Yale University Program for the Study of Health Care Relationships - Pilot Study of Rapid Electronic Adherence Reporting with Provider Feedback on Adherence to Antiretroviral Therapy (HCR 01-12)

FUNDED PROJECTS:

The Program for the Study of Health Care Relationships funded 11 researchers in CT to study various aspects of health care relationships and adherence. The individuals listed below each received grants for one year (2001-02). They were awarded funds on the basis of a competitive review process and the scientific merit of their proposals. The wide range of disciplines they represent and their research interests demonstrated the diversity and complexity inherent in studying health care relationships.

Researcher's Summaries

Pilot Study of Rapid Electronic Adherence Reporting with Provider Feedback on Adherence to Antiretroviral Therapy (HCR 01-12)

Kevin Dieckhaus, MD, Carol Bova PhD, RN, ANP Kristopher Fennie, PhD, Ann Williams, EdD, APRN

The widespread use of antiretroviral medication (ARVs) has led to a dramatic decrease in AIDS-related mortality and morbidity. It has become apparent they must be taken with extreme diligence dosing instructions. Seemingly trivial deviations from strict adherence may result in sub-therapeutic plasma concentrations of drug with resultant loss of viral inhibition, ultimately leading to virologic breakthrough, development of drug resistance, immune deterioration, and clinical disease progression. If a medication is dosed improperly, the events resulting from virologic breakthrough may be irreversible and already in process before the problem can be identified at routine clinical visits.

The project used a novel technology to measure antiviral use with MEMS caps. Adherence data was transmitted daily to an internet site to be reviewed by an adherence nurse. The nurse provided both scheduled and problem-focused contact with the subject and the primary health care provider as based on alga rhythms developed by the investigator with guidance by the HIV-infected community.

The specific aims of the study were to 1. evaluate the feasibility of such an intervention, 2. measure changes in adherence to HIV medications, 3. assess the effect of the intervention on contextual factors that may be associated with improved medication taking behavior, 4. evaluate the effect of the intervention on clinical outcomes, and 5. assess if any improvements in adherence to a primary medication correlates to other medications that are being measured in a similar manner but are not the focus of such intense intervention.

Seventy-nine subjects were screened, 39 subjects met initial protocol-defined eligibility criteria. Fifteen had evaluable baseline, intervention, and washout phase data. Two of these underwent provider-directed treatment discontinuations during the active intervention, and are thus excluded from adherence analyses. Mean CD4 count at entry was 460 cells/ul; mean viral load was 1284 copies/ml with 60% having non-detectable serum viremia.

Assessing the feasibility of this exploratory intervention was an important aspect of the project. Feasibility was determined by general measured of adherence, rates of study completion, subjective commentary and frequency of technical difficulties. Subjects were defined as "responders" if adherence increased by 5%. By this measure, 53.8% of subjects responded to total doses taken, and 61.5% responded by improving dose-time (4-hour window) adherence. Subjects were then characterized as "responders" if the intervention phase adherence was >95% to total doses. Only 30.7% attained this goal. However, 100% of subjects who completed the intervention did have some measurable improvement to total doses and/or 4-hour dosing window. Additional feasibility data was obtained by eliciting opinions as to the usefulness of this intense intervention. 92.3% reported that they felt the intervention improved their adherence, and 61.6% felt that it improved their communication with their clinical provider. 58.3% reported agreed somewhat or very much that they would wish to continue the intervention outside of a study protocol. However, 23.1% thought that the intervention interfered with their lives in a significant way. Technical problems with the telecommunications devices were frequently encountered. 10% of subjects given HomeLink devices were unable to transmit data to the central server. 47% had some interruptions in reporting during the intervention phase. The mean duration of active internet-based adherence observation and feedback was 6.3 weeks of a planned 8-week intervention.

Changes in adherence were measured in two manners: 1. percentage of doses taken, and 2. percentage taken within a 4-hour "dosing time window". Adherence to the index ARV increased 9.6%, from 81.2% to 90.8% (p<.01) by total doses. Adherence increased 11.2%, from 57.7% to 68.9% (p=.06) by four-hour time-window. Improvements in adherence extinguished rapidly (to 82.7% and 62.4% respectively) after the active intervention concluded.

Changes in contextual factors that may affect adherence were assessed by the "trust in provider scale" with additional questions specific to trust in HIV. Trust in Provider scores were not changed pre-and post-intervention (49.8 to 48.5 on standard scale, 70.7 to 72.9 on HIV-focused scale).

No statistically significant effects in viral load, percentage of patients with undetectable viremia, CD4 counts, or results of the MOS-20 HIV Short form Quality of Life were noted between baseline and end of intervention.

Adherence to secondary medication generally correlated with adherence to the primary medication, with a statistical difference found in the baseline period (81% compared to 87% to total doses, p=.01). This is likely due to the protocol-defined selection of the medication with the poorest adherence as the primary medication. This difference loses statistical significance in the intervention and washout phase. There was no significant difference between changes in primary and changes in secondary medication adherence to doses taken or dose-time over the course of the intervention.

In conclusion, this pilot project demonstrated a significant improvement in adherence to ARVs by implementing an intense internet-based rapid adherence reporting system with rapid feedback. Adherence improved during the active intervention but was not sustained. This may indicate that such technology-based interventions may be most effective when maintained over a long term rather than a shorter burst of intensive follow-up. Limitations of the study included frequent technical difficulties, issues with integrating a technologically sophisticated device into an urban poor population, and resultant small sample size. Further research exploring this device and other more culturally accepted technologies (e.g. cell phones, pagers, etc.) should be directed to larger HIV+ populations.