Research has generally failed to examine the specific ways in which families manage children's ART regimens and how those family factors relate statistically to adherence. Rather, studies have focused on individual characteristics of a child or caregiver that may indirectly impact adherence, such as depression or substance use,^4,5 or interpersonal factors, such as a caregiver's perceived social support.^6,7 While such factors may be important determinants of adherence, they provide little information about families' roles and processes that may affect adherence. Understanding families' experiences—especially who is responsible for regimen-related tasks, what keeps them from accomplishing those tasks, and what helps them succeed—may be critical for developing effective adherence-promoting interventions. For example, it may be important to know whether adult caregivers or children are responsible for tasks related to adherence (e.g., refilling prescriptions, remembering to take the medicine, etc.) so that we can both understand how differences in regimen responsibility affect adherence and know to whom interventions should be directed. Research among adolescents with a variety of chronic illnesses suggests that children assume increasing levels of responsibility for disease management as they mature,^8–11 and at least one study suggests that caregivers may assume less responsibility as children become older.¹² At the same time, adherence problems increase as children get older,^13–15 especially during adolescence.^15–18 Thus, children and adolescents may be given responsibility for regimen-related tasks when they may be unlikely to complete them successfully. Studies have reported on some aspects of responsibility for medication-related behaviors within families of children on ART.^11,19,20 In a study by Boni and colleagues,¹⁹ 86% of 25 caregivers of children on ART reported that they “watched children while they took their therapy.” No additional information about regimen responsibility was provided.¹⁹ Wrubel and colleagues,²⁰ in a qualitative interview study, reported on experiences of 71 mothers of children (ages 1–18 years) on ART; while 89% of mothers reported taking responsibility for their children's adherence on at least one occasion, 21% indicated that they shared the responsibility with their children at least part of the time, and nearly 25% of mothers indicated a desire for their children to be responsible for ART adherence. Only 2 mothers recognized that their adolescents may not be able to take their medications consistently without maternal reminders.

Only one known study has directly assessed the relationship between regimen responsibility and ART adherence. Martin and colleagues¹¹ studied ART adherence among 24 children ages 8 to 18 years and used both child and caregiver report to assess the extent to which the caregiver, child, or both assumed responsibility for regimen-related tasks. From their report, it was not clear what percentage of children and caregivers were solely responsible for regimen-related tasks. Neither child nor caregiver reports of regimen responsibility were directly associated with adherence, although greater caregiver–child discrepancies about regimen responsibility were associated with worse adherence at the second time-point. Also, children's regimen responsibility increased significantly with child age.

Multiple regimen-related tasks must be completed in order to achieve adherence, including: (1) remembering that it is time to take the medications; (2) opening the bottles and retrieving the medication; (3) swallowing the medication; (4) noticing when refills are needed; and (5) calling to arrange for refills. It is possible that responsibility for some of these tasks could be allocated to children or youth without detrimental effects on adherence. For example, it may be that adherence could be achieved easily if a child is responsible for opening the bottles and retrieving the medication, as long as an adult ensures that the medications are swallowed. The importance of children versus caregivers assuming responsibility for these specific tasks has not been well studied.

As suggested by the Health Belief Model,^21,22 identifying barriers to adherence may also help to inform intervention efforts, especially to the extent that barriers are associated with adherence outcomes. Studies have examined families' barriers to pediatric ART adherence regimens,^{19,20,23–34} but only three small studies have examined the statistical association between reported barriers and adherence, and those have produced conflicting results. When a checklist was used to assess barriers to adherence, more barriers were associated with lower levels of caregiver-reported adherence^29,30; yet when 10 barriers were rated on a Likert-type scale, the mean barrier frequency score was not associated with adherence measured by caregiver report, pill counts, or electronic monitoring.³³ Studies with larger sample sizes are needed to clarify the relationship between barrier reporting and adherence, as well as to identify which specific barriers statistically account for adherence difficulties.

Once we identify the critical barriers, it is then important to identify strategies for overcoming those barriers. Understanding which strategies help families successfully manage ART regimens is important for helping other families improve adherence behavior. Some such strategies have been enumerated, including: using reminder devices (e.g., calendars, timers, beepers), using physical reminders such as putting bottles in an obvious location or posting reminder notes, asking other family members to help with reminding, incorporating medication-taking into daily routines, and improving or masking unpleasant tastes.^24,35,36

Studies have generally not supported the effectiveness of several commonly recommended strategies for improving adherence among adults. A recent meta-analytic review showed that three interventions using electronic reminder devices (i.e., beepers or pagers) failed to increase the likelihood of participants obtaining >95% adherence relative to controls.³⁷ One of those studies found more one-log decreases in plasma HIV RNA loads among intervention versus control participants, suggesting that electronic reminders may have some adherence-enhancing effects.³⁸ Another controlled study found improvements with the use of an online pager system, but adherence was still low in the intervention group (64%).³⁹ Other interventions have included teaching patients to use cues as reminder of medication times. One controlled study found that using electronic medication monitoring feedback and having participants identify cues that would help them remember to take their ART medications was not effective in improving adherence or plasma HIV RNA loads unless it was paired with a monetary reward.⁴⁰ Another study tested a single-session and follow-up phone call “Life-Steps” intervention that included assistance identifying cues for medication-taking and guided imagery review of medication-taking in response to identified cues. Compared to participants who used daily dairies to monitor their adherence, Life Steps participants showed better adherence outcomes after two weeks but no difference at 12-week follow-up.⁷⁵

Few studies have examined the use of memory strategies for increasing adherence among children on ART. A recent cross-sectional study of children and adolescents found increased odds of adherence among those who used a buddy system to help them remember to take medications, although they found no greater likelihood of adherence among those who tied medication-taking into other activities or used pill boxes to remember doses versus those who did not.¹⁵ An intervention pilot study with youth ages 15–22 provided participants with two reminder devices: a beeper that vibrates and can hold pills, and an alarm watch; provision of these devices did not significantly improve adherence or reduce the problem of youth forgetting doses, although youth believed that the multiple alarm watch was helpful.⁴¹ These studies provide little support for using cue-based strategies to improve adherence among children and adolescents with HIV. Nevertheless, the literature in this area is too sparse to be conclusive. While the adult literature suggests that memory strategies are of limited effectiveness for improving adherence to ART, it is important to examine the value of cue-based strategies for facilitating adherence among children, as pediatric clinicians continue to recommend these strategies with very limited pediatric data regarding how the use of such strategies impacts adherence.

Confirming our hypotheses as well as a previous finding in the pediatric ART literature,¹¹ children who shared responsibility with their caregivers were older than children who did not share responsibility for each of five regimen tasks. Caregivers often allow children to assume increasing responsibility for medication adherence as they become older and show signs of maturity,^8–11 which for some may be an appropriate means of preparing children for self-care during adulthood. Children and adolescents may appreciate the opportunity to control whether or not and when they get their medications, leaving them with less anger and greater regimen-specific self-efficacy,⁵² while caregivers may be relieved to relinquish medication responsibilities.⁵³ Unfortunately, across chronic conditions, data consistently show that for many, adolescence—the time when caregivers are most likely to grant their children greater responsibility—is a time of poor adherence to medical regimens.^15–18 Thus, it may be important for clinicians to offer caregivers anticipatory guidance that stresses the importance of supervising and monitoring their child's completion of regimen-related tasks,^11,54 even when they allocate primary responsibility for medication-taking to their child or adolescent.

Older age was not significantly associated with nonadherence in this study. Other studies in the pediatric HIV literature have failed to find a direct relationship between age and adherence^{30,32,33,55,56} or have found that caregiver reports of adherence increase with child age.²⁶ Like the present study, most such studies have included primarily children under age 13 years^30,32,55,56 at least partially excluding adolescents, who may have the greatest difficulty adhering. Therefore, the failure to find an age effect may be due to truncation of the sample. Moreover, the relationship between age and adherence is likely mediated by other factors, including regimen responsibility, in which case identifying and changing those mediators may be the most pertinent objectives.

Consistent with findings of two previous studies, caregivers who reported fewer barriers to adherence were more likely to have adherent children.^29,30 Several individual barriers were significantly associated with nonadherence, including three barriers that may reflect disorganization (forgot, busy with other things, and change in daily routine), and one barrier associated with parent–child dynamics (child refused). As they are the most commonly reported barriers in this sample, determining whether or not families experience these barriers may be particularly important for identifying adherence problems and creating a dialogue with families about their medication-related experiences.

Once adherence problems are identified, providers may recommend commonly used strategies for remembering medication. Our findings suggest that this approach may not always be effective. Neither using more memory strategies overall nor using at least one (versus no) cue-based strategy was significantly associated with adherence. Although strategies such as tying medication-taking into daily routines and using reminder devices may assist some families with adherence, using such strategies is insufficient for some families to achieve adherence.^15,40,41,57 In fact, our results show that caregivers who used more strategies (versus less) for remembering medications were more likely to report a problem with forgetting doses. Similarly, Kalichman and colleagues⁵⁸ found that some adults on ART who used pill boxes also used other memory strategies and were more likely than non-pillbox users to report missing medications due to forgetting. It seems unlikely that using memory strategies contributes to problems of forgetting; rather, the problem of forgetting may trigger the use of multiple strategies to address adherence problems. Families who miss doses due to forgetting may try multiple strategies for remembering medication dose times.

Families who struggle with forgetting, multiple demands, or changes in their routine may have difficulties with organization in general. A beeper or other reminder device is only helpful if, upon receiving the reminder, the child takes the medication. The reminder device will be ineffective if the reminder is given but the child does not get the medicine. For example, a child may not get the medicine after a reminder due to a supply problem (e.g., they ran out of medication) or a proximity problem (e.g., they did not take the reminder device or the medication along when they left the house). Families who experience such situations may benefit from assistance in organizing their lives so that they have a dose of medication available at all times. In the case of proximity problems, families may also benefit from learning how to set their reminder devices to alert them later, when they are likely to be at home.

