- Email: [email protected]
Deep-learning-based fault detection and diagnosis of air-handling units
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Research in Social and Administrative Pharmacy journal homepage: www.elsevier.com/locate/rsap
Triple strength utility of the Modified Drug Adherence Work-Up (M-DRAW) tool in a veterans affairs outpatient diabetes clinic Sun Leea, Yuna H. Bae-Shaawb, Hyma Goginenic, Marcia M. Worleyd, Anandi V. Lawc,∗ a
High Point University, School of Pharmacy, Department of Clinical Sciences, One University Parkway, High Point, NC, 27268, USA University of Southern California, School of Pharmacy, Schaeffer Center for Health Policy and Economics, 635 Downey Way, Verna & Peter Dauterive Hall (VPD), 2nd Floor, Los Angeles, CA, 90089, USA c Western University of Health Sciences, College of Pharmacy, Department of Pharmacy Practice and Administration, 309 East Second Street., Pomona, CA, 91766, USA d The Ohio State University, College of Pharmacy, Division of Pharmacy Practice and Science, A214 Parks Hall, 500 W. 12th Avenue, Columbus, OH, 43210, USA b
A B S T R A C T
Objective: The purpose of this study was to test the psychometric properties of the M-DRAW tool and to examine its applicability and utility at a primary clinic setting in patients with diabetes. Methods: A prospective, pre-post interview design study was conducted at the VA Loma Linda Health System (VALLHS) from 03/2017–03/2018. Eligibility criteria consisted of English-speaking patrons who were 18 years and older, diagnosed with Type 2 diabetes mellitus, residing in non-institutional setting, and having 1 + prescriptions for diabetes. A priming question about self-reported adherence was used to assign participants to control (Group A) or intervention (Group N). Pharmacist-led interventions were thus directed to those who recognized their medication nonadherence issue. The M-DRAW tool consisted of 13 statements about barriers to adherence on a 4-point frequency scale. A “3 = sometimes” or “4 = often” on each item indicated a barrier to adherence that was then addressed using the GUIDE strategy using motivational interviewing with the participant. Results: Of the 200 eligible individuals, 88 participants completed both baseline and follow-up assessments (Group A, n = 63; Group N, n = 25). Participants were male (98.8%), taking 7–8 medications on average, and using insulin (79.5%). The tool yielded good internal consistency (Cronbach's alpha = 0.873). Using confirmatory factor analysis, four factors were extracted with items loading as hypothesized. At baseline, group N identified three times greater number of barriers from the M-DRAW tool compared to Group A (5.1 items vs. 1.7, p < 0.05). At 3-month follow-up, a decrease in the number of barriers was observed among Group N. Both PDC and HbA1c did not result in statistically significant reduction in pre-post change. Conclusions: The M-DRAW tool is shown to be reliable and valid. A tailored intervention reduced the number of barriers contributing to medication nonadherence and resulted in a trend of improved clinical outcomes.
Introduction One of the most effective ways to delay progression of chronic conditions is by patients taking medication as prescribed.1 This statement has even more significance in patients with diabetes since medication adherence can delay the serious health complications that result from poor glycemic control, ranging from vision loss and amputation of limbs to heart disease, stroke, kidney failure, and death.2,3 In 2015, the prevalence of diabetes in the US was a record high of 30.3 million Americans, with an incidence rate of 1.5 million people annually.4 Yet, as seen from various studies, four in ten patients with type 2 diabetes mellitus are not adherent to their prescribed therapy, and about 50% of patients with diabetes fail to achieve optimal glycemic control.5,6 Various interventions have been used to encourage patient adherence to medications, including refill reminder systems, medication synchronization programs, education sessions, and monetary incentives
provided by the insurance plan.7–11 In spite of the these efforts, diabetes mellitus remains the 7th leading cause of death in US3,4 and medication nonadherence continues to be a problem costing $300 billion a year in the US.7 One of the approaches being used by various entities is to identify reasons for medication nonadherence so as to provide an individualized and customized solution to address these reasons.11–16 The American Medical Association issued a STEPS Forward™ module encouraging clinicians to develop a process to routinely assess reasons for medication adherence.12 Researchers have identified and categorized key reasons for nonadherence into five domains: patient, therapy, condition, socioeconomic, and health system-related factors.17 Many selfreported adherence measures (i.e., Morisky scale,14 Brief Medication Questionnaire,15 Adherence Barrier Questionnaire18) may serve as an initial screening step for patients who may be at risk for medication nonadherence; however, they are lengthy, take a one-size-fits-all
∗
Corresponding author. E-mail addresses: [email protected] (S. Lee), [email protected] (Y.H. Bae-Shaaw), [email protected] (H. Gogineni), [email protected] (M.M. Worley), [email protected] (A.V. Law). https://doi.org/10.1016/j.sapharm.2019.09.063 Received 14 March 2019; Received in revised form 17 September 2019; Accepted 26 September 2019 1551-7411/ © 2019 Published by Elsevier Inc.
Please cite this article as: Sun Lee, et al., Research in Social and Administrative Pharmacy, https://doi.org/10.1016/j.sapharm.2019.09.063
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
Participant allocation
approach, and/or lack strategies for addressing the identified reasons for medication nonadherence. The Drug Adherence Work-up (DRAW) tool covers areas commonly contributing to nonadherence, such as side effects, complexity of drug regimen, and cost consideration.13 The tool also provides a standardized (Guide) strategy to address each problem. Our research team modified the DRAW tool to include additional factors such as beliefs, attitudes, and concerns related to health and medication use.19 The use of motivational interviewing (MI) to help strategize and guide patients to improve medication adherence was limited in previous research.13 Therefore, a MI-trained ambulatory care pharmacist was involved to provide face-and-content validity of the tool and the Guide Strategy during the development of the tool prior to the pilot study. The resultant Modified Drug Adherence Work-up (M-DRAW) tool (Appendix A) was recently tested and found to be reliable (measured by internal consistency) and possess discriminant validity.19 Construct validity was not performed due to limited sample size and high attrition rate in the validation study. Studies have shown the value of pharmacist-led intervention in diabetes care management; therefore, the current study was conducted to assess the usefulness of the M-DRAW tool in a primary clinic followed by a targeted pharmacist-led intervention to address medication adherence issue among patients with diabetes.11,20,21 The current study also further tested the psychometric properties of the MDRAW tool and examined the applicability and utility of the tool in a primary clinic setting.
Similar to the initial validation study,19 the priming question was used to group participants into self-reported adherer (Group A) vs. selfreported non-adherers (Group N). The following single-item question was used to capture self-reported adherence level: Priming Question: You have been prescribed medication(s) for your health condition(s) which is to be taken regularly. How would you describe your past experience with taking your medication(s)? (1) I take my medication(s) regularly (9 out of 10 times). (2) I want to be very regular in taking my medication(s), but I am not always good with it due to some challenges. (3) I am not very regular in taking my medication(s) because I feel unwilling. The priming question was self-administered on a pencil-and-paper format. Participants who selected (1) were allocated into Group A. Participants who answered (2) or (3) were allocated into Group N. This self-selection group assignment method was used so that the interventional strategies could be directed to those who recognize their medication nonadherence issue. Data collection Participants were recruited for the study from March 2017 to March 2018. At initial enrollment, each participant was asked to complete (1) informed consent, (2) Health Insurance Portability and Accountability Act (HIPAA) form approved by the VA IRB, (3) California Experimental Subject's Bill of Rights form, (4) priming question and, (5) M-DRAW tool. Shortly after the enrollment, the research pharmacist (SL) reviewed the Electronic Health Record (EHR) to collect baseline characteristics (age, gender, number of active prescription medications), clinical measures (HbA1c, blood pressure, and lipid panel), and refill history to calculate proportion of days covered (PDC)22 as an objective measure of adherence level for each participant. At or after 3 months from the enrollment date, participants were followed up over the phone or in person to complete the M-DRAW tool again. During the follow-up, two additional questions were included, that related to helpfulness and usefulness of the study content and tool . Participant clinical measures and refill history were reviewed via EHR (Fig. 1). The research pharmacist (SL) reviewed the refill history and manually calculated PDC using Microsoft Excel. One notable point to mention regarding the refill history was that there were no auto-refills available to patients at the VALLHS; therefore, all refill history was based on patients' request of medication refills.
Method Study design A prospective, pre-post interview design study was conducted at a Community-Based Outpatient Clinic (CBOC) in VA Loma Linda Health System (VALLHS) in Loma Linda, California. Eligibility criteria consisted of English-speaking patrons who were 18 years and older, diagnosed with type 2 diabetes mellitus, receiving medical care at the VALLHS, residing in non-institutional setting, and taking 1 or more prescription medication(s) for diabetes. Eligible participants were referred to the study by providers and nurses at the clinic. These potential participants were contacted in person promptly after their clinic visits or over-the-phone. This study protocol was approved by the Institutional Review Board at VALLHS (ID: 01187). Eligible and willing participants were informed about the study and provided informed consent to participate in the study. Participants were provided with a $10 gift card at the end of the follow-up visit as a token of appreciation for their time and effort.
