Evaluating adherence to medication in children and adolescents with HIV
- 1Division of Infectious Disease, The George Washington University School of Medicine and Health Sciences, Washington, DC, USA
- 2Departments of Pediatrics, The George Washington University School of Medicine and Health Sciences, Washington, DC, USA
- 3Epidemiology and Global Health, The George Washington University School of Public Health and Health Services, Washington, DC, USA
- 4Epidemiology and Biostatistics, The George Washington University School of Public Health and Health Services, Washington, DC, USA
- 5Children’s National Medical Center, Washington, DC, USA
- 6Biostatistics and Informatics Unit, Children’s National Medical Center, Washington, DC, USA
- Correspondence to Dr N Rakhmanina, Department of Special Immunology, Children’s National Medical Center, 111 Michigan Avenue NW, Washington, DC 20010, USA;
- Accepted 17 August 2009
- Published Online First 31 August 2009
Objective: The study was aimed to evaluate the relationship between pharmacy supply, self-reported treatment adherence and HIV viral load in HIV-infected children.
Methods: A retrospective (52 weeks) cohort study was conducted through the review of the existing databases. Pharmacy supply was classified as “home delivery” when the medications were delivered home and as “in pharmacy pick-up” when they were picked up at the pharmacy. Adherence was assessed through retrospective (3 days recall) self-report. Fisher’s exact model, univariate and multivariate logistic regression analyses were used.
Settings: The study collected data on 140 HIV-infected children (<18 years). Adherence, pharmacy supply information and HIV viral loads were obtained from clinical and research databases.
Patients: The data from 127 HIV-infected children (60 boys and 67 girls; mean age 9.9 years) were collected.
Main Outcome Measures: Complete adherence (100%) was reported in only 24% of patients. With 40% of patients being rarely or never completely adherent, 64% of children achieved undetectable viral loads during the study period.
Results: No association between pharmacy supply and self-reported adherence was found (p = 0.605). Self-reported adherence (p = 0.0328) and age (p = 0.025) were the significant predictors of reaching undetectable viral loads. Adolescents (>13 years) were significantly less likely to reach undetectable viral loads than children under 13 years (odds ratio 0.38; 95% CI 0.16 to 0.89).
Conclusion: In our study, pharmacy supply was not associated with self-reported adherence. Most importantly, adherence and age were significant predictors of reaching undetectable viral loads.
Combination antiretroviral therapy (ART) has led to dramatic changes in the clinical course of HIV infection in paediatric patients.1 2 High ART adherence is required to achieve a complete and sustained suppression of viral replication.3 4 5 Poor adherence leads to treatment failure, the development of viral resistance with subsequent reduction in treatment options, increased morbidity and mortality.3 4 5 6 In adult patients, the mean ART adherence has been reported to be between 80% and 91%, and multiple adherence barriers have been identified.6 7 In paediatric HIV-infected patients, ART adherence has been reported to be lower (58%),8 9 due to the combination of patient or caregiver factors including age and maturity-related tolerability and parental involvement.10 11 12 13 Several population-specific ART adherence barriers have been established in paediatric HIV such as dependence on a caregiver, poor palatability (particularly of liquid preparations), large pill size and pill swallowing capacity relative to the child’s developmental stage.9 11
While some of the barriers related to the patients’ ability to tolerate and receive ART are difficult to change, others such as pharmacy supply can be modified through thoughtful evaluation and interventions by the providers, pharmacies and insurance companies. The examination of the relationship between pharmacy factors and treatment adherence in adults has shown the significance of pharmacy supply in patient adherence.14 15 Among pharmacy-related adherence barriers patients reported that either the pharmacy was out of the medication or the medications were not refilled in a timely manner.14 As children’s access to ART is dependent on their caregivers, the efficient mechanism of medications supply and positive experience of the caregiver with the pharmacy may lead to increased ART adherence and improved outcome in paediatric patients with HIV.16 17 To date, no studies have been published on the effects of pharmacy factors on treatment adherence and outcome in HIV-infected children and adolescents.
This study was conducted to characterise the methods of obtaining antiretroviral medications by caregivers and to evaluate the relationship between pharmacy supply, self-reported ART adherence and virological outcome in children and adolescents with HIV.
What is already known on this topic
Examination of the relationship between pharmacy-based factors and treatment adherence in adults has shown the significance of pharmacy supply in patients’ adherence.
The efficient mechanism of medications supply may lead to increased adherence to ART and improved outcome in paediatric patients with HIV. To date no studies have been published on the effects of pharmacy-related factors on treatment adherence and outcome in paediatric HIV infection.
What this study adds
Adolescents were less likely than younger children to reach undetectable viral load levels due to poorer adherence.
A direct relationship between self-reported adherence and virological outcome in children and adolescents with HIV has been demonstrated.
