Association of Self-Reported Medication Adherence with Potentially Inappropriate Medications in Elderly Patients: A Cross-Sectional Pilot Study

Background: Polypharmacy (PP) and potentially inappropriate medications (PIMs) cause problematic drug-related issues in elderly patients; however, little is known about the association between medication adherence and PP and PIMs. This study evaluated the association of self-reported medication adherence with PP and PIMs in elderly patients. Methods: A cross-sectional pilot study was conducted using data collected from electronic medical records of 142 self-administering patients aged ≥65 years, excluding emergency hospitalization cases. Self-reported medication adherence was assessed using the visual analogue scale (VAS). Results: Of the 142 patients, 91 (64.1%) had PP and 80 (56.3%) used at least one PIM. In univariate analysis, patients with a VAS score of 100% had a significantly higher number of female patients and ≥1 PIM use compared to other patients. We found no association between the VAS score and PP. In multivariable analysis, the use of PIMs was significantly associated with a VAS score of 100% (odds ratio = 2.32; 95% confidence interval = 1.16–4.72; p = 0.017). Conclusions: Use of PIMs by elderly patients is significantly associated with self-reported medication adherence. Pharmacists should pay more attention to prescribed medications of self-administering elderly patients in order to improve their prescribing quality.


Introduction
In Japan, the demand for drug therapy for elderly patients (age ≥ 65 years) is increasing with an increase in the number of the elderly [1]. However, the pharmacokinetics and drug responsiveness in elderly patients differ from those in younger adult patients because of aging-related physiological changes, and interactions among drugs administered to treat multiple comorbidities might cause adverse drug events (ADEs) [2,3]. Polypharmacy (PP), the use of five or more medications, is a major problem in terms of increasing risk of ADEs in elderly patients [4][5][6][7][8][9][10][11]. Onoue et al. (2018), in a nationwide retrospective study using 240 million pharmacy claim data items, reported that

Study Setting and Patients
We conducted a cross-sectional pilot study in the Fukuoka University Chikushi Hospital (Chikushino, Japan). The study subjects were patients admitted to the hospital between 1 September 2019, and 29 February 2020. Pharmacists interviewed patients soon after they were admitted to the hospital on weekdays. The questions involved prescribed drugs, over-the-counter (OTC) drug use, history of ADEs, and medication adherence (i.e., VAS). The pharmacists interviewed~20 patients/day for 5-10 min per patient and recorded the interviews in electronic medical records (EMRs). Most pharmacists were irregularly engaged in this work, but one pharmacist (Ph. X) did this every day.
After the patient completed the hospitalization procedure, the administrative department requested the pharmacist to interview the patient, and the patient was randomly selected. To minimize selection bias (i.e., to ensure the consistency of the method of recording the interview with the patient), we targeted only patients interviewed by Ph. X. Patients < 65 years old, admitted for an emergency, unable to self-medicate, or not taking oral medications were excluded. Finally, we included 142 elderly patients in this study.
The study was approved by the Fukuoka University Medical Ethics Review Board (C20-04-002).

Patient Characteristics and Medication Information
Patients' data were collected on admission. Patient characteristics and medication information were obtained from their EMRs. Patient characteristics included sex, age, height, body weight, body mass index, and history of ADEs. Medication information included OTC drug use, possession of prescription records, one-dose package (ODP) dispensing, the number, type, and duration of prescribed oral medications, and the number of consulting medical institutions. VAS data were obtained from the pharmacist's records in the EMRs. At the time of the interview, a pharmacist asked each patient, "What percentage of your medications did you take exactly as your doctor prescribed them?" to evaluate the adherence for all medications. The VAS tool was presented to each patient with a continuous line ranging from 0% to 100%, and he or she was asked to mark the line at his or her best guess about self-medication adherence. A VAS score of 100% was defined as full medication adherence, as described previously [52,53].
As recommended by the JGS, the patients were divided into the following two groups: pre-old (65-74 years) and old (≥75 years) [30].
The list of target drugs, target patient populations, and recommendations developed by the JGS, available on the JGS web page for the STOPP-J in Japanese [55], was used for PIM screening [31]. The list used in this study was shown in Table S1. We checked the administration period of the prescribed oral medications by using patients' prescription records or patient referral documents scanned and saved in EMRs; however, we could not obtain accurate data on the long-term combined use of multiple antithrombotic agents. Thus, the recommended administration period of the combined use of multiple antithrombotic agents was excluded from screening. We used the Kyoto Encyclopedia of Genes and Genomes drug database (https://www.kegg.jp/kegg/drug/) to search and classify the generic names of medications each patient was taking.