Description of the study site
Intervention
The primary care clinic consisted of 6 physicians, 8 registered nurses, 1 clinic manager (Registered Nurse), 4 lab technicians, and 2 support staff members. Any of the providers could refer eligible participants to the study. The telehealth pharmacist (HG) held the DM-clinic via computer monitor in CBOC while physically located at the main campus of VALLHS. The telehealth pharmacist referred eligible patients to the research pharmacist (SL) located at the CBOC. When patients visited the clinic during their usual care appointment with the providers (including physicians, nurses, and a telehealth pharmacist), the providers screened patients for their eligibility and referred to the study. The research fellow (SL) who is also a MI-trained licensed pharmacist, administered the survey and intervention in both initial and follow-up contact with the participants. At the time of survey administration, each participant was given a thorough explanation about the purpose of the study being medication adherence.
Once participants were enrolled, the research fellow (SL) reviewed responses to the priming question and assigned participants to their relevant groups. Participants in both groups completed the research consent, priming question, and the M-DRAW tool. Group A did not receive any intervention, but was informed that a pharmacist would be in contact in 3 months to gather the follow-up response. Group N received a tailored intervention led-by the pharmacist (SL). Responses to M-DRAW tool items were dichotomized and recoded to count the number of barriers (i.e., each item answered ‘1 = never’ or ‘2 = rarely’ was recoded to ‘0’ and each item answered ‘3 = sometimes’ or ‘4 = often’ were recoded to ‘1’). Any items that were answered either “3 = sometimes” or “4 = often” were identified to be a potential problem contributing to the participant's nonadherence issue; therefore, the pharmacist provided interventions as specified by the Guided Strategies for Increasing Adherence (Guide Strategy) as a part of the MDRAW tool. (Appendix 1). 2
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
Fig. 1. Study design.
Statistical analysis
Table 1 Baseline characteristics between Groups A and N identified using the priming question.
Data were collected through Microsoft Excel and analyzed using SPSS version 24.0.23 The study was designed to detect a moderate effect size of 0.5 between the control group and intervention group. We determined that total sample size of 64 was needed for each group in order to provide a power of 80% using a two-sided alpha level of 0.05. Baseline characteristics were compared between the two groups using independent t-tests for continuous variables (i.e., age, number of barriers identified using the M-DRAW tool, proportion of days covered, Hemoglobin A1c, and other clinical markers). Chi-square test was used to compare the categorical variables (i.e., gender, two quality improvement questions asking about satisfaction with pharmacist interaction). A Difference-in-difference (D-I-D) model estimated the difference in pre-post changes of outcomes between the two groups. Itemspecific analysis was performed using Fisher's Exact Test to compare the number of participants in each group who answered “3 = sometimes” or “4 = often” on each item (Section 3.5). These outcomes included the number of barriers identified using the M-DRAW tool, 3-month PDC, and HbA1c. Confirmatory Factor Analysis was performed to test for construct validity.
Age, mean ± SD, yr Gender, male, n (%) Clinical Markers HbA1C (%) Systolic BP (mmHg) Diastolic BP (mmHg) # Rx Medications + SD PDC (%) Chronic Conditions, n (%) Diabetes Hypertension Dyslipidemia Use of insulin, n (%) Use of any medication organizational method, n (%)
Group A
Group N
(n = 63)
(n = 25)
65.8 ± 8.4 62 (98.4)
60.4 ± 8.7 25 (100.0)
0.011* 0.526
8.7 ± 1.7 131.5 ± 10.2 69.6 ± 6.1 7.5 ± 2.6 74.8 ± 21.9
9.5 ± 2.2 138.2 ± 12.6 74.1 ± 5.5 7.2 ± 4.1 67.1 ± 22.9
0.094 0.21 0.058 0.764 0.163
63 57 59 51 23
25 21 21 19 11
0.493 0.388 0.156 0.603 0.515
(100.0) (90.5) (93.7) (81.0) (36.5)
(100.0) (84.0) (84.0) (76.0) (44.0)
P-value
PDC, Proportion of Days Covered; HbA1c, Hemoglobin A1C; *p < 0.05.
Psychometric properties
Results
The internal consistency was measured by Cronbach's coefficient alpha (alpha = 0.873). The Kaiser-Meyer-Olkin (KMO) test yielded 0.829, deeming an adequate sample size to perform the confirmatory factor analysis (Bartlett's test of sphericity, p < 0.05). The researchers accepted factor loading above 0.4. Four factors were extracted with items loading as hypothesized; confirming the hypothesized domains (Table 2). Discriminant validity was confirmed based on the higher number of barriers contributing to medication adherence issue in Group N compared to Group A. Additional information about the validity of the M-DRAW tool was included in a previous publication.19
Response rate and allocation Of approximately 200 eligible individuals referred to the study, 91 agreed to enroll. A total of 88 participants completed both baseline and follow-up assessments. Reasons for not participating were not assessed as those patients were not seen by the research pharmacist (SL); however, among those who saw the research pharmacist, some felt unease of sharing their social security number in the HIPAA consent form as a part of the approved protocol by the VALLHS. Reasons for not completing the follow-up assessment were hospitalization (n = 2) and death (n = 1). Of the 88 participants included in the final analysis, 63 (71.6%) participants assigned themselves to Group A, and 25 (28.4%) to Group N. Of the 25 participants enrolled in Group N, only 3 participants reported to be non-adherent to medications due to unwillingness (Choice C of the Priming Question). A small number of people with self-identified nonadherence issues who answered option “C” were grouped along with Group N.
Outcome measures At baseline, group N identified three times as many barriers from the M-DRAW tool compared to Group A (5.1 items vs. 1.7, p < 0.05). There was a decrease in the number of barriers at the 3-month period among those who received the pharmacist-led intervention according to the Guide Strategy (Fig. 2). The baseline PDC level was not significantly different between the Group A vs. Group N (74.6 vs. 67.0, p > 0.05) (Fig. 3). The difference in pre-post change of the PDC level was not significantly different between the two groups (Δ = 6.3 vs. Δ = 10.5, p > 0.05). There was no difference in baseline HbA1c value between the two groups (9.5 vs. 8.7, p > 0.05) (Fig. 4). Likewise, the difference in the pre-post change of the outcome was not significantly different between
Demographic characteristics Baseline characteristics were comparable between the two groups, except lower mean age with Group N. In general, participants were male in gender, taking 7–8 prescription medications, using insulin to treat diabetes, and had other chronic conditions such as high blood pressure and dyslipidemia (Table 1). 3
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
Table 2 Confirmatory factor analysis result showing in factor matrix using Varimax rotation. Item
Description
Therapy-related
1 2 2a 2b 2c 3
Do you feel unsure about how/when to take your medications? Do you have any difficulty getting your medications on time from the pharmacy? Is it difficult for you to get to the pharmacy to pick up your medication? Is paying for your medications a burden on your finances? Do you forget to place refill request on time? Do you have difficulty keeping track of all your medication schedules throughout the day? (e.g., when to take each medication) Do your medications give you side effects that make you NOT want to take it? Do you worry about what foods or other medications might interact with your medication? Do you feel that you can take more or less of your medication than the prescribed dose to fit your lifestyle? Do you feel like you don't get any benefits from taking your medication? Do you feel uncomfortable about taking your medication while you are out with family and friends? Do you consider it a burden that you have to take your medications for the rest of your life? Do you have doubts about whether your health condition needs to be treated? Do you have doubts if taking your medication will improve your health condition in the long term? Do you feel that you are NOT receiving the best possible treatment available from your healthcare provider?
0.892
4 5 6 7 8 9 10 11 12
the two groups (Δ = −0.4 vs. Δ = −0.4, p > 0.05).
Socioeconomic
Patient-related
Condition-related
0.851 0.651 0.650 0.588 0.858 0.696 0.738 0.583 0.584 0.651 0.798 0.833 0.797 0.606
medication since the initial interview. Figs. 5 and 6 represent the results.
Item-specific analysis Discussion As described in Section 2.5, responses to M-DRAW tool items were dichotomized and recoded to count the number of barriers (i.e., each item answered ‘1 = never’ or ‘2 = rarely’ was recoded to ‘0’ and each item answered ‘3 = sometimes’ or ‘4 = often’ were recoded to ‘1’). Participants in Group N reported experiencing barriers contributing toward medication non-adherence on all items, except having difficulty accessing the pharmacy on time (Q2) and adjusting the prescribed dose to fit their lifestyle (Q6). Participants who received the intervention (Group N) showed a trend toward reduced identification of items as a barrier contributing to medication nonadherence at the follow-up assessment. Items assessed in Q1, Q5, Q8, and Q9 showed a statistically significant reduction.