Patients and methods
This was a retrospective (52 weeks) cohort study of 140 HIV-infected paediatric patients (0–18 years old) treated at the Special Immunology Program at the Children’s National Medical Center, Washington DC. The Special Immunology Program provides care to perinatally HIV-infected children and adolescents from Washington DC metropolitan area including suburban areas, and is a site of multiple clinical studies including the National Institutes of Health and industry-sponsored research. Paediatric HIV-infected patients are seen in the clinic every 3 months as routine follow-up. Patients are given verbal and written instructions on antiretroviral medications at each clinic visit. The adherence assessment, pharmacy supply information and HIV-RNA viral load are routinely obtained as part of standard of care and clinical research studies. The majority of patients (95%) are black with approximately equal gender distribution. More than 90% of patients receive ART consisting of two nucleoside reverse transcriptase inhibitors and a single protease inhibitor (60%) or non-nucleoside reverse transciptase inhibitor (30%).
The records of self-reported adherence, HIV-RNA viral load, and pharmacy use were reviewed in a retrospective follow-up of 52 weeks of clinic and research records. Patients without ART, with incomplete pharmacy and adherence information in the clinical and research databases, were not included. Parents/legal guardians enrolled in the research studies have consented to personal health information storage in a database for future HIV research. The protocol and the Health Insurance Portability and Accountability Act/Institutional Review Board (IRB) authorisation for waiver of consent were reviewed and approved by the IRB at the Children’s National Medical Center.
Self-reported adherence was assessed through interactive 3 days recall questionnaire-based interview by the clinic staff with the caregiver and older children (>10 years of age) at routine clinic visit (every 3 months). During the interview caregivers and older children were asked about the medications names, dosing schedule, time of last dose, and the number of missed doses in the 3 days preceding the clinic visit (yesterday, 2 days ago and 3 days ago). They were also asked to identify adherence difficulties such as “forgetting”, “running out of medications”, “having difficulties taking the medications”, “missing medications due to pharmacy-related factors” and “others”. Three days recall was used since the available data suggest that patients or caregivers cannot accurately recall missed doses beyond a few dates.18 The history of pharmacy use was obtained from the clinical database, which contained information on the pharmacies used in the previous 12 months (including a history of problems or switching pharmacy), patient satisfaction with the pharmacy and the history of missing antiretroviral doses due to pharmacy errors or delays. Adherence was calculated as the number of all antiretroviral doses taken divided by the number of doses prescribed ×100(%) during the 72 h preceding the visit. The patients were graded into four categories based on the percentage of the 100% adherence reports (complete adherence) during the 52 weeks of follow-up: (1) always adherent (96–100% of visits completely adherent); (2) mostly adherent (50–95% of visits completely adherent); (3) rarely adherent (25–49% of visits completely adherent) and (4) never adherent (<25% of visits completely adherent). The pharmacy supply was classified as “home delivery” (HD) when the caregiver received antiretroviral medications at home (through mail or personal delivery) and as “in pharmacy pick-up” (IPP), when the caregiver picked up the medications at the pharmacy. HIV-RNA viral load (Roche Amplicor; Roche, Molecular Systems, Inc, Branchburg, New Jersey, USA) was considered to be undetectable with less than 400 copies/ml (the lowest limit of quantitation during the study period). The demographic characteristics including race, sex and age of the subjects were collected. All analyses were performed using SAS version 8.2.
Descriptive statistics, such as means and standard deviations were calculated for each continuous variable, and frequency distributions were generated for each categorical variable as appropriate. Frequency distribution characterised the pharmacy supply of obtaining antiretroviral medications (IPP vs HD). Fisher’s exact tests were used to estimate the association between pharmacy supply and self-reported adherence. Univariate and multivariate logistic regression analyses were used to estimate the association between pharmacy supply, self-reported adherence, age and virological outcome.
Data were collected on 127 paediatric patients with perinatally acquired HIV. There were 60 (47%) girls and 67 (53%) boys with a mean age of 9.9 years (SD 4.3) and the majority (n = 117; 93%) of patients was black. The mean number of visits per patient was 4.5 during the 52 weeks. Seventy-seven (61%) patients were children under 13 years, whereas 50 (39%) were adolescents (13–18 years old). Eighty-one patients (64%) have achieved an HIV-RNA viral load of less than 400 copies/ml during the study period.
A significant proportion of the children and adolescents reported suboptimal adherence during the study period (table 1). The majority of caregivers of adolescents (n = 29; 58%) used HD, whereas among the caregivers of younger children 36 (47%) used IPP. The caregivers of younger children were more likely to use the combination (IPP plus HD) supply than the caregivers of adolescents (8% vs 2%, respectively). Interestingly, pharmacy factors as a barrier to adherence were reported more often among HD (n = 8; 67%) than IPP (n = 4; 33%) patients. A history of missing antiretroviral doses due to the pharmacy was more frequently reported among the patients who used HD (n = 14; 70%) than those who used IPP (n = 6; 30%).