Stigma—anticipated or experienced— is a potential barrier to the success of memory strategies for promoting adherence. It is uncommon for families to disclose a child's HIV status to most people outside of the immediate family; thus, even when families have effective strategies for remembering doses, children may miss doses due to concerns that others will inadvertently learn about the child's HIV status.^28,53 Yet stigma does not explain the lack of relationship between memory strategies and adherence in this study, given that only 2% of caregivers reported that nonprivacy was a barrier.

Untested in this study is the hypothesis that motivational barriers may keep some children from receiving the medication in response to a reminder. The child or caregiver might turn off a medication alarm with some vague intention of taking or administering the medication in the ensuing minutes or hours, and then “forget” to take or give the medication.²⁸ “Forgetting” may be a proxy for psychological processes such as avoidance and denial.⁵⁹ Depression^5,60 and posttraumatic stress related to the HIV diagnosis,⁶¹ other mental health problems,⁵⁶ and limited adherence-specific or global social support⁵⁶ may fuel this avoidance or otherwise make dose administration difficult if not impossible. Surely most caregivers want to their child to get this critical medication, although some caregivers may avoid dosing either: (1) because the medication reminds them of HIV—a life-threatening, stigmatized chronic illness that burdens their family, and in some cases, which they passed to their child⁵³; (2) out of compassion for their child's reluctance to take the medicine⁵³ or (3) to avoid stressful caregiver–child conflicts around medication-taking.^28,53 Studies have examined attitudes and beliefs related to ART medication^62–64 (see Fisher et al.⁶⁵ for a brief review), but these psychological barriers to adherence have not been well explored.

Research is needed to determine families' goals or behavioral intentions for medication-taking and how they relate to adherence.⁶⁶ Although health care providers often have near-perfect adherence as a goal, some families may be content with the child getting their medication most of the time as long as their child appears healthy. Bauman⁶⁷ distinguishes between types of nonadherers: volitional and inadvertent. Volitional non-adherers consciously choose to not follow a regimen. Inadvertent nonadherers make a decision to follow a medical regimen but fail to do so; they may fall into one of three subtypes: (1) those who are nonadherent because they misunderstand the regimen; (2) those who want to adhere but face barriers; and (3) those who miss doses at times, but feel they are adhering sufficiently. In a study of young women with HIV, Kalichman and colleagues⁶⁸ found that intentions to adhere were significantly associated with adherence; similar work is needed to better understand the role of intentions in adherence to pediatric regimens. Such research should attempt to reduce social desirability bias, because women may be concerned about the consequences of reporting socially undesirable intentions for adhering to their child's regimen.⁶⁹

This study has multiple strengths, including a large sample relative to most studies in the literature.⁷⁰ Data were drawn from seven pediatric HIV clinics in four U.S. cities; thus, these findings are likely generalizable to a large portion of children living with HIV in the United States. Moreover, this study is unique in its examination of three family experience factors as they relate to adherence. Several limitations must be noted as well. First, conclusions about causality are limited due to the cross-sectional design. Additionally, this study used a single caregiver self-report question to assess adherence over a 6-month period. Although the association between this measure and virologic outcomes supports its validity, replication of this finding is important, and findings will be enhanced when multiple measures are used.⁷¹ For example, incorporating child as well as caregiver reports may provide a fuller, if more complex, picture of adherence. More objective assessment tools such as electronic monitoring devices, while expensive and imperfect, can provide useful triangulation of data, as well.²⁵

Both the present study and a previous study¹¹ adapted a measure originally used to assess responsibility for diabetes regimen tasks,⁵⁰ which asks caregivers whether the child, caregiver, or both are primarily responsible for regimen tasks. The current study found limited variability in responses to most items, suggesting that a different response set may improve the measure. A similar measure to assess responsibility for asthma management has the respondent rate each task on a 5-point Likert-type scale ranging from 1 (parent is completely responsible) to 5 (child is completely responsible),¹² although only limited support was found for the relationship between that measure and children's adherence.^12,72 Another option would have respondents report how frequently over a set period of time a caregiver (versus child) actually completed each task. For some regimen tasks it may be valuable to reframe the construct from responsibility to supervision and monitoring.⁵⁴ Presumably, the child is typically the one who puts the medication in his/her mouth, so the important question may be whether or not a caregiver was present and watched it occur.⁷³ With either modification, efforts to limit response biases due to social desirability will be critical to the future success of this measure.

PACTS and PACTS-HOPE were funded by the U.S. Centers for Disease Control and Prevention through cooperative agreements U64/CCU207228 (MHRA of New York City), U64/CCU202219 (UMDNJ-New Jersey Medical School), U64/CCU306825 (University of Maryland School of Medicine), and U64/CCU404456 (Emory University School of Medicine). The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.

During the production of this manuscript, Dr. Marhefka was supported by the Center Grant P30 MH43520 from the National Institute of Mental Health to the HIV Center for Clinical and Behavioral Studies, Anke A. Ehrhardt, Ph.D., Principal Investigator, and NRSA T32 MH19139, Behavioral Sciences Research in HIV Infection, Anke A. Ehrhardt, Ph.D., Program Director.