This study examined the psychometric properties of the M-DRAW tool and its utility in an outpatient clinic setting. The M-DRAW tool was previously validated for reliability but not completely for validity.19 The current study showed the tool to be both reliable (internal consistency) and valid (construct and discriminant validity). An important contribution of the M-DRAW tool is the single-item priming question which helps examine self-perceived intentional or unintentional nonadherence. The priming question was able to identify participants with poor medication adherence. The M-DRAW tool identified a higher number of barriers to adherence among the non-adherers, supporting the tool's discriminant validity. Confirmatory factor analysis yielded four factors containing items aligning with the hypothesized domains (i.e., patient-related, therapy-related, condition-related, and socioeconomic factors) (Table 3). In terms of the utility of the M-DRAW tool in a primary care setting, tailored education provided by a pharmacist following the targeted Guide Strategy helped reduce the number of barriers contributing to medication adherence at the 3-month follow-up (Fig. 2).
Helpfulness and usefulness of the intervention At the 3-month follow-up, participants were asked if they found it helpful to enroll in the study and to meet with a pharmacist; they also were queried on if the intervention resulted in a change in their medication use behavior. About one third of participants in both groups indicated that the interview was very helpful. More than 50% of the patients in Group A reported no change in their medication-taking behaviors in the past 3 months. On the other hand, more than 70% of the patients who received the intervention (Group N) reported that they experienced ‘little’ or ‘significant’ change in the way they took the
Use of priming question asa conversation starter The priming question used to allocate participants into Group A (self-reported adherers) versus Group N (self-reported non-adherers),
Fig. 2. Number of barriers identified at the initial and follow-up visits. 4
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
Fig. 3. Changes of adherence level measured by the proportion of days covered ratio.
was specific in identifying patients who were ready to admit their medication nonadherence and take responsibility to improve their diabetes care management. One third of the study participants admitted that they were not adherent to their prescribed therapy, and more than 70% had a poor medication refill history (PDC < 80%). When the refill history was reviewed among Group A, it was noticeable that half of the Group A (50.8%) had poor refill history. That is, the priming question was specific in identifying nonadherence issues but had limited sensitivity in identifying adherers due to the self-reported nature of the group allocation. The medication refill history was reviewed after the initial interview, so the pharmacist who was leading the intervention had no objective measure of adherence level. In retrospect, it was evident that participants who were ready to admit their medication nonadherence problem were more engaged in the conversation and interested in learning about strategies to improve their medication taking behavior. Although not statistically significant, there was a trend of improved PDC in both groups. It is possible that being seen by a research pharmacist (SL) and participation in a medication adherencerelated study may have led to a mild Hawthorne effect or participants may have truly appreciated the benefit of subsequent medication refils. The priming question could be used as a conversation starter about medication adherence in a busy clinic setting as it takes less than a minute to answer the question. Therefore, the priming question could be incorporated as part of routine care in primary care settings to assess medication adherence.12
Fig. 5. Helpfulness of the pharmacist-led interview.
intervention.25 Although the trend of drop in HbA1c was observed at the 3-month follow-up, the magnitude of change observed in the intervention group was not significantly greater than the usual care group (Fig. 4). This finding may be due to small sample size or short follow-up period; therefore, a similar future study with a larger sample size and longer follow up may be needed to examine the change. Similarly, participants in both groups refilled their medications more regularly after being enrolled to the study, which may be explained by a Hawthorne effect increasing motivation to refill prescriptions after visiting the clinic (Fig. 3). The effectiveness of the intervention was not
Use of M-DRAW tool in a primary care setting Diabetes was chosen as the disease state for the study because it affects nearly one fourth of the VA's patient population.24 In addition, routine assessment of HbA1c can signal change in disease control potentially achieved by the study intervention. Similar to a review article by Doggrell, our study finding did not support the clinical benefit of the
Fig. 4. Changes of hemoglobin A1c value. 5
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
taking medication’ (Q9). Previous studies have identified modifiable, common patient-related factors contributing to medication adherence issue such as having less knowledge about the disease or drug and incongruence in health belief contributing to lack of motivation to adhere to drug regimen.26,27 Similar to the existing body of evidence, our study findings suggest that targeted education based on identified modifiable factors would help clinicians guide the conversation in a patient-specific manner. When specific questions were identified as potential barriers to medication adherence, the pharmacist followed the Guide Strategy to provide a customized approach to solve the problem based on the participant's need. If a study participant was having difficulty keeping track of a medication, medication reconciliation was performed to review and create a medication list with a patient. When participants allowed, the pharmacist helped set up an alarm on their cell phone to remind them when to take their medication. A pillbox was provided at the clinic, and the pharmacist reviewed how to use the pillbox. Participants often were concerned about drug-food or drug-drug interaction, and the pharmacists used reliable resources such as the Micromedex and Natural Medicines Comprehensive Database to show the result and discuss in a patient-friendly language.28,29 Depending on the patient's engagement in the conversation, the intervention lasted approximately 15 min–45 min. These standardized interventions may have helped both with timely identification of problems contributing to medication adherence and implementation of practical strategies tailored to the individuals' identified barrier. Overall, the M-DRAW can serve as an adherence tool due to its three strengths: discriminant validity of the priming question, identification of barriers, and Guide Strategy to resolve the barriers.
Fig. 6. Usefulness of the intervention.
remarkable in either the clinical marker or refill history; however, participants identified fewer number of barriers contributing to medication adherence using the M-DRAW tool at follow-up compared to baseline (Fig. 2). It is possible that the 3-month follow-up may have been too short to see the benefit of the intervention clinically and objectively. It is also possible that participants who agreed to participate in the study recognized the importance of ‘medication adherence’. At the 3-month follow-up, participants were asked to rate the helpfulness and usefulness of the initial interaction. While both groups answered that meeting a pharmacist at the study enrollment was helpful in their care management, only those participants who received the intervention found it useful in changing their medication intake behavior.
Strength, limitation, and future research The strengths of the study itself included a low attrition rate (88/91 were followed up) and completeness of the clinical markers collected using the electronic health record. The user-friendly language of the tool has enabled participants to independently read and answer both the priming questions and the thirteen statements on the M-DRAW tool. This finding points to the applicability of the tool in future studies to screen for study participants who are medication nonadherent. The results of the study should be interpreted with caution due to some limitations. In spite of the low attrition rate, the small sample size precluded multivariate regression analysis. The duration of follow-up may have been too short to observe effectiveness of the intervention via a clinical marker and refill history. The nature of the study site, which
Use of Guide Strategy to provide targeted intervention On average, participants in Group N identified all items, except Q2 and Q6, as potential barriers contributing to adherence issues (Table 3). Items that showed a statistically significant reduction between the initial and follow-up screening included ‘feeling unsure about how/when to take medication’ (Q1), ‘food or drug interaction’ (Q5), ‘uncomfortable taking medication in public’ (Q8), and ‘life-long burden of Table 3 Changes in answers before and after in each group. Group
Group A (n = 63)
Group N (n = 25)
Items
PRE (%)
POST (%)
PRE (%)
POST (%)
Q1 - Do you feel unsure about how/when to take your medications?*+ Q2 - Do you have any difficulty getting your medications on time from the pharmacy? Q3 - Do you have difficulty keeping track of all your medication schedules throughout the day? (e.g., when to take each medication)* Q4 - Do your medications give you side effects that make you NOT want to take it?* Q5 - Do you worry about what foods or other medications might interact with your medication?* + Q6 - Do you feel that you can take more or less of your medication than the prescribed dose to fit your lifestyle? Q7 - Do you feel like you don't get any benefits from taking your medication?* Q8 - Do you feel uncomfortable about taking your medication while you are out with family and friends?* + Q9 - Do you consider it a burden that you have to take your medications for the rest of your life?* + Q10 - Do you have doubts about whether your health condition needs to be treated?* Q11 - Do you have doubts if taking your medication will improve your health condition in the long term?* Q12 - Do you feel that you are NOT receiving the best possible treatment available from your healthcare provider?*
6.5 17.7 9.7
4.9 10.0 4.9
40.0 36.0 28.0
30.0 25.0 20.0
4.8 19.4 21.0 14.5 16.1 32.3 6.5 11.3 8.1
7.3 4.9 14.6 9.8 7.3 26.8 4.9 17.1 12.2
44.0 48.0 28.0 40.0 48.0 68.0 40.0 56.0 28.0
15.0 25.0 30.0 25.0 30.0 55.0 15.0 40.0 10.0
*p < 0.05 of Fisher's Exact Test comparing response between Group A and Group N at baseline (PRE-response); +p < 0.05 of Fisher's Exact Test comparing response between Group A and Group N at baseline (POST-response); As described in Section 2.5, responses to M-DRAW tool items were dichotomized and recoded to count the number of barriers (i.e., each item answered ‘1 = never’ or ‘2 = rarely’ was recoded to ‘0’ and each item answered ‘3 = sometimes’ or ‘4 = often’ were recoded to ‘1’). 6
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
provides care for predominantly older white males, may have limited the generalizability of the tool. Additionally, the convenience of having a private room to speak with participants for 15–45 min may limit transferability of the study procedure in a busy community pharmacy setting. Future studies with a longer follow-up and multiple clinic settings may help further evaluate the applicability and generalizability of the M-DRAW tool.
