Abstract
Background
Deprescribing, intentional medication discontinuation or dose reduction, can reduce potentially inappropriate medication use and medication-related harms. Engaging patients in deprescribing discussions may increase likelihood of deprescribing and promote shared decision-making.
Objective
To examine the impact of patient-directed educational brochures on patient engagement and deprescribing discussions with primary care providers (PCPs).
Design
We mailed medication-specific brochures 2 weeks prior to each patient’s PCP appointment (4/12/2021–10/7/2022), followed by a mailed survey 2 weeks after scheduled PCP visits.
Participants
Patients from three Veterans Affairs facilities with scheduled PCP appointments eligible for one of three medication-based cohorts (proton pump inhibitor, gabapentin, diabetes-hypoglycemia risk).
Main Measures
Our primary outcome was patient-reported deprescribing discussions with their PCP (yes/no). Descriptive statistics characterized engagement with and reactions to the brochure. Multivariable logistic regression models determined associations of patient characteristics, attitudes, and brochure-engagement with reported deprescribing discussions.
Key Results
Adjusting only for patient characteristics, discussions were less likely if respondents were Black (vs. White: OR 0.47, 95% CI 0.29–0.78) and more likely with higher education level (e.g., advanced degree vs. high school or less: OR 2.39, 95% CI 1.53–3.73), and adequate health literacy (OR 1.84, 95% CI 1.16–2.92). After further adjusting for general deprescribing attitudes and brochure engagement, discussions were more likely if respondents completed brochure activities (vs. did not read brochures: OR 2.23, 95% CI 1.39–3.59), contacted their PCPs prior to their visits (OR 2.47, 95% CI 1.34, 4.58), or discussed the brochure with family/friends (OR 1.72, 95% CI 1.22–2.41) or a healthcare provider (OR 3.18, 95% CI 2.08–4.85).
Conclusions
Patient characteristics and brochure engagement were associated with deprescribing discussions. Patient-centered deprescribing brochures can foster patient engagement and inclusion of patient perspectives into deprescribing decisions. Future studies should explore implementation strategies that promote greater deprescribing reach and adoption among patients with lower health literacy.
Supplementary Information
The online version contains supplementary material available at 10.1007/s11606-024-09346-w.
KEY WORDS: deprescribing, shared-decision making, patient education, polypharmacy, patient engagement
Background
Patients with chronic conditions are often prescribed multiple medications, resulting in polypharmacy. Polypharmacy, typically defined as five or more medications,1 is most prevalent in older adults.2 Harms associated with polypharmacy include adverse drug events, drug-drug interactions, falls, hospital admissions, and mortality.3 Deprescribing, the intentional discontinuation or dose reduction of medications where risks outweigh benefits, is a practice that reduces potentially inappropriate medication use and associated harms.4
Successful deprescribing increases with both patient and clinician engagement. Many patients express interest in taking fewer medications,5 but patients and clinicians may be reluctant to discuss deprescribing.6,7 Time constraints can make deprescribing discussions logistically challenging for clinicians. Both patients6 and clinicians8 may fear that deprescribing will lead to an increase in the symptoms for which the medication was prescribed. Some patients and clinicians choose the status quo in the absence of clear reason to change care.9 Finally, clinicians and patients must weigh the relative importance of managing a specific condition (e.g., acid reflux) with other health conditions and medications.10
Deprescribing discussions should be based upon principles of shared decision-making (SDM), with open communication concerning patients’ preferences and health goals.11,12 SDM models highlight the value of patient-provider discussion and patient engagement for effective care management.13 Several key elements increase patient engagement in SDM; among these is sufficient patient knowledge of the health condition.14 Thus, materials and decision-making aids that educate patients on their health condition and help them identify and communicate their overall health goals, such as patient education brochures, can effectively promote SDM in clinical care and deprescribing.15–17 High-quality deprescribing educational materials teach patients about medications by presenting a balanced perspective (benefits and harms of medications) and promote patient-clinician communication.18
In the current study, we tested the effectiveness of a medication-specific educational brochure on increasing patient engagement and deprescribing discussions. We selected three medication-based cohorts because of potential low benefit or high risk: proton pump inhibitors (PPIs), gabapentin, and diabetes medications with risk for hypoglycemia. PPIs treat gastric conditions but can be low-benefit due to lack of guideline concordance (e.g., longer than recommended use). Gabapentin is commonly used off-label19 and at high doses,20 conferring risk for falls.21 Insulin and sulfonylureas can cause hypoglycemia (i.e., low blood sugar), which may lead to seizures and loss of consciousness.22 We examined how patients engaged with the brochures and if brochure engagement was associated with a deprescribing discussion at their next primary care appointment.
METHODS
Method Overview
During the 18-month intervention period (April 2021–October 2022), patients with upcoming primary care appointments for one of our three medication-based cohorts were sent brochures. After the appointment, patients were mailed a survey about their experiences. The study was deemed exempt by the Institutional Review Board, and oversight was provided by the Research and Development Committee at Veterans Affairs (VA) Boston Healthcare System.
Brochure Development
The Canadian Deprescribing Network originally developed Eliminating Medications Through Patient Ownership of End Results (EMPOWER) brochures.17,23–25 EMPOWER brochures were developed using theories of patient activation, adult learning, and cognitive dissonance,23 and they facilitate active patient communication such as asking questions, communicating assertively, and expressing concerns.26 To better reflect the Veteran population, we adapted images and vignettes for two existing EMPOWER brochures (PPIs and sulfonylureas), and the sulfonylurea brochure was further expanded to include insulin to align with ongoing VA efforts to reduce hypoglycemia. We created a gabapentin brochure since one had not been developed at the time of this study. All brochures were modified to include space for Veterans to report how their medication practices helped them reach their personal health goals.
Survey Development
The survey was developed to focus on patient factors, patient-clinician communication, and downstream behaviors. First, we adapted an interview guide from our previous study using EMPOWER brochures,23 replacing open-ended questions with closed-ended items. Next, we conducted cognitive interviews with eight Veterans using a “think aloud” protocol in which we asked them to explain their understanding of and the thought processes that occur when responding to survey items. We used strategic probes as necessary to elicit understanding of key terms and concepts and then iteratively revised the survey.
Veterans were asked about their demographic and health characteristics (e.g., age, living arrangement, overall health) and three items to assess health literacy (coded as adequate: yes/no). To understand general attitudes toward their health care and deprescribing, the survey included two sub-scales from the Altarum Consumer Engagement (ACE) measure27 (navigation and ownership), along with two scales from the Short-Form Patient Perceptions of Deprescribing (SF-PPoD) instrument (motivation for deprescribing and primary care provider relationship).28 Other survey items queried brochure engagement and actions taken after brochure receipt, as well as concerns triggered by the brochure, relative importance to discuss the medicine at their primary care visit, and preference prior to their visit for whether they wanted to stop the medication. Finally, the survey asked whether respondents discussed the medication of interest at their primary care visit, and if so, who initiated the discussion, and if the brochure specifically was discussed. (See Appendix for full survey instrument.)
Study Setting and Population
We identified study subjects with upcoming primary care appointments at three VA Medical Centers and their associated community-based outpatient clinics. Sites were selected based on geographic variability and projected ability to yield adequate sample sizes. We included all primary care providers (PCPs, n = 158) but excluded resident physicians given anticipated turnover during the study. We included all patients with a visit with one of the 158 PCPs during our 18-month intervention period. The intervention period was divided into three sequential 6-month waves so that a given PCP’s patients were only eligible to be enrolled in one medication group per wave, and patients were enrolled only once regardless of subsequent visits. The visit needed to be scheduled at least 2 weeks in advance to provide sufficient time to receive the study intervention by postal mail.
To be enrolled, Veterans had to meet eligibility criteria for one of three medication-based cohorts. For all medications of interest, we required a minimum of 90 consecutive days of use in the prior year and an active prescription at the time of subject identification (i.e., 14–20 days prior to a scheduled primary care visit). The PPI cohort included patients with prescriptions for any PPI at any dose; we excluded patients for whom PPI continuation could be potentially appropriate (e.g., diagnosis of Barrett’s esophagus, or prescribed chronic glucocorticoids or nonsteroidal anti-inflammatory drugs (NSAIDs)). The Gabapentin cohort included patients prescribed a total daily dose of gabapentin > 1800 mg; we excluded patients with approved or commonly used indications for gabapentin (e.g., diagnoses of trigeminal neuralgia, neuropathic pain, seizure disorder). The Diabetes-Hypoglycemia Risk (DM-HR) cohort included patients with diabetes prescribed insulin or sulfonylureas, or both; most recent HbA1c < 7%; and one or more of the following criteria: (1) age 65 years or older, (2) renal insufficiency, defined as creatinine > 2 mg/dL, or (3) cognitive impairment, defined by a diagnosis of cognitive impairment and/or prescription for an acetylcholinesterase inhibitor (e.g., donepezil). There were 3206 patients who met the inclusion criteria (PPI: N = 2624; gabapentin: N = 121; DM-HR: N = 461).
Data Sources
All primary care appointment and medication data were obtained from the VA Corporate Data Warehouse (CDW). Additional variables obtained from CDW included patient demographics; specified comorbidities based on International Classification of Diseases (Tenth Revision) Clinical Modification (ICD-10-CM) diagnosis codes; medications based on dispensing date and days supplied; and utilization in the year prior that may have been an indication or contraindication for the target medication. All other data were obtained from self-report survey responses. Data obtained from CDW were used only to compare survey respondents to non-respondents; self-reported data were used for all other analyses.
Study Procedures
Eligible patients were mailed a cover letter and the applicable EMPOWER brochure. The cover letter explained that the mailing was to gather information about their medication, and that they would subsequently receive a survey asking their opinion about the brochure. One week after the scheduled PCP visit date, patients were mailed a medication-specific survey with a $5 incentive to encourage survey completion and an opt-out card. We used a modified Dillman approach,29 whereby the brochure was the first point of contact, and the survey was sent subsequently. If the survey or opt-out card was not returned after 2 weeks, we sent a reminder letter, and then after another week of non-response, we sent a second survey copy as a final attempt. Patients who returned opt-out cards and for whom we received notifications of death were removed from the administration cycle. All responses were entered into a database by two trained research staff (JP, MS), and 10% of responses were checked to ensure accuracy (AML).
Analysis
We ran descriptive analyses to characterize the survey sample and their engagement and reactions to the brochure. For all inferential analyses, our primary outcome variable was patient-reported occurrence of a deprescribing discussion with their PCP (yes/no). We examined bivariate associations between deprescribing discussions and patient characteristics, general attitudes toward deprescribing, and brochure engagement and reactions, using chi-square and t-tests as appropriate. We then used multivariable binary logistic regression analyses to assess these associations when controlling for other variables. We used a stepped approach to multivariable model building: Model 1 included only patient characteristics as predictor variables, model 2 added general attitudes toward deprescribing, and model 3 added items specifically related to brochure engagement and reactions. This stepped approach enabled us to assess potential contributions of general attitudes about deprescribing to having a deprescribing discussion before and after controlling for more specific behaviors and reactions related to the brochure. All analyses were conducted with SAS EG, version 8.3.
RESULTS
Patient Characteristics
There were 1382 survey respondents across the three medication cohorts (PPI: n = 1116/2624 (43%), gabapentin: n = 40/121 (33%), DM-HR: n = 226/461 (49%)), reflecting an overall response rate of 43% (1382/3206). All respondent characteristics are seen in Table 1. As seen in the first column of Table 1, survey respondents were primarily White (75%), male (95%), aged 66–75 years (45%), lived with at least one other person (63%), and had adequate health literacy (89%). Respondents were more likely than non-respondents to be White, married, and older in age; they were less likely to be Black, Hispanic, or female; respondents also had lower 30-day all cause readmission risk (see Table 2).
Table 1.
Respondent Characteristics, Deprescribing Attitudes, and Brochure Engagement (N = 1382)
| Characteristic | N (%) |
Reported discussion N (%) |
Did not report discussion N (%) |
P | |
| Medication cohort | < 0.001* | ||||
| PPI | 1116 (80.7) | 353 (31.6) | 763 (68.4) | ||
| Gabapentin | 40 (2.9) | 12 (30.0) | 28 (70) | ||
| DM—HR | 226 (16.4) | 37 (16.4) | 189 (83.6) | ||
| Race/ethnicity (missing n = 64) | 0.001* | ||||
| White (non-Hispanic) | 986 (74.8) | 314 (31.9) | 672 (68.2) | ||
| Black (non-Hispanic) | 129 (9.8) | 21 (16.3) | 108 (83.7) | ||
| Hispanic | 120 (9.1) | 29 (24.2) | 91 (75.8) | ||
| Other/Multiracial | 83 (6.3) | 22 (26.5) | 61 (73.5) | ||
| Age (missing n = 45) | 0.5 | ||||
| ≤ 65 | 236 (17.7) | 78 (33.1) | 158 (67.0) | ||
| 66–75 | 606 (45.3) | 168 (27.7) | 438 (72.3) | ||
| 76–85 | 369 (27.6) | 110 (29.8) | 259 (70.2) | ||
| ≥ 86 | 126 (9.4) | 37 (29.4) | 89 (70.6) | ||
| Educational level (missing n = 24) | < 0.001* | ||||
| High school or less | 426 (31.4) | 84 (19.7) | 342 (80.3) | ||
| Some college or vocational school | 629 (46.3) | 201 (32.0) | 428 (68.0) | ||
| College degree | 181 (13.3) | 62 (34.3) | 119 (65.8) | ||
| Graduate/advanced degree | 122 (9.0) | 50 (41.0) | 72 (59.0) | ||
| Gender (missing n = 52) | 0.201 | ||||
| Male | 1261 (94.8) | 366 (29.0) | 895 (71.0) | ||
| Female | 69 (5.2) | 25 (36.2) | 44 (63.8) | ||
| Living arrangement (missing n = 60) | 0.682 | ||||
| Live alone | 431 (32.6) | 130 (30.2) | 301 (69.8) | ||
| Lives with ≥ 1 person | 891 (67.4) | 259 (29.1) | 632 (70.9) | ||
| Overall health (missing n = 49) | 0.008* | ||||
| Excellent/very good/good | 837 (62.8) | 268 (32.0) | 569 (68.0) | ||
| Fair/poor | 496 (37.2) | 125 (25.2) | 371 (74.8) | ||
| Adequate health literacy (missing n = 18) | < 0.001* | ||||
| Yes | 1214 (89.0) | 371 (30.6) | 843 (69.4) | ||
| No | 150 (11.0) | 25 (16.7) | 125 (83.3) | ||
| Brochure engagement | < 0.001* | ||||
| None/missing | 397 (28.7) | 44 (11.1) | 353 (88.9) | ||
| Read only | 399 (28.9) | 94 (23.6) | 305 (76.4) | ||
| Engage with activities | 461 (33.4) | 202 (43.8) | 259 (56.2) | ||
| Contact PCP | 125 (9.0) | 62 (49.6) | 63 (50.4) | ||
| Discuss brochure with family/friends? | < 0.001* | ||||
| Yes | 363 (26.3) | 170 (46.8) | 193 (53.2) | ||
| No/missing | 1019 (73.7) | 232 (22.8) | 787 (77.2) | ||
| Discuss brochure with healthcare provider other than PCP? | < 0.001* | ||||
| Yes | 175 (12.7) | 102 (58.3) | 73 (41.7) | ||
| No/missing | 1207 (87.3) | 300 (24.9) | 907 (75.1) | ||
| Concern about medication | < 0.001* | ||||
| Not at all/missing | 641 (46.4) | 84 (13.1) | 557 (86.9) | ||
| A little/somewhat/a lot | 741 (53.6) | 318 (42.9) | 423 (57.1) | ||
| Importance of discussing deprescribing | < 0.001* | ||||
| Low/not at all/missing | 761 (55.1) | 80 (10.5) | 681 (89.5) | ||
| Most/very/moderately | 621 (44.9) | 322 (51.9) | 299 (48.2) | ||
| Intention to deprescribe | < