family responsibilities may require adults to place the health care needs of others before their own. However, these are issues to be considered and resolved on an individual basis, and there is good evidence that dedicated programmes can assist adherence in a range of populations who may be assumed to have special needs, including drug users and the homeless31.

Psychological factors: Mental health problems such as depression have been associated with low adherence in HIV-positive adults32 and adolescents33, as have other psychological variables such as perception of one’s ability to follow a medication regimen, or ‘self-efficacy’34 35.

Health beliefs: Beliefs about health and illness, in particular about the necessity of medication to ward off illness, and concerns about its potential adverse events, have been found influential in both HIV disease36 37, and in other disease areas including asthma, diabetes, kidney disease, heart disease and cancer38 39 40 41. The effect of symptom experiences on views about medication may be complex. At one level symptoms may stimulate medication use by acting as a reminder or reinforcing beliefs about its necessity. However, patients’ expectations of symptom relief are also likely to have an important effect. This could be problematic if expectations are unrealistic, or where treatment is prescribed for asymptomatic disease, as is frequently the case in HIV infection42. In addition, patients’ concerns about HAART’s potential harm may be entirely rational, and indeed Horne and colleagues have proposed that for some individuals, missed doses may be a logical attempt to moderate this risk by taking less medication43. People with HIV themselves frequently attribute low adherence to the experience of side-effects44; both transient events and longer-term problems such as lipodystrophy45.

Recommendations

  1. Engage the individual in discussion about their knowledge and understanding of the need for HAART and their concerns about taking it. Regularly evaluate their concerns about anti-HIV therapy and address specific concerns and misconceptions promptly.
  2. Evaluate the patient’s motivation and commitment to taking anti-HIV therapy prior to starting, and at regular intervals following initiation. If this raises persistent concerns, motivational interviewing techniques can be employed to help patients strengthen their intentions and adherence behaviour46 47. Peer interventions have been argued to be effective in a number of sexual behaviour settings48 49 50 51, and the use of credible peers to discuss and promote adherence could be adapted to influence group adherence motivation.
  3. Mental health factors should be assessed regularly and significant mental health problems managed through appropriate referral to clinical psychology, psychiatry or other allied health professional, prior to starting anti-HIV therapy. All HIV health care professionals should be aware of signs of depression and other forms of psychological distress.
  4. Assess the extent to which the patient has the behavioural skills to adhere to therapy. Where skills deficits are an issue, input from nursing, pharmacy, health advising and psychology can help develop skills needed to manage adherence behaviour better. Such difficulties include establishing pill-taking routines, prompts, timetabling, assertiveness, strategies for travel, managing disclosure or discovery by others, and problem solving.
  5. The manner in which environmental and social factors may influence adherence should be assessed regularly. Such factors include housing, employment, relationships, drug and alcohol use, and immigration status. Health care professionals should be familiar with social care services provided by local statutory and community-based organisations, and provide referrals as appropriate.

Recommendations

  1. Engage the individual in discussion about their knowledge and understanding of the need for HAART and their concerns about taking it. Regularly evaluate their concerns about anti-HIV therapy and address specific concerns and misconceptions promptly.
  2. Evaluate the patient’s motivation and commitment to taking anti-HIV therapy prior to starting, and at regular intervals following initiation. If this raises persistent concerns, motivational interviewing techniques can be employed to help patients strengthen their intentions and adherence behaviour46 47. Peer interventions have been argued to be effective in a number of sexual behaviour settings48 49 50 51, and the use of credible peers to discuss and promote adherence could be adapted to influence group adherence motivation.
  3. Mental health factors should be assessed regularly and significant mental health problems managed through appropriate referral to clinical psychology, psychiatry or other allied health professional, prior to starting anti-HIV therapy. All HIV health care professionals should be aware of signs of depression and other forms of psychological distress.
  4. Assess the extent to which the patient has the behavioural skills to adhere to therapy. Where skills deficits are an issue, input from nursing, pharmacy, health advising and psychology can help develop skills needed to manage adherence behaviour better. Such difficulties include establishing pill-taking routines, prompts, timetabling, assertiveness, strategies for travel, managing disclosure or discovery by others, and problem solving.
  5. The manner in which environmental and social factors may influence adherence should be assessed regularly. Such factors include housing, employment, relationships, drug and alcohol use, and immigration status. Health care professionals should be familiar with social care services provided by local statutory and community-based organisations, and provide referrals as appropriate.

4.2 Provider factors

Offer support universally: Given that low adherence is widespread and that predictions about future adherence behaviour cannot be made on the basis of demographic characteristics, there can be no justification for withholding treatment solely on the grounds of assumptions about an individual’s personal circumstances. Adherence support services must be provided to all patients in recognition that any individual is capable of both low and high adherence.

Support when changing therapy: Virological response in patients who change their failing antiretroviral therapy is improved in those who report high adherence to their new regimen. This suggests that therapy change is a key time-point for the type of adherence support interventions which may be more routinely available to individuals initiating therapy. Data from the ARGENTA study suggest that the use of genotypic testing when changing a failing regimen is enhanced in patients who report high adherence compared to those reporting low adherence52. ACTG 370, which switched patients on dual nucleoside therapy to a HAART regimen, found that low adherence to prior nucleoside therapy predicted low adherence to HAART, and consequently to an increased risk of failure on HAART53. Further, analysis of both antiretroviral naïve and experienced participants in two large CPCRA studies found adherence to be poorer in those taking second-line therapy54.

Patient education: Patients who understand the rationale for anti-HIV therapy and the role of adherence in the development of drug resistance and treatment failure report higher adherence levels than those without this information55 56 57 58. Reinforcing information provided verbally with written information to take home, and by checking that information delivered has been heard correctly is likely to be beneficial as patients commonly misunderstand their health care providers’ instructions. One study found that 13% of patients prescribed HAART were not taking their medication correctly, despite believing that they were59. In addition, the ATHENA study reported that 50% of patients with low plasma nelfinavir levels who were counselled regarding nelfinavir’s dietary requirements, had drug levels above the target level when next measured60. The HIV voluntary sector is resourced to provide patient education. Partnerships between clinical settings and community-based organisations, including those who target different ethnic communities, can improve the uptake of information resources61.

Medication alerts: Numerous medication alert strategies are available, from programmable wrist watches to SMS text messaging software, and the frequency with which HIV-positive people cite forgetfulness as the cause of missed doses suggests these may play an important role in establishing a routine when beginning a new regimen62 63 64. Telephone or postcard reminders, pill diaries and charts, medication containers, and the enlistment of family and friends as reminders have all been shown to improve adherence in other disease areas65 66.

Multi-disciplinary approach: The limited time available for doctor-patient consultation and the appropriate use of skilled professionals allied to medicine, point towards a multidisciplinary approach to HIV care; a model already in relatively common use within the UK67. A pilot study investigating the effects of a six month, nurse-led educational intervention reported improved virological responses in treatment experienced patients68, whilst pharmacist-led programmes have also been shown to improve adherence levels in people receiving HAART69.

Provision of ongoing support: Low adherence even to very short medication courses is common. A 1960s study of antibiotic use found that only 19% of patients were able to follow a ten day treatment course70. High adherence is a process, not a single event, and adherence support must therefore be integrated into clinical follow-up. Data from large studies involving both naïve and experienced patients indicate that the likelihood of maintaining 100% adherence diminishes over time71. Tuldra and colleagues, who undertook a randomised, controlled trial (RCT) of a ‘psychoeducative’ intervention to improve adherence to HAART (see table 2), reported a need for ongoing support if high adherence is to be maintained72. A number of other RCTs have demonstrated that one-off interventions have only a short-lived benefit73 74 75.

Professional development: The move from the use of the word compliance to adherence in the late 1990s was in part a recognition that effective relationships between health care professionals and patients should not be expected where there is undue emphasis on the submission of patients to their doctors’ demands. There is growing recognition that optimal use of medicines involves a partnership between clinicians, patients and others. Adherence to antiretroviral therapy is improved where patients view their relationship with their doctor positively76. Patients vary in their desire to play an active part in treatment decisions. Exploring patients’ preferences for involvement and structuring consultations accordingly may act as a driver to adherence. A validated scale has recently been developed to address these issues in HIV therapy77.

It should be recognised however, that few HIV professionals have received specific training around treatment adherence78, and there may be benefits in the establishment of accredited programmes run by professional bodies.

Provision of Post-exposure prophylaxis (PEP): Tolerability of both indinavir- and nevirapine-based HAART regimens prescribed following occupational or sexual exposure to HIV has been reported to be low, raising concerns about adherence in such circumstances (though adherence levels required for successful PEP are not known)79 80. Participants in the San Francisco PEP study received medication adherence counselling at baseline, week 1, and week 2 of a four week antiretroviral treatment course, most frequently involving dual therapy with AZT/3TC, given as Combivir81. Counselling involved a review of the treatment regimen, identification of possible dosing cues, and discussion of potential barriers to adherence. In addition, medication was dispensed on a weekly or two weekly basis. Participants self-reported their adherence once each week by recalling the number of doses taken over the prior four days. Seventy-eight percent of individuals completed the four week course. During this period, those reporting having taken all of their medication varied between 78% and 84%, despite significant adverse events.

Obstetric care: Low adherence to antiretroviral therapy has been reported to be common amongst both pregnant and postpartum women82 83.These studies involved a large cohort of women in New York state who were in receipt of Medicaid-supported care during the period between 1993-96. Adherence was measured using pharmacy reports. There seem to be no data on interventions aimed specifically at pregnant women. Pregnant women’s concerns about the effects of medication on the foetus are likely to be an additional factor in this context, and should be addressed.

Paediatric care: Though many aspects of this document are relevant to the care of HIV-positive children, a number of distinct issues should also be noted, and these are reviewed in more detail elsewhere84. In the UK, most HIV infection in children has been acquired vertically. As such, HIV in children is generally considered and represented within a family context, and managed using a family-centred, multi-disciplinary model of care. This approach can be crucial in dealing with medical, nursing, social, psychological and nutritional aspects of antiretroviral treatment. Children are not a homogenous group, however. Age-specific differences are important in terms of needs, including those which relate to adherence. For example, food and fasting restrictions can be particularly difficult to enforce in infants and young children, whilst school age children may need to avoid regimens which may disclose HIV status. Swallowing tablets or unpalatable liquids may be difficult for children of all ages, and developmental progress may bring with it behavioural changes which challenge routine medication use. Observational data from a small group of HIV-infected children treated with combination antiretroviral therapy suggest that improved formulations, counselling of families and children, and tailoring regimen selection may improve adherence in this group85. Where more than one family member is being treated, there may be benefits in aligning their regimens where appropriate.

Recommendations

  1. Every HIV treatment centre should have a written adherence strategy in place which should be audited regularly in the context of the growing knowledge base.
  2. Interventions to improve adherence must be offered to all patients taking HAART.
  3. Adherence support interventions should be offered when starting therapy, changing therapy, and as part of ongoing routine follow-up.
  4. Providers must ensure that patients have sufficient understanding of HIV pathogenesis; the rationale for anti-HIV therapy; the relationship between adherence and resistance; the requirements of their regimen; and its potential side-effects (see appendix I). The support of voluntary sector resources should be enlisted, and verbal information should be supported by written information. Early follow-up, including a telephone assessment (performed by, for example, a community nurse 24-48 hours after medication has been dispensed) should evaluate whether patients have continuing information needs (though this has not been demonstrated to improve adherence to HAART).
  5. Medication alerts and containers should be provided as appropriate.
  6. Interventions offered by treatment centres should involve multi-disciplinary input.
  7. HIV professionals require continuing education in adherence issues, including professional development and skills-based training.
  8. Multiple interventions are likely to be necessary to maintain high adherence amongst clinic populations who are administered chronic therapy (see tables I and II).

Recommendations

  1. Every HIV treatment centre should have a written adherence strategy in place which should be audited regularly in the context of the growing knowledge base.
  2. Interventions to improve adherence must be offered to all patients taking HAART.
  3. Adherence support interventions should be offered when starting therapy, changing therapy, and as part of ongoing routine follow-up.
  4. Providers must ensure that patients have sufficient understanding of HIV pathogenesis; the rationale for anti-HIV therapy; the relationship between adherence and resistance; the requirements of their regimen; and its potential side-effects (see appendix I). The support of voluntary sector resources should be enlisted, and verbal information should be supported by written information. Early follow-up, including a telephone assessment (performed by, for example, a community nurse 24-48 hours after medication has been dispensed) should evaluate whether patients have continuing information needs (though this has not been demonstrated to improve adherence to HAART).
  5. Medication alerts and containers should be provided as appropriate.
  6. Interventions offered by treatment centres should involve multi-disciplinary input.
  7. HIV professionals require continuing education in adherence issues, including professional development and skills-based training.
  8. Multiple interventions are likely to be necessary to maintain high adherence amongst clinic populations who are administered chronic therapy (see tables I and II).

4.3 Regimen factors

Lifestyle assessment: BHIVA recommend that, in the absence of definitive data regarding choice of therapy, regimens should be selected on an individual basis86. In addition to differing drug histories, patients will have varying needs and preferences in relation to toxicity, dosing, and avoidance of harmful drug interactions. A range of methods may be adopted to assist patients to consider the likely fit of any given regimen with their own lifestyle. The use of these methods prior to starting or changing treatment may help patients and health care workers foresee likely problems and devise coping strategies.

Dosing frequency: A recent systematic review indicates that medication which is dosed three or four times a day is associated with lower adherence levels than regimens dosed once or twice daily87. An open-label trial reported that virological responses to a twice daily HAART regimen (abacavir/Combivir) were at least comparable to one dosed three times daily (indinavir/Combivir) in naïve patients after 48 weeks therapy88. Adherence to therapy, measured by self-reported recall, was significantly higher in abacavir recipients (72% versus 45% reported taking more than 95% of their medication). Regimen factors which may be assumed to have influenced adherence in this study include twice daily versus three times daily dosing, differences in pill burden89, and differences in food and fluid requirements90. However, this is not an exhaustive list, and the relative weights which individuals may place on each of these factors is not reflected in a crude analysis of this kind.