https://doi.org/10.1186/1472-6963-14-219. 9. Dao N, Lee S, Hata M, Sarino Lord. Impact of appointment-based medication synchronization on proportion of days covered for chronic medications. Pharm (Basel, Switzerland). 2018;6(2):44. https://doi.org/10.3390/pharmacy6020044. 10. Lee S, Khare MM, Olson HR, Chen AMH, Law AV. The TEACH trial: tailored education to assist label comprehension and health literacy. Res Soc Adm Pharm. 2018;14(9):839–845. https://doi.org/10.1016/j.sapharm.2018.05.015. 11. Lau SR, Kriegbaum M. Medication non-adherence in the context of situated uncertainty: moving beyond simple, dichotomous approaches. Res Soc Adm Pharm. 2018;14(8):742–748. https://doi.org/10.1016/J.SAPHARM.2017.09.003. 12. Medication adherence | STEPS forward. https://www.stepsforward.org/modules/ medication-adherence, Accessed date: 10 September 2018. 13. Doucette WR, Farris KB, Youland KM, Newland BA, Egerton SJ, Barnes JM. Development of the drug adherence work-up (DRAW) tool. J Am Pharm Assoc. 2012;52(6):e199–e204. https://doi.org/10.1331/JAPhA.2012.12001. 14. Moon SJ, Lee W-Y, Hwang JS, Hong YP, Morisky DE. Accuracy of a screening tool for medication adherence: a systematic review and meta-analysis of the Morisky Medication Adherence Scale-8. PLoS One. 2017;12(11):e0187139https://doi.org/10. 1371/journal.pone.0187139. 15. Svarstad BL, Chewning BA, Sleath BL, Claesson C. The Brief Medication Questionnaire: a tool for screening patient adherence and barriers to adherence. Patient Educ Couns. 1999;37(2):113–124http://www.ncbi.nlm.nih.gov/pubmed/ 14528539, Accessed date: 10 September 2018. 16. Jaam M, Awaisu A, Mohamed Ibrahim MI, Kheir N. A holistic conceptual framework model to describe medication adherence in and guide interventions in diabetes mellitus. Res Soc Adm Pharm. 2018;14(4):391–397. https://doi.org/10.1016/J. SAPHARM.2017.05.003. 17. World Health Organization(WHO). WHO - ADHERENCE TO LONG-TERM THERAPIES: EVIDENCE FOR ACTION. WHO; 2015. 18. Müller S, Kohlmann T, Wilke T. Validation of the Adherence Barriers Questionnaire an instrument for identifying potential risk factors associated with medication-related non-adherence. BMC Health Serv Res. 2015;15:153. https://doi.org/10.1186/ s12913-015-0809-0. 19. Lee S, Bae YH, Worley M, Law A. Validating the modified drug adherence work-up (M-DRAW) tool to identify and address barriers to medication adherence. Pharm (Basel, Switzerland). 2017;5(3):52. https://doi.org/10.3390/pharmacy5030052. 20. Presley B, Groot W, Pavlova M. Pharmacy-led interventions to improve medication adherence among adults with diabetes: a systematic review and meta-analysis. Res Soc Adm Pharm. October 2018. https://doi.org/10.1016/J.SAPHARM.2018.09.021. 21. Fredericksen RJ, Gibbons L, Brown S, et al. Medication understanding among patients living with multiple chronic conditions: implications for patient-reported measures of adherence. Res Soc Adm Pharm. 2018;14(6):540–544. https://doi.org/ 10.1016/j.sapharm.2017.06.009. 22. National Center for Chronic Disease Prevention and Health Promotion. Calculating Proportion of Days Covered (PDC) for Antihypertensive and Antidiabetic Medications: An Evaluation Guide for Grantees. United States; 2015https://www.cdc.gov/dhdsp/docs/ med-adherence-evaluation-tool.pdf. 23. IBM Corp. IBM SPSS Statistics for Windows, Version 24. Armonk, NY. Released. IBM Corp IBM Corp - Google Scholar; 2015 2015. 24. Lehnert H, Wittchen H-U, Pittrow D, et al. Prevalence and pharmacotherapy of diabetes mellitus in primary care. Dtsch Medizinische Wochenschrift. 2005. https:// doi.org/10.1055/s-2005-863050. 25. Doggrell SA. A review of interventions ≥ 6 months by pharmacists on adherence to medicines in cardiovascular disease: characteristics of what works and what doesn't. Res Soc Adm Pharm. 2019;15(2):119–129. https://doi.org/10.1016/J.SAPHARM. 2018.04.003. 26. Gellad WF, Grenard JL, Marcum ZA. A systematic review of barriers to medication adherence in the elderly: looking beyond cost and regimen complexity. Am J Geriatr Pharmacother. 2011;9(1):11–23. https://doi.org/10.1016/j.amjopharm.2011.02. 004. 27. Roebuck MC, Liberman JN. Impact of pharmacy benefit design on prescription drug utilization: a fixed effects analysis of plan sponsor data. Health Serv Res. 2009;44(3):988–1009. https://doi.org/10.1111/j.1475-6773.2008.00943.x. 28. Chatfield AJ. Lexicomp online and Micromedex 2.0. J Med Libr Assoc. 2015. https:// doi.org/10.3163/1536-5050.103.2.016. 29. Tomasulo P. Natural medicines Database. J Consum Health Internet. 2004. https://doi. org/10.1080/15424065.2016.1225541.
Conclusion Results from this study support the reliability and validity of the MDRAW tool, and its usefulness in a clinical practice setting to identify and address barriers to patient medication adherence. The priming question showed good discriminant validity with limited sensitivity. A tailored education followed by the M-DRAW tool's Guide Strategy in patients with diabetes mellitus appears to reduce self-reported barriers to medication adherence. Integrating the M-DRAW tool in a primary clinic setting can provide a timely assessment of medication adherence issues and encourage meaningful patient-provider interactions that can impact patient outcomes. Acknowledgement The authors would like to acknowledge Drs. Bill Doucette, Jan Kevookjian, and Emmanuelle Schwartzman for their advice and assistance with the tool and motivational interviewing techniques. Appendix A. Supplementary data Supplementary data to this article can be found online at. References 1. Centers for Disease Control and Prevention. National Diabetes Statistics Report 2017; 2017https://doi.org/10.1177/1527154408322560 2017. 2. Krass I, Schieback P, Dhippayom T. Adherence to diabetes medication: a systematic review. Diabet Med. 2015;32(6):725–737. https://doi.org/10.1111/dme.12651. 3. Data & statistics | diabetes | CDC. https://www.cdc.gov/diabetes/data/index.html, Accessed date: 7 September 2018. 4. Statistics About Diabetes: American Diabetes Association®. http://www.diabetes. org/diabetes-basics/statistics/, Accessed date: 2 May 2019. 5. Krass I, Schieback P, Dhippayom T. Adherence to diabetes medication: a systematic review. Diabet Med. 2015. https://doi.org/10.1111/dme.12651. 6. Ali MK, McKeever Bullard K, Imperatore G, Barker L, Gregg EW. Centers for Disease Control and Prevention (CDC). Characteristics associated with poor glycemic control among adults with self-reported diagnosed diabetes–National Health and Nutrition Examination Survey, United States, 2007-2010. MMWR Suppl. 2012;61(2):32–37http://www.ncbi.nlm.nih.gov/pubmed/22695461, Accessed date: 7 September 2018. 7. Viswanathan M, Golin CE, Jones CD, et al. Interventions to improve adherence to self-administered medications for chronic diseases in the United States. Ann Intern Med. 2012;157(11):785. https://doi.org/10.7326/0003-4819-157-11-20121204000538. 8. Kooy MJ, van Geffen EC, Heerdink ER, van Dijk L, Bouvy ML. Effects of a TELephone Counselling Intervention by Pharmacist (TelCIP) on medication adherence, patient beliefs and satisfaction with information for patients starting treatment: study protocol for a cluster randomized controlled trial. BMC Health Serv Res. 2014;14(1):219.
