To our knowledge, this would be the first study to apply this algorithm-based memory profile classification of individuals to understand an important everyday functioning outcome . Further, no study thus far has evaluated the HIV+ memory deficit profiles associated with non-adherence while using different types of verbal stimuli . The format in which information is presented may have a direct impact on memory performance, particularly in the context of medication management. For example, multiple studies have reported improved medication adherence when traditional prescriptions were enhanced with more descriptive instructions about the medication regimen instead of simply listing the information . The current study seeks to expand upon previous research to determine the associations between different episodic memory profiles and antiretroviral non-adherence by utilizing traditional CVLT scoring methods with the inclusion of retention and recognition indices and a concurrent measure of passage recall . In addition to standard group-level analyses of these clinical memory tasks, we also endeavored to look at individual profiles of word list learning as a risk factor for non-adherence using a previously established algorithm for classifying normal, encoding and retrieval profiles. Based on the literature reviewed above, we expect that antiretroviral non-adherence will be associated with a mixed encoding/retrieval profile. These associations may be particularly pronounced on word lists versus story memory given the former’s greater reliance on executive processes , such as strategic encoding, which is reliably impaired in HIV+ individuals and has been strongly associated with non-adherence .The study sample was comprised of 202 HIV+ participants recruited from the general San Diego area . The study was approved by the University of California, San Diego’s Human Research Protections Program. All participants provided written informed consent prior to study participation. Participants were included in the analyses if they maintained at least one current prescription for an antiretroviral medication and completed an approximately month-long observation of adherence using Medication Event Monitoring System.
Based on a parent project ,vertical farming pros and cons exclusion from the study occurred if a participant had a severe psychiatric disorder , major neurological disease , estimated verbal IQ < 70 determined using the Wechsler Test of Adult Reading , diagnosis of a current substance use disorder at baseline, positive urine toxicology screen for illicit substances at baseline. Adherence was tracked using MEMS, which utilizes a microchip-containing bottle cap that records the time, date, and frequency with which the bottle is opened. All participants were provided the MEMS bottle the day of neuropsychological testing and continued use for approximately 5 weeks . Participants were asked to use only the bottle provided to dispense their medication, to open the bottle only for the purpose of dispensing their medication, and to dispense only one dose of medication at a time. The ARV selected for the MEMS bottle was based on whether the participant would continue the medication through the course of the study, the frequency with which the medication is prescribed in the general population , and its pharmacokinetic properties. Blind reviewers used the MEMS data to categorize participants as either adherent or non-adherent. Participants were determined adherent if they followed their target antiretroviral prescription regimen ≥ 90% of the time . Table 1 displays descriptive data for non-cognitive variables that differed between the Adherent and Non-Adherent groups. There were no significant between group differences in demographics , HIV disease characteristics , or medication regimen complexity . The Adherence groups were also comparable on most psychiatric characteristics, including current Major Depressive Disorder , current and lifetime Generalized Anxiety Disorder , current mood as measured by the Profile of Mood States , lifetime substance dependence, and positive urine toxicology screen for cannabis on the day of testing . Because a positive marijuana urine toxicology screen was allowed on the day of testing, between-group differences were evaluated to determine the potential confound of acute cannabis effects on adherence. No significant differences were found between adherent and non-adherent participants or between the normal , encoding deficit , and retrieval deficit algorithm-derived clinical classification profiles.
There was a significantly higher prevalence of lifetime MDD in the Non-Adherent group; however, this variable was not associated with any of the primary memory variables of interest in the entire study cohort , with the sole exception of CVLT-II Total Trials 1-5 . Follow-up logistic regressions showed that LT MDD did not affect the magnitude of the association between CVLT Total Trials 1-5 and adherence . Moreover, CVLT-II Total Trials 1-5 was not associated with current depression or the depression/dejection sub-scale of the Profile of Mood States . As such, we did not include LT MDD as a covariate in any of our subsequent analyses. Medication non-adherence is common in HIV disease, particularly among individuals with memory deficits .Consistent with earlier studies on this topic, our data showed that episodic verbal learning and memory have broadly medium associations with adherence in HIV+ individuals, with CVLT-II surprisingly showing smaller average effects than WMS-III LM . The profile of memory findings in the non-adherent cohort was characterized by deficits in initial acquisition and delayed free recall, but not in retention or recognition. This pattern of findings is consistent with what has been described as a “retrieval” deficit, which is commonly found in persons with HAND and other conditions that affect prefronto-striato-thalamo-cortical circuits . Of note, algorithm-derived clinical classifications of the CVLT-II showed that HIV+ non-adherers have a higher prevalence of sub-cortical retrieval deficit profiles and that those with such retrieval profiles were three times more likely to be non-adherent than those with an encoding deficit profile. This is interpretively consistent with the work of Wright and colleagues , who found that non-adherent HIV+ persons differed from healthy adults on the ISDA index of retrieval, but extends that prior paper by: using traditional CVLT-II indices that include retention and recognition; incorporating a measure of passage recall, providing classifications of individual profiles using the algorithm-derived methods of Massman and colleagues along with traditional CVLT scoring methods and demonstrating associations between the algorithm profile approach and ecologically relevant measures of everyday functioning. The significant associations between adherence and profile scores in the context of insignificant standard deficit scores underscores the clinical relevance of understanding individual profiles of memory deficits when evaluating everyday functioning outcomes.