Association of PP with Patient Characteristics and Medication Information
In univariate analysis comparing patient characteristics and medication information for PP(+) and PP(−) patients (Table 3), we found a statistically significant difference in the age group (p = 0.003), ODP dispensing (p = 0.003), number of consulting medical institutions (p = 0.002), and PIM use (p = 0.001).

Association of PIMs with Patient Characteristics and Medication Information
In univariate analysis comparing PIM(+) and PIM(−) patients (Table 4), we found a statistically significant difference in sex (p = 0.037), ODP dispensing (p = 0.020), number of prescribed oral medications (p < 0.001), and number of consulting medical institutions (p = 0.045).   Table 5 shows the results of univariate and multivariable analyses for a VAS score of 100%. In univariate analysis, patients with a VAS score of 100% had a significantly higher number of female patients and ≥1 PIM use compared to patients with a VAS score < 100% (p = 0.060 and 0.008, respectively). Multivariable analysis showed that PIM use (odds ratio [OR] = 2.58; 95% confidence interval [CI] = 1.20-5.72; p = 0.015) was significantly associated with a VAS score of 100%-that is, full medication adherence.

Discussion
Self-reported medication adherence is a practical method of measuring a patient's medication adherence because it is quick and cheap and has the potential to be easily implemented into the clinical workflow. In our hospital, since 1 September 2019, pharmacists have measured the VAS score of hospitalized patients at admission as routine work and, as a result, have provided better medication counseling to these patients than before. The VAS is used to assess medication adherence and shows high median or mean scores in a variety of populations: patients taking antidiabetes (median, 95.9%) and lipid-modifying (median, 95.2%) drugs [48]; hypertension/type 2 diabetes mellitus/dyslipidemia patients (mean, 91.3%) [51]; patients taking at least one hypertensive medication (median, 100%) [56]; IBD, including ulcerative colitis or Crohn's disease, patients (median, 91-100%) [50,54,57,58]; rheumatoid arthritis patients taking methotrexate (median, 94%) [59]; patients taking warfarin (mean, 92.2-96.6%) [49,60]; patients admitted to the psychiatric ward (mean, 86%) [61]; glaucoma patients (median, 95.0%) [62]; postmenopausal women with hormone receptor-positive breast cancer taking aromatase inhibitors (median, 100%) [63]; human immunodeficiency virus patients undergoing antiretroviral therapy (94-100%) [53,64]. Our median VAS score of 98% was high, similar to previous studies, probably because the study participants were self-administering elderly patients who were highly motivated to take their medications, leading to a high VAS score.
PP and PIMs are important, closely drug-related issues, especially in elderly patients. To the best of our knowledge, this is the first study to evaluate the association of self-reported medication adherence with PP and PIMs in elderly patients in Japan. PP is generally defined as regular use of multiple drugs and is associated with an increased risk of ADEs, hospital admission, and mortality, especially in elderly patients [12,[65][66][67][68]. PP is prevalent in elderly patients in Japan. In a nationwide retrospective study using 240 million pharmacy claims data items, Onoue et al. (2018) reported that 69.0% of elderly patients were PP(+) [9]. Ishizaki et al. (2020) also reported that non-excessive PP (5-9 medications) and excessive PP (≥10 medications) was seen, respectively, in 45.3% and 18.2% elderly patients (≥75 years old) [7]. In our study (as shown in Table 2), old (≥75 years) patients had a significantly higher number of PP(+) patients as compared to pre-old (65-74 years) patients (76.5% vs. 52.7%, p = 0.003), and 64.1% PP(+) patients had a higher median age (75 years) compared to PP(-) patients (72 years), which is supported by previous studies [7,9].
In our study, ODP dispensing was also associated with PP. PP leads to medication non-adherence [16,17], and ODP dispensing prevents the chance of the unintentional missing of doses, promoting patient safety, and improves medication adherence [69]. In addition, the number of consulting medical institutions was associated with PP and PIM use. Cross-sectional studies in Japan have reported that patients who were prescribed by two or more physicians or who consulted more medical institutions are more likely to have PP and PIM use [8,70], as shown in our study. Furthermore, our study demonstrated that there is also a significant association between PP and PIM use, which is supported by previous reports [8,[20][21][22][23][24][25][26][27]. The frequency of elderly PIM(+) patients in Japan varies from 22.9% to 67.3% [8,22,23,27,31,35,68,71,72] because of different study populations and settings, different definitions of PIMs, or different timings of the investigation.