Delayed dosing and pharmacokinetic issues: An analysis of adherence to PI-based HAART found no difference between twice daily and three times daily regimens with regard to the number of doses taken, but found that fewer doses were taken on time in the three times daily group91. Inter-patient and inter-regimen pharmacokinetic variability make the effects of late dosing difficult to predict, though data from a comparative study of different electronic MEMS (Medication Event Monitoring System) systems in people taking PI-containing HAART, found treatment failure was predicted by dosing late by one or by two hours in multivariate analysis92. While the longer half-lives of NNRTIs are often assumed to offer greater leeway over the timing of doses, in the absence of data regarding the effect of late dosing on treatment outcomes, it may be safer to advise all patients to space their antiretroviral doses consistently. Such regimen factors are also likely to be crucial to the outcome of treatment interruptions93.

Drug class: A longitudinal study involving over 1,100 naïve and experienced HAART recipients found higher levels of adherence in NNRTI recipients compared to PI recipients, after four and eight months of follow-up94. However, participants in these studies were taking a variety of regimens, and it is unclear therefore if this broad distinction is helpful. Moreover, a prospective, randomised comparison of antiretroviral-experienced individuals found no difference in adherence levels between those treated with an efavirenz-containing regimen versus a protease inhibitor-containing regimen95. The advent of improved formulations and pharmacokinetic enhancement means that regimens with low pill burden, and with no more than twice daily dosing, are now available in all classes.

Pill burden: Meta-analysis of eighteen randomised trials of nineteen HAART regimens, in treatment naïve individuals, found that lower pill burden was associated with the likelihood of having viral load below 50 copies after 48 weeks96. The extent to which pill burden may be a marker for other ‘adherence factors’ is unclear, however. The longitudinal study noted above found adherence was not associated with daily pill burden97.

Reliance on compact regimens: Though the term ‘treatment simplification’ has become commonplace in HIV therapy, whether any antiretroviral regimen prescribed for chronic use should truly be considered ‘simple’ is a moot point. An open-label trial found that 72% of abacavir/Combivir recipients reported adherence levels above 95%, and the remainder below this98. These data suggest that compact regimens (in this case, two pills, dosed twice daily, without regard to food) are not a panacea. Moreover, ‘treatment simplification’ is also regimen change and as such is a critical point for adherence review and support.

Drug interactions and side-effects: Additional adherence factors relating to specific regimens include the potential for harmful drug interactions (which should be expected to be particularly important in people with co-morbid conditions), and side-effects, which are frequently cited by patients as a common cause of missed doses99. The influence of side-effects may be both cognitive and practical, for example where doses are missed due to nausea, vomiting, diarrhoea, or where fatigue causes individuals to sleep through doses.

Recommendations

  1. Evaluate the extent to which lifestyle factors such as eating, sleeping and working patterns may impede adherence to a proposed regimen.
  2. Individual preferences regarding regimen characteristics, including formulations and pill size, should be assessed rather than assumed; what is simple for one person may not be for another100. Regimens which are dosed once or twice daily, which do not require fasting or dietary modification, and which have lower pill burdens, are likely to enable higher adherence and should be prescribed where appropriate, according to patient preference. Patients should always see the pills they may be prescribed prior to regimen selection.
  3. Undergoing a ‘dry run’, where antiretrovirals are substituted by placebos, may be useful in identifying possible adherence barriers101.
  4. Potential for drug interactions should be assessed prior to starting therapy. Ongoing assessment should be performed by physician and/or pharmacist whenever drugs are dispensed.
  5. Side-effects require prompt palliation. Patients can facilitate this process where they have a clear understanding of the likely toxicities associated with their regimen, appropriate self-management, and processes for accessing clinical support both in and out of hours.

Recommendations

  1. Evaluate the extent to which lifestyle factors such as eating, sleeping and working patterns may impede adherence to a proposed regimen.
  2. Individual preferences regarding regimen characteristics, including formulations and pill size, should be assessed rather than assumed; what is simple for one person may not be for another100. Regimens which are dosed once or twice daily, which do not require fasting or dietary modification, and which have lower pill burdens, are likely to enable higher adherence and should be prescribed where appropriate, according to patient preference. Patients should always see the pills they may be prescribed prior to regimen selection.
  3. Undergoing a ‘dry run’, where antiretrovirals are substituted by placebos, may be useful in identifying possible adherence barriers101.
  4. Potential for drug interactions should be assessed prior to starting therapy. Ongoing assessment should be performed by physician and/or pharmacist whenever drugs are dispensed.
  5. Side-effects require prompt palliation. Patients can facilitate this process where they have a clear understanding of the likely toxicities associated with their regimen, appropriate self-management, and processes for accessing clinical support both in and out of hours.