There was no association between pharmacy supply (IPP vs HD) and self-reported adherence (Fisher’s exact test p = 0.605). Examination of the multivariate logistic regression to determine the association between pharmacy supply, self-reported adherence and virological outcome revealed the overall model as significant (p<0.001). Moreover, adherence (p = 0.0328) and age (p = 0.025) were the only significant predictors of ever reaching an undetectable viral load during the study period (table 2). Adolescents (13–18 years old) were significantly less likely to reach an undetectable viral load than younger children (<13 years old) (odds ratio (OR) 0.38; 95% CI 0.16 to 0.89). For every year increase in age, the odds of reaching an undetectable viral load decreased by 10% after controlling for self-reported adherence and refill mechanism.
Those patients who chose a combination of IPP and HD were four times more likely to become undetectable (OR 3.85; 95% CI 0.38 to 38.58) than those who used HD; however, due to small numbers (n = 7), this association was not statistically significant. For every percentage increase in adherence, the odds of reaching an undetectable viral load increased by 3% after controlling for age and pharmacy supply.
Whereas the overall model of association between pharmacy supply, self-reported adherence and virological outcome in our study was shown to be significant, self-reported adherence and age were the only significant predictors of ever reaching an undetectable viral load during the 52 weeks of the study period. Previous studies have demonstrated the complexity of adherence assessment in children and adolescents with HIV.19 Direct measures of adherence, such as therapeutic drug monitoring, are expensive and difficult to interpret due to limited knowledge of the pharmacodynamics of paediatric ART. Indirect methods (self-reports, electronic drug monitoring, refill verification) have disadvantages ranging from the accuracy of the data to cost and practicality issues.19 Our data support the use of interactive self-report as an efficient tool in assessing adherence in paediatric patients, despite a well-recognised potential for overreporting.19 20 We must acknowledge that the study was conducted within a well-established paediatric HIV programme with vigorous adherence interventions, and adherence reports were collected through interactive interview with the caregivers and older children by familiar clinic personnel in a non-biased non-judgmental style that has the potential to decrease the overestimation of adherence. The study also used repeated adherence assessments, with an average number of reports of 4.5 per study period instead of a single report, which increased the reliability of self-reports.
In accordance with previously published studies, the higher level of adherence was significantly associated with the likelihood of reaching virological suppression when compared with a lower level of adherence. Rarely adherent (25–49%) and never adherent (<25%) patients were equally unlikely to achieve a viral load of less than 400 copies/ml.
The finding that adolescents were significantly less likely to reach an undetectable viral load than younger children in our study correlates with previously published data on the lower rates of ART adherence among HIV-infected youth.3 10 Although we acknowledge that the adolescents in the study had perinatally acquired HIV, were highly treatment experienced, and therefore had higher chances of HIV resistance and virological failure, we must recognise many additional obstacles to ART adherence emerging during puberty. Transition to adolescence by children with perinatally acquired infection leads to changes in lifestyle involving growing independence, separation from parental involvement, increased peer pressure and fear of stigmatisation, increased risk-taking behaviour, psychiatric problems and substance abuse.6 19 20 21 22 For providers and caregivers loss of adherence during puberty in adolescents with perinatally acquired HIV represents a difficult and emotional challenge that requires a team approach and close collaboration. While several strategies (directly observed therapy, regimen-related, education and counselling interventions) have been suggested to maximise ART adherence during this transition period,6 19 20 21 22 clinicians frequently seek guidance from research and practice in other paediatric chronic illnesses such as asthma, juvenile rheumatoid arthritis and type I diabetes mellitus. Given the lack of a well-defined adherence intervention model in adolescents with HIV, more research on adherence among HIV-infected youth with interdisciplinary collaboration is warranted.
We recognise limitations of our study such as the retrospective study design, small sample size and lack of direct patient/caregiver interviews. In addition, all of the pharmacy information gathered from self-reports was not confirmed by the verification of refill histories. Other factors, such as drug resistance and treatment history, could account for the differences in reaching an undetectable viral load. In summary, self-reported adherence and age were the only significant predictors of ever reaching an undetectable HIV viral load during the 52 weeks of follow-up.
The authors express sincere gratitude to Drs Veronica Miller, PhD, Research Professor, Department of Prevention and Community Health, The George Washington University (GWU) School of Public Health and Health Services (SPHHS) and Ann Goldman, MA, MPH, Research Instructor, Department of Epidemiology and Biostatistics, GWU SPHHS for their guidance and critical review of the study. The data analysis was completed with the participation of Dante Verme, PhD, MS, Professor of Epidemiology and Biostatistics, and Gregory Phillips II, MS, PhD, Department of Epidemiology and Biostatistics, GWU SPHHS.
Funding The authors were supported by Department of Health and Human Services, National Institutes of Health Public Health Service grants NCRR 1K12 RR017613 and NICHD 1U10 HD45993 (NR).
Competing interests None.
Ethics approval The protocol and the Health Insurance Portability and Accountability Act/Institutional Review Board (IRB) authorisation for waiver of consent were reviewed and approved by the IRB at the Children’s National Medical Center.
Provenance and peer review Not commissioned; externally peer reviewed.