7
Contents lists available at ScienceDirect
Research in Social and Administrative Pharmacy journal homepage: www.elsevier.com/locate/rsap
Triple strength utility of the Modified Drug Adherence Work-Up (M-DRAW) tool in a veterans affairs outpatient diabetes clinic Sun Leea, Yuna H. Bae-Shaawb, Hyma Goginenic, Marcia M. Worleyd, Anandi V. Lawc,∗ a
High Point University, School of Pharmacy, Department of Clinical Sciences, One University Parkway, High Point, NC, 27268, USA University of Southern California, School of Pharmacy, Schaeffer Center for Health Policy and Economics, 635 Downey Way, Verna & Peter Dauterive Hall (VPD), 2nd Floor, Los Angeles, CA, 90089, USA c Western University of Health Sciences, College of Pharmacy, Department of Pharmacy Practice and Administration, 309 East Second Street., Pomona, CA, 91766, USA d The Ohio State University, College of Pharmacy, Division of Pharmacy Practice and Science, A214 Parks Hall, 500 W. 12th Avenue, Columbus, OH, 43210, USA b
A B S T R A C T
Objective: The purpose of this study was to test the psychometric properties of the M-DRAW tool and to examine its applicability and utility at a primary clinic setting in patients with diabetes. Methods: A prospective, pre-post interview design study was conducted at the VA Loma Linda Health System (VALLHS) from 03/2017–03/2018. Eligibility criteria consisted of English-speaking patrons who were 18 years and older, diagnosed with Type 2 diabetes mellitus, residing in non-institutional setting, and having 1 + prescriptions for diabetes. A priming question about self-reported adherence was used to assign participants to control (Group A) or intervention (Group N). Pharmacist-led interventions were thus directed to those who recognized their medication nonadherence issue. The M-DRAW tool consisted of 13 statements about barriers to adherence on a 4-point frequency scale. A “3 = sometimes” or “4 = often” on each item indicated a barrier to adherence that was then addressed using the GUIDE strategy using motivational interviewing with the participant. Results: Of the 200 eligible individuals, 88 participants completed both baseline and follow-up assessments (Group A, n = 63; Group N, n = 25). Participants were male (98.8%), taking 7–8 medications on average, and using insulin (79.5%). The tool yielded good internal consistency (Cronbach's alpha = 0.873). Using confirmatory factor analysis, four factors were extracted with items loading as hypothesized. At baseline, group N identified three times greater number of barriers from the M-DRAW tool compared to Group A (5.1 items vs. 1.7, p < 0.05). At 3-month follow-up, a decrease in the number of barriers was observed among Group N. Both PDC and HbA1c did not result in statistically significant reduction in pre-post change. Conclusions: The M-DRAW tool is shown to be reliable and valid. A tailored intervention reduced the number of barriers contributing to medication nonadherence and resulted in a trend of improved clinical outcomes.
Introduction One of the most effective ways to delay progression of chronic conditions is by patients taking medication as prescribed.1 This statement has even more significance in patients with diabetes since medication adherence can delay the serious health complications that result from poor glycemic control, ranging from vision loss and amputation of limbs to heart disease, stroke, kidney failure, and death.2,3 In 2015, the prevalence of diabetes in the US was a record high of 30.3 million Americans, with an incidence rate of 1.5 million people annually.4 Yet, as seen from various studies, four in ten patients with type 2 diabetes mellitus are not adherent to their prescribed therapy, and about 50% of patients with diabetes fail to achieve optimal glycemic control.5,6 Various interventions have been used to encourage patient adherence to medications, including refill reminder systems, medication synchronization programs, education sessions, and monetary incentives
provided by the insurance plan.7–11 In spite of the these efforts, diabetes mellitus remains the 7th leading cause of death in US3,4 and medication nonadherence continues to be a problem costing $300 billion a year in the US.7 One of the approaches being used by various entities is to identify reasons for medication nonadherence so as to provide an individualized and customized solution to address these reasons.11–16 The American Medical Association issued a STEPS Forward™ module encouraging clinicians to develop a process to routinely assess reasons for medication adherence.12 Researchers have identified and categorized key reasons for nonadherence into five domains: patient, therapy, condition, socioeconomic, and health system-related factors.17 Many selfreported adherence measures (i.e., Morisky scale,14 Brief Medication Questionnaire,15 Adherence Barrier Questionnaire18) may serve as an initial screening step for patients who may be at risk for medication nonadherence; however, they are lengthy, take a one-size-fits-all
∗
Corresponding author. E-mail addresses: [email protected] (S. Lee), [email protected] (Y.H. Bae-Shaaw), [email protected] (H. Gogineni), [email protected] (M.M. Worley), [email protected] (A.V. Law). https://doi.org/10.1016/j.sapharm.2019.09.063 Received 14 March 2019; Received in revised form 17 September 2019; Accepted 26 September 2019 1551-7411/ © 2019 Published by Elsevier Inc.
Please cite this article as: Sun Lee, et al., Research in Social and Administrative Pharmacy, https://doi.org/10.1016/j.sapharm.2019.09.063
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
Participant allocation
approach, and/or lack strategies for addressing the identified reasons for medication nonadherence. The Drug Adherence Work-up (DRAW) tool covers areas commonly contributing to nonadherence, such as side effects, complexity of drug regimen, and cost consideration.13 The tool also provides a standardized (Guide) strategy to address each problem. Our research team modified the DRAW tool to include additional factors such as beliefs, attitudes, and concerns related to health and medication use.19 The use of motivational interviewing (MI) to help strategize and guide patients to improve medication adherence was limited in previous research.13 Therefore, a MI-trained ambulatory care pharmacist was involved to provide face-and-content validity of the tool and the Guide Strategy during the development of the tool prior to the pilot study. The resultant Modified Drug Adherence Work-up (M-DRAW) tool (Appendix A) was recently tested and found to be reliable (measured by internal consistency) and possess discriminant validity.19 Construct validity was not performed due to limited sample size and high attrition rate in the validation study. Studies have shown the value of pharmacist-led intervention in diabetes care management; therefore, the current study was conducted to assess the usefulness of the M-DRAW tool in a primary clinic followed by a targeted pharmacist-led intervention to address medication adherence issue among patients with diabetes.11,20,21 The current study also further tested the psychometric properties of the MDRAW tool and examined the applicability and utility of the tool in a primary clinic setting.
Similar to the initial validation study,19 the priming question was used to group participants into self-reported adherer (Group A) vs. selfreported non-adherers (Group N). The following single-item question was used to capture self-reported adherence level: Priming Question: You have been prescribed medication(s) for your health condition(s) which is to be taken regularly. How would you describe your past experience with taking your medication(s)? (1) I take my medication(s) regularly (9 out of 10 times). (2) I want to be very regular in taking my medication(s), but I am not always good with it due to some challenges. (3) I am not very regular in taking my medication(s) because I feel unwilling. The priming question was self-administered on a pencil-and-paper format. Participants who selected (1) were allocated into Group A. Participants who answered (2) or (3) were allocated into Group N. This self-selection group assignment method was used so that the interventional strategies could be directed to those who recognize their medication nonadherence issue. Data collection Participants were recruited for the study from March 2017 to March 2018. At initial enrollment, each participant was asked to complete (1) informed consent, (2) Health Insurance Portability and Accountability Act (HIPAA) form approved by the VA IRB, (3) California Experimental Subject's Bill of Rights form, (4) priming question and, (5) M-DRAW tool. Shortly after the enrollment, the research pharmacist (SL) reviewed the Electronic Health Record (EHR) to collect baseline characteristics (age, gender, number of active prescription medications), clinical measures (HbA1c, blood pressure, and lipid panel), and refill history to calculate proportion of days covered (PDC)22 as an objective measure of adherence level for each participant. At or after 3 months from the enrollment date, participants were followed up over the phone or in person to complete the M-DRAW tool again. During the follow-up, two additional questions were included, that related to helpfulness and usefulness of the study content and tool . Participant clinical measures and refill history were reviewed via EHR (Fig. 1). The research pharmacist (SL) reviewed the refill history and manually calculated PDC using Microsoft Excel. One notable point to mention regarding the refill history was that there were no auto-refills available to patients at the VALLHS; therefore, all refill history was based on patients' request of medication refills.
Method Study design A prospective, pre-post interview design study was conducted at a Community-Based Outpatient Clinic (CBOC) in VA Loma Linda Health System (VALLHS) in Loma Linda, California. Eligibility criteria consisted of English-speaking patrons who were 18 years and older, diagnosed with type 2 diabetes mellitus, receiving medical care at the VALLHS, residing in non-institutional setting, and taking 1 or more prescription medication(s) for diabetes. Eligible participants were referred to the study by providers and nurses at the clinic. These potential participants were contacted in person promptly after their clinic visits or over-the-phone. This study protocol was approved by the Institutional Review Board at VALLHS (ID: 01187). Eligible and willing participants were informed about the study and provided informed consent to participate in the study. Participants were provided with a $10 gift card at the end of the follow-up visit as a token of appreciation for their time and effort.
Description of the study site
Intervention
The primary care clinic consisted of 6 physicians, 8 registered nurses, 1 clinic manager (Registered Nurse), 4 lab technicians, and 2 support staff members. Any of the providers could refer eligible participants to the study. The telehealth pharmacist (HG) held the DM-clinic via computer monitor in CBOC while physically located at the main campus of VALLHS. The telehealth pharmacist referred eligible patients to the research pharmacist (SL) located at the CBOC. When patients visited the clinic during their usual care appointment with the providers (including physicians, nurses, and a telehealth pharmacist), the providers screened patients for their eligibility and referred to the study. The research fellow (SL) who is also a MI-trained licensed pharmacist, administered the survey and intervention in both initial and follow-up contact with the participants. At the time of survey administration, each participant was given a thorough explanation about the purpose of the study being medication adherence.
Once participants were enrolled, the research fellow (SL) reviewed responses to the priming question and assigned participants to their relevant groups. Participants in both groups completed the research consent, priming question, and the M-DRAW tool. Group A did not receive any intervention, but was informed that a pharmacist would be in contact in 3 months to gather the follow-up response. Group N received a tailored intervention led-by the pharmacist (SL). Responses to M-DRAW tool items were dichotomized and recoded to count the number of barriers (i.e., each item answered ‘1 = never’ or ‘2 = rarely’ was recoded to ‘0’ and each item answered ‘3 = sometimes’ or ‘4 = often’ were recoded to ‘1’). Any items that were answered either “3 = sometimes” or “4 = often” were identified to be a potential problem contributing to the participant's nonadherence issue; therefore, the pharmacist provided interventions as specified by the Guided Strategies for Increasing Adherence (Guide Strategy) as a part of the MDRAW tool. (Appendix 1). 2
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
Fig. 1. Study design.
Statistical analysis
Table 1 Baseline characteristics between Groups A and N identified using the priming question.
Data were collected through Microsoft Excel and analyzed using SPSS version 24.0.23 The study was designed to detect a moderate effect size of 0.5 between the control group and intervention group. We determined that total sample size of 64 was needed for each group in order to provide a power of 80% using a two-sided alpha level of 0.05. Baseline characteristics were compared between the two groups using independent t-tests for continuous variables (i.e., age, number of barriers identified using the M-DRAW tool, proportion of days covered, Hemoglobin A1c, and other clinical markers). Chi-square test was used to compare the categorical variables (i.e., gender, two quality improvement questions asking about satisfaction with pharmacist interaction). A Difference-in-difference (D-I-D) model estimated the difference in pre-post changes of outcomes between the two groups. Itemspecific analysis was performed using Fisher's Exact Test to compare the number of participants in each group who answered “3 = sometimes” or “4 = often” on each item (Section 3.5). These outcomes included the number of barriers identified using the M-DRAW tool, 3-month PDC, and HbA1c. Confirmatory Factor Analysis was performed to test for construct validity.
Age, mean ± SD, yr Gender, male, n (%) Clinical Markers HbA1C (%) Systolic BP (mmHg) Diastolic BP (mmHg) # Rx Medications + SD PDC (%) Chronic Conditions, n (%) Diabetes Hypertension Dyslipidemia Use of insulin, n (%) Use of any medication organizational method, n (%)
Group A
Group N
(n = 63)
(n = 25)
65.8 ± 8.4 62 (98.4)
60.4 ± 8.7 25 (100.0)
0.011* 0.526
8.7 ± 1.7 131.5 ± 10.2 69.6 ± 6.1 7.5 ± 2.6 74.8 ± 21.9
9.5 ± 2.2 138.2 ± 12.6 74.1 ± 5.5 7.2 ± 4.1 67.1 ± 22.9
0.094 0.21 0.058 0.764 0.163
63 57 59 51 23
25 21 21 19 11
0.493 0.388 0.156 0.603 0.515
(100.0) (90.5) (93.7) (81.0) (36.5)
(100.0) (84.0) (84.0) (76.0) (44.0)
P-value
PDC, Proportion of Days Covered; HbA1c, Hemoglobin A1C; *p < 0.05.
Psychometric properties
Results
The internal consistency was measured by Cronbach's coefficient alpha (alpha = 0.873). The Kaiser-Meyer-Olkin (KMO) test yielded 0.829, deeming an adequate sample size to perform the confirmatory factor analysis (Bartlett's test of sphericity, p < 0.05). The researchers accepted factor loading above 0.4. Four factors were extracted with items loading as hypothesized; confirming the hypothesized domains (Table 2). Discriminant validity was confirmed based on the higher number of barriers contributing to medication adherence issue in Group N compared to Group A. Additional information about the validity of the M-DRAW tool was included in a previous publication.19
Response rate and allocation Of approximately 200 eligible individuals referred to the study, 91 agreed to enroll. A total of 88 participants completed both baseline and follow-up assessments. Reasons for not participating were not assessed as those patients were not seen by the research pharmacist (SL); however, among those who saw the research pharmacist, some felt unease of sharing their social security number in the HIPAA consent form as a part of the approved protocol by the VALLHS. Reasons for not completing the follow-up assessment were hospitalization (n = 2) and death (n = 1). Of the 88 participants included in the final analysis, 63 (71.6%) participants assigned themselves to Group A, and 25 (28.4%) to Group N. Of the 25 participants enrolled in Group N, only 3 participants reported to be non-adherent to medications due to unwillingness (Choice C of the Priming Question). A small number of people with self-identified nonadherence issues who answered option “C” were grouped along with Group N.
Outcome measures At baseline, group N identified three times as many barriers from the M-DRAW tool compared to Group A (5.1 items vs. 1.7, p < 0.05). There was a decrease in the number of barriers at the 3-month period among those who received the pharmacist-led intervention according to the Guide Strategy (Fig. 2). The baseline PDC level was not significantly different between the Group A vs. Group N (74.6 vs. 67.0, p > 0.05) (Fig. 3). The difference in pre-post change of the PDC level was not significantly different between the two groups (Δ = 6.3 vs. Δ = 10.5, p > 0.05). There was no difference in baseline HbA1c value between the two groups (9.5 vs. 8.7, p > 0.05) (Fig. 4). Likewise, the difference in the pre-post change of the outcome was not significantly different between
Demographic characteristics Baseline characteristics were comparable between the two groups, except lower mean age with Group N. In general, participants were male in gender, taking 7–8 prescription medications, using insulin to treat diabetes, and had other chronic conditions such as high blood pressure and dyslipidemia (Table 1). 3
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
Table 2 Confirmatory factor analysis result showing in factor matrix using Varimax rotation. Item
Description
Therapy-related
1 2 2a 2b 2c 3
Do you feel unsure about how/when to take your medications? Do you have any difficulty getting your medications on time from the pharmacy? Is it difficult for you to get to the pharmacy to pick up your medication? Is paying for your medications a burden on your finances? Do you forget to place refill request on time? Do you have difficulty keeping track of all your medication schedules throughout the day? (e.g., when to take each medication) Do your medications give you side effects that make you NOT want to take it? Do you worry about what foods or other medications might interact with your medication? Do you feel that you can take more or less of your medication than the prescribed dose to fit your lifestyle? Do you feel like you don't get any benefits from taking your medication? Do you feel uncomfortable about taking your medication while you are out with family and friends? Do you consider it a burden that you have to take your medications for the rest of your life? Do you have doubts about whether your health condition needs to be treated? Do you have doubts if taking your medication will improve your health condition in the long term? Do you feel that you are NOT receiving the best possible treatment available from your healthcare provider?
0.892
4 5 6 7 8 9 10 11 12
the two groups (Δ = −0.4 vs. Δ = −0.4, p > 0.05).
Socioeconomic
Patient-related
Condition-related
0.851 0.651 0.650 0.588 0.858 0.696 0.738 0.583 0.584 0.651 0.798 0.833 0.797 0.606
medication since the initial interview. Figs. 5 and 6 represent the results.
Item-specific analysis Discussion As described in Section 2.5, responses to M-DRAW tool items were dichotomized and recoded to count the number of barriers (i.e., each item answered ‘1 = never’ or ‘2 = rarely’ was recoded to ‘0’ and each item answered ‘3 = sometimes’ or ‘4 = often’ were recoded to ‘1’). Participants in Group N reported experiencing barriers contributing toward medication non-adherence on all items, except having difficulty accessing the pharmacy on time (Q2) and adjusting the prescribed dose to fit their lifestyle (Q6). Participants who received the intervention (Group N) showed a trend toward reduced identification of items as a barrier contributing to medication nonadherence at the follow-up assessment. Items assessed in Q1, Q5, Q8, and Q9 showed a statistically significant reduction.