Future research comparing this algorithm across HIV and other well-established cortical and sub-cortical deficits using both neuropsychological and ecologically valid real world tests appears warranted to better understand the true impact of these deficit profiles across disorders. From a memory process perspective, these findings indicate that missing a medication dose is not simply a “forgetting” problem , as there was no evidence of an association between adherence and memory savings scores or recognition performance. Instead, non-adherence in HIV is at least partly associated with deficits in the complex process of efficiently accessing information from episodic verbal memory with minimal retrieval cues. Such retrieval deficits may be a function of weak encoding and/or difficulties with self-initiated memory search and access to stored information. Therefore,air racking even if an HIV+ individual understands and encodes the relevant information regarding their cART regimen, he/she may still not remember to take his/her medication as prescribed at least in part due to difficulties retrieving the proper information at the later dosing time. These memory processes did not appear to change depending on the complexity of one’s regimen, instead, difficulty retrieving information was associated with non-adherence regardless of the number of pills an individual was prescribed. Importantly, these complex memory process failures may also occur in concert with failures in other aspects of cognition , psychiatric comorbidities, and various psychosocial factors , which may further complicate and exacerbate the observed retrieval deficits. Indeed, there are numerous neurobehavioral aspects involved in managing one’s medication regimen . The absence of an association between executive functions and adherence may at first appear contrary to the known pattern of deficits seen in individuals with fronto-striatal dysregulation, such as in HIV disease. Indeed, prior studies have shown that executive functions are related to cART non-adherence . However, the free recall memory variables that differentiated the adherence groups were associated with executive functions in the non-adherence group, which supports the hypothesis that a deficiency in the higher-order, organizational aspects of memory processes may play an important role in medication non-adherence. Specifically, Moscovitch described a component process theory that included a higher-order control process by which the frontal systems work “with” encoding and retrieval to properly organize information for efficient storage and recall. Poor strategic encoding and retrieval has been observed in HIVseropositive individuals , and therefore the profile of deficits and the observed associations with executive dysfunction findings suggest that these strategically driven aspects of memory processes are important for successful medication adherence in HIV.
The stronger association between non-adherence and story recall versus list learning was surprising, as we had expected the purported executive demands of list learning to be more important in the context of managing medications . However, this unexpected finding may reflect the relevance of contextual information to understanding memory’s effects on important aspects of everyday functioning, such as medication adherence. In other words, the stronger associations between story memory and adherence underscores the importance of an individual’s ability to understand and independently recall medication information presented in a contextual format for proper regimen adherence. There is a long history of episodic memory research on the role of context in facilitating recall and the match between the context of encoding and retrieval . In this instance, therefore, story recall may better match the everyday demands of memory on adherence than does list learning. Non-adherers may have greater difficulty remembering and applying regimen information presented conversationally by their healthcare providers and in the context of related information as opposed to being presented with a listed dosing schedule. This provides one explanation for why the LM sub-test has historically provided a more sensitive test for subtle memory deficits compared to the CVLT . Morgan and colleagues , to this end, demonstrated that individuals with HAND evidenced difficulty with contextually-based health information. Specifically, the HAND group performed more poorly on an applied task in which health-related information was provided in the full context of a detailed scenario involving a health choice . This finding is consistent with the results of the present study, suggesting that medication regimen instructions presented in the larger context of health-related information may interfere with, rather than facilitate, memory for that information and subsequently decrease adherence. These findings suggest that when complex, contextualized health information is provided to patients it should be supplemented with more focused, targeted instructions that facilitate adherence to the prescribed health plan. Several limitations of this study are important to highlight. This study focused exclusively on verbal episodic memory. Due to a heavy reliance on reading prescription labels to obtain medication regimen information as well as visual cues when dispensing and dosing medications , understanding memory profiles associated with visual stimuli could also be of value. Because MEMS caps were utilized for adherence observations, the effects of commonly used compensatory behaviors, such as using pillbox dispensers , on adherence were not evaluated, therefore reducing some generalizability to everyday life. Further, future studies could improve our understanding of these profiles on medication non-adherence by evaluating the efficacy of commonly used compensatory strategies among individuals with specific profiles of memory deficits. This study also examined memory associated with common words and simple stories. However, drug labels often include complex medical and numeric information , which may influence adherence. Multiple studies found that retention of medical information has been as low as 25% . Therefore, it would be more ecologically relevant to examine memory in the context of medical information and complex descriptions rather than the commonly understood information presented on the clinical memory tasks used in this study. Despite these limitations, findings from this study may have implications for the development and implementation of compensatory strategies to improve cART adherence. Given that non-adherers and adherers exhibited comparable recognition abilities, compensatory adherence strategies associated with simple recognition, such as placing a pillbox by a commonly used area, may help to improve adherence in individuals with retrieval difficulties.