Any PIM use was significantly associated with self-reported full medication adherence in elderly patients. The proportion of PIMs increases with increasing age [23], and medication non-adherence is problematic for elderly patients [36], which leads us to hypothesize that patients taking any PIM are non-adherent to medication; however, our result indicates the opposite. This is probably because we selected elderly patients who can self-administer their medications as study patients. The drug list in STOPP-J is a "List of Medications That Require Particularly Careful Administration," so patients who can self-administer their medications might have been educated and proactively checked by a physician or pharmacists to prevent ADEs or worsening outcomes. A previous study on medication adherence in atrial fibrillation patients taking direct oral anticoagulants reported a higher adherence of PP in elderly patients compared to younger patients [47]. Our study participants might have been highly motivated regarding their medications, as shown by their VAS scores. Patients who can self-administer their medications and have good medication adherence need to start drug therapy carefully and should be carefully monitored to avoid ADEs caused by continuous use of PIMs.
This study had a few limitations. First, this was a pilot study conducted in a single university hospital, and the study cohort was relatively small. Therefore, our findings might not be generalized to other hospitals or countries. In the future, a multicenter study is necessary to obtain enough sample sizes of the patients. Second, the VAS is a subjective adherence measurement tool and has never been validated in Japan. We defined a VAS score of 100% as a self-reported full medication adherence, as previously described [52,53]; however, previous studies have used a cut-off value of 80% to divide medication adherence into good or poor [48][49][50][51]. The evaluation of medication adherence from the perspective of a pharmacist and the application of an objective measurement tool could not be performed in this study. There are several methods of assessing medication adherence; however, there is no gold standard for measuring adherence [73]. The triangulation of methods is recommended to increase the validity and reliability of the adherence data collected [73]. Further studies are needed to evaluate the relationship between the VAS and other methods, such as administrative claims or electronic pill monitoring in Japan. Third, because of the retrospective study design, we did not examine clinical factors involved in medication adherence. Many potential factors might affect medication adherence, such as education level, severity and duration of illness, patients' understanding and beliefs about their illness, and medical cost [36]. Lastly, 64.1% and 56.3% patients had polypharmacy and at least one PIM, respectively, but these percentages may be difficult to compare to the previous reports, because we excluded individuals admitted for an emergency or unable to self-medicate. In previous studies which examined individuals using PP and PIMs, it is likely that the incidence rate of unplanned re-admissions and advanced cognitive impairment are associated with PP and PIMs [67,72]. Furthermore, we did not examine the administration period of the combined use of multiple antithrombotic agents, which may have been overestimated in the percentage of PIMs.

Conclusions
To the best of our knowledge, this is the first study to evaluate the association of self-reported medication adherence with PP and PIMs in a limited elderly patient population excluding emergency hospitalization cases in Japan. PP and PIM prevalence was not uncommon, and self-reported medication adherence was extremely high in elderly patients who can self-administer their medications. There was a significant association between PIM prescription and self-reported full medication adherence. Pharmacists should pay more attention to prescribed medications of self-administering elderly patients in order to improve their prescribing quality.

Conflicts of Interest:
The authors declare no conflict of interest.