This study examined the psychometric properties of the M-DRAW tool and its utility in an outpatient clinic setting. The M-DRAW tool was previously validated for reliability but not completely for validity.19 The current study showed the tool to be both reliable (internal consistency) and valid (construct and discriminant validity). An important contribution of the M-DRAW tool is the single-item priming question which helps examine self-perceived intentional or unintentional nonadherence. The priming question was able to identify participants with poor medication adherence. The M-DRAW tool identified a higher number of barriers to adherence among the non-adherers, supporting the tool's discriminant validity. Confirmatory factor analysis yielded four factors containing items aligning with the hypothesized domains (i.e., patient-related, therapy-related, condition-related, and socioeconomic factors) (Table 3). In terms of the utility of the M-DRAW tool in a primary care setting, tailored education provided by a pharmacist following the targeted Guide Strategy helped reduce the number of barriers contributing to medication adherence at the 3-month follow-up (Fig. 2).
Helpfulness and usefulness of the intervention At the 3-month follow-up, participants were asked if they found it helpful to enroll in the study and to meet with a pharmacist; they also were queried on if the intervention resulted in a change in their medication use behavior. About one third of participants in both groups indicated that the interview was very helpful. More than 50% of the patients in Group A reported no change in their medication-taking behaviors in the past 3 months. On the other hand, more than 70% of the patients who received the intervention (Group N) reported that they experienced ‘little’ or ‘significant’ change in the way they took the
Use of priming question asa conversation starter The priming question used to allocate participants into Group A (self-reported adherers) versus Group N (self-reported non-adherers),
Fig. 2. Number of barriers identified at the initial and follow-up visits. 4
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
Fig. 3. Changes of adherence level measured by the proportion of days covered ratio.
was specific in identifying patients who were ready to admit their medication nonadherence and take responsibility to improve their diabetes care management. One third of the study participants admitted that they were not adherent to their prescribed therapy, and more than 70% had a poor medication refill history (PDC < 80%). When the refill history was reviewed among Group A, it was noticeable that half of the Group A (50.8%) had poor refill history. That is, the priming question was specific in identifying nonadherence issues but had limited sensitivity in identifying adherers due to the self-reported nature of the group allocation. The medication refill history was reviewed after the initial interview, so the pharmacist who was leading the intervention had no objective measure of adherence level. In retrospect, it was evident that participants who were ready to admit their medication nonadherence problem were more engaged in the conversation and interested in learning about strategies to improve their medication taking behavior. Although not statistically significant, there was a trend of improved PDC in both groups. It is possible that being seen by a research pharmacist (SL) and participation in a medication adherencerelated study may have led to a mild Hawthorne effect or participants may have truly appreciated the benefit of subsequent medication refils. The priming question could be used as a conversation starter about medication adherence in a busy clinic setting as it takes less than a minute to answer the question. Therefore, the priming question could be incorporated as part of routine care in primary care settings to assess medication adherence.12
Fig. 5. Helpfulness of the pharmacist-led interview.
intervention.25 Although the trend of drop in HbA1c was observed at the 3-month follow-up, the magnitude of change observed in the intervention group was not significantly greater than the usual care group (Fig. 4). This finding may be due to small sample size or short follow-up period; therefore, a similar future study with a larger sample size and longer follow up may be needed to examine the change. Similarly, participants in both groups refilled their medications more regularly after being enrolled to the study, which may be explained by a Hawthorne effect increasing motivation to refill prescriptions after visiting the clinic (Fig. 3). The effectiveness of the intervention was not
Use of M-DRAW tool in a primary care setting Diabetes was chosen as the disease state for the study because it affects nearly one fourth of the VA's patient population.24 In addition, routine assessment of HbA1c can signal change in disease control potentially achieved by the study intervention. Similar to a review article by Doggrell, our study finding did not support the clinical benefit of the
Fig. 4. Changes of hemoglobin A1c value. 5
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
taking medication’ (Q9). Previous studies have identified modifiable, common patient-related factors contributing to medication adherence issue such as having less knowledge about the disease or drug and incongruence in health belief contributing to lack of motivation to adhere to drug regimen.26,27 Similar to the existing body of evidence, our study findings suggest that targeted education based on identified modifiable factors would help clinicians guide the conversation in a patient-specific manner. When specific questions were identified as potential barriers to medication adherence, the pharmacist followed the Guide Strategy to provide a customized approach to solve the problem based on the participant's need. If a study participant was having difficulty keeping track of a medication, medication reconciliation was performed to review and create a medication list with a patient. When participants allowed, the pharmacist helped set up an alarm on their cell phone to remind them when to take their medication. A pillbox was provided at the clinic, and the pharmacist reviewed how to use the pillbox. Participants often were concerned about drug-food or drug-drug interaction, and the pharmacists used reliable resources such as the Micromedex and Natural Medicines Comprehensive Database to show the result and discuss in a patient-friendly language.28,29 Depending on the patient's engagement in the conversation, the intervention lasted approximately 15 min–45 min. These standardized interventions may have helped both with timely identification of problems contributing to medication adherence and implementation of practical strategies tailored to the individuals' identified barrier. Overall, the M-DRAW can serve as an adherence tool due to its three strengths: discriminant validity of the priming question, identification of barriers, and Guide Strategy to resolve the barriers.
Fig. 6. Usefulness of the intervention.
remarkable in either the clinical marker or refill history; however, participants identified fewer number of barriers contributing to medication adherence using the M-DRAW tool at follow-up compared to baseline (Fig. 2). It is possible that the 3-month follow-up may have been too short to see the benefit of the intervention clinically and objectively. It is also possible that participants who agreed to participate in the study recognized the importance of ‘medication adherence’. At the 3-month follow-up, participants were asked to rate the helpfulness and usefulness of the initial interaction. While both groups answered that meeting a pharmacist at the study enrollment was helpful in their care management, only those participants who received the intervention found it useful in changing their medication intake behavior.
Strength, limitation, and future research The strengths of the study itself included a low attrition rate (88/91 were followed up) and completeness of the clinical markers collected using the electronic health record. The user-friendly language of the tool has enabled participants to independently read and answer both the priming questions and the thirteen statements on the M-DRAW tool. This finding points to the applicability of the tool in future studies to screen for study participants who are medication nonadherent. The results of the study should be interpreted with caution due to some limitations. In spite of the low attrition rate, the small sample size precluded multivariate regression analysis. The duration of follow-up may have been too short to observe effectiveness of the intervention via a clinical marker and refill history. The nature of the study site, which
Use of Guide Strategy to provide targeted intervention On average, participants in Group N identified all items, except Q2 and Q6, as potential barriers contributing to adherence issues (Table 3). Items that showed a statistically significant reduction between the initial and follow-up screening included ‘feeling unsure about how/when to take medication’ (Q1), ‘food or drug interaction’ (Q5), ‘uncomfortable taking medication in public’ (Q8), and ‘life-long burden of Table 3 Changes in answers before and after in each group. Group
Group A (n = 63)
Group N (n = 25)
Items
PRE (%)
POST (%)
PRE (%)
POST (%)
Q1 - Do you feel unsure about how/when to take your medications?*+ Q2 - Do you have any difficulty getting your medications on time from the pharmacy? Q3 - Do you have difficulty keeping track of all your medication schedules throughout the day? (e.g., when to take each medication)* Q4 - Do your medications give you side effects that make you NOT want to take it?* Q5 - Do you worry about what foods or other medications might interact with your medication?* + Q6 - Do you feel that you can take more or less of your medication than the prescribed dose to fit your lifestyle? Q7 - Do you feel like you don't get any benefits from taking your medication?* Q8 - Do you feel uncomfortable about taking your medication while you are out with family and friends?* + Q9 - Do you consider it a burden that you have to take your medications for the rest of your life?* + Q10 - Do you have doubts about whether your health condition needs to be treated?* Q11 - Do you have doubts if taking your medication will improve your health condition in the long term?* Q12 - Do you feel that you are NOT receiving the best possible treatment available from your healthcare provider?*
6.5 17.7 9.7
4.9 10.0 4.9
40.0 36.0 28.0
30.0 25.0 20.0
4.8 19.4 21.0 14.5 16.1 32.3 6.5 11.3 8.1
7.3 4.9 14.6 9.8 7.3 26.8 4.9 17.1 12.2
44.0 48.0 28.0 40.0 48.0 68.0 40.0 56.0 28.0
15.0 25.0 30.0 25.0 30.0 55.0 15.0 40.0 10.0
*p < 0.05 of Fisher's Exact Test comparing response between Group A and Group N at baseline (PRE-response); +p < 0.05 of Fisher's Exact Test comparing response between Group A and Group N at baseline (POST-response); As described in Section 2.5, responses to M-DRAW tool items were dichotomized and recoded to count the number of barriers (i.e., each item answered ‘1 = never’ or ‘2 = rarely’ was recoded to ‘0’ and each item answered ‘3 = sometimes’ or ‘4 = often’ were recoded to ‘1’). 6
Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx
S. Lee, et al.
provides care for predominantly older white males, may have limited the generalizability of the tool. Additionally, the convenience of having a private room to speak with participants for 15–45 min may limit transferability of the study procedure in a busy community pharmacy setting. Future studies with a longer follow-up and multiple clinic settings may help further evaluate the applicability and generalizability of the M-DRAW tool.
https://doi.org/10.1186/1472-6963-14-219. 9. Dao N, Lee S, Hata M, Sarino Lord. Impact of appointment-based medication synchronization on proportion of days covered for chronic medications. Pharm (Basel, Switzerland). 2018;6(2):44. https://doi.org/10.3390/pharmacy6020044. 10. Lee S, Khare MM, Olson HR, Chen AMH, Law AV. The TEACH trial: tailored education to assist label comprehension and health literacy. Res Soc Adm Pharm. 2018;14(9):839–845. https://doi.org/10.1016/j.sapharm.2018.05.015. 11. Lau SR, Kriegbaum M. Medication non-adherence in the context of situated uncertainty: moving beyond simple, dichotomous approaches. Res Soc Adm Pharm. 2018;14(8):742–748. https://doi.org/10.1016/J.SAPHARM.2017.09.003. 12. Medication adherence | STEPS forward. https://www.stepsforward.org/modules/ medication-adherence, Accessed date: 10 September 2018. 13. Doucette WR, Farris KB, Youland KM, Newland BA, Egerton SJ, Barnes JM. Development of the drug adherence work-up (DRAW) tool. J Am Pharm Assoc. 2012;52(6):e199–e204. https://doi.org/10.1331/JAPhA.2012.12001. 14. Moon SJ, Lee W-Y, Hwang JS, Hong YP, Morisky DE. Accuracy of a screening tool for medication adherence: a systematic review and meta-analysis of the Morisky Medication Adherence Scale-8. PLoS One. 2017;12(11):e0187139https://doi.org/10. 1371/journal.pone.0187139. 15. Svarstad BL, Chewning BA, Sleath BL, Claesson C. The Brief Medication Questionnaire: a tool for screening patient adherence and barriers to adherence. Patient Educ Couns. 1999;37(2):113–124http://www.ncbi.nlm.nih.gov/pubmed/ 14528539, Accessed date: 10 September 2018. 16. Jaam M, Awaisu A, Mohamed Ibrahim MI, Kheir N. A holistic conceptual framework model to describe medication adherence in and guide interventions in diabetes mellitus. Res Soc Adm Pharm. 2018;14(4):391–397. https://doi.org/10.1016/J. SAPHARM.2017.05.003. 17. World Health Organization(WHO). WHO - ADHERENCE TO LONG-TERM THERAPIES: EVIDENCE FOR ACTION. WHO; 2015. 18. Müller S, Kohlmann T, Wilke T. Validation of the Adherence Barriers Questionnaire an instrument for identifying potential risk factors associated with medication-related non-adherence. BMC Health Serv Res. 2015;15:153. https://doi.org/10.1186/ s12913-015-0809-0. 19. Lee S, Bae YH, Worley M, Law A. Validating the modified drug adherence work-up (M-DRAW) tool to identify and address barriers to medication adherence. Pharm (Basel, Switzerland). 2017;5(3):52. https://doi.org/10.3390/pharmacy5030052. 20. Presley B, Groot W, Pavlova M. Pharmacy-led interventions to improve medication adherence among adults with diabetes: a systematic review and meta-analysis. Res Soc Adm Pharm. October 2018. https://doi.org/10.1016/J.SAPHARM.2018.09.021. 21. Fredericksen RJ, Gibbons L, Brown S, et al. Medication understanding among patients living with multiple chronic conditions: implications for patient-reported measures of adherence. Res Soc Adm Pharm. 2018;14(6):540–544. https://doi.org/ 10.1016/j.sapharm.2017.06.009. 22. National Center for Chronic Disease Prevention and Health Promotion. Calculating Proportion of Days Covered (PDC) for Antihypertensive and Antidiabetic Medications: An Evaluation Guide for Grantees. United States; 2015https://www.cdc.gov/dhdsp/docs/ med-adherence-evaluation-tool.pdf. 23. IBM Corp. IBM SPSS Statistics for Windows, Version 24. Armonk, NY. Released. IBM Corp IBM Corp - Google Scholar; 2015 2015. 24. Lehnert H, Wittchen H-U, Pittrow D, et al. Prevalence and pharmacotherapy of diabetes mellitus in primary care. Dtsch Medizinische Wochenschrift. 2005. https:// doi.org/10.1055/s-2005-863050. 25. Doggrell SA. A review of interventions ≥ 6 months by pharmacists on adherence to medicines in cardiovascular disease: characteristics of what works and what doesn't. Res Soc Adm Pharm. 2019;15(2):119–129. https://doi.org/10.1016/J.SAPHARM. 2018.04.003. 26. Gellad WF, Grenard JL, Marcum ZA. A systematic review of barriers to medication adherence in the elderly: looking beyond cost and regimen complexity. Am J Geriatr Pharmacother. 2011;9(1):11–23. https://doi.org/10.1016/j.amjopharm.2011.02. 004. 27. Roebuck MC, Liberman JN. Impact of pharmacy benefit design on prescription drug utilization: a fixed effects analysis of plan sponsor data. Health Serv Res. 2009;44(3):988–1009. https://doi.org/10.1111/j.1475-6773.2008.00943.x. 28. Chatfield AJ. Lexicomp online and Micromedex 2.0. J Med Libr Assoc. 2015. https:// doi.org/10.3163/1536-5050.103.2.016. 29. Tomasulo P. Natural medicines Database. J Consum Health Internet. 2004. https://doi. org/10.1080/15424065.2016.1225541.
Conclusion Results from this study support the reliability and validity of the MDRAW tool, and its usefulness in a clinical practice setting to identify and address barriers to patient medication adherence. The priming question showed good discriminant validity with limited sensitivity. A tailored education followed by the M-DRAW tool's Guide Strategy in patients with diabetes mellitus appears to reduce self-reported barriers to medication adherence. Integrating the M-DRAW tool in a primary clinic setting can provide a timely assessment of medication adherence issues and encourage meaningful patient-provider interactions that can impact patient outcomes. Acknowledgement The authors would like to acknowledge Drs. Bill Doucette, Jan Kevookjian, and Emmanuelle Schwartzman for their advice and assistance with the tool and motivational interviewing techniques. Appendix A. Supplementary data Supplementary data to this article can be found online at. References 1. Centers for Disease Control and Prevention. National Diabetes Statistics Report 2017; 2017https://doi.org/10.1177/1527154408322560 2017. 2. Krass I, Schieback P, Dhippayom T. Adherence to diabetes medication: a systematic review. Diabet Med. 2015;32(6):725–737. https://doi.org/10.1111/dme.12651. 3. Data & statistics | diabetes | CDC. https://www.cdc.gov/diabetes/data/index.html, Accessed date: 7 September 2018. 4. Statistics About Diabetes: American Diabetes Association®. http://www.diabetes. org/diabetes-basics/statistics/, Accessed date: 2 May 2019. 5. Krass I, Schieback P, Dhippayom T. Adherence to diabetes medication: a systematic review. Diabet Med. 2015. https://doi.org/10.1111/dme.12651. 6. Ali MK, McKeever Bullard K, Imperatore G, Barker L, Gregg EW. Centers for Disease Control and Prevention (CDC). Characteristics associated with poor glycemic control among adults with self-reported diagnosed diabetes–National Health and Nutrition Examination Survey, United States, 2007-2010. MMWR Suppl. 2012;61(2):32–37http://www.ncbi.nlm.nih.gov/pubmed/22695461, Accessed date: 7 September 2018. 7. Viswanathan M, Golin CE, Jones CD, et al. Interventions to improve adherence to self-administered medications for chronic diseases in the United States. Ann Intern Med. 2012;157(11):785. https://doi.org/10.7326/0003-4819-157-11-20121204000538. 8. Kooy MJ, van Geffen EC, Heerdink ER, van Dijk L, Bouvy ML. Effects of a TELephone Counselling Intervention by Pharmacist (TelCIP) on medication adherence, patient beliefs and satisfaction with information for patients starting treatment: study protocol for a cluster randomized controlled trial. BMC Health Serv Res. 2014;14(1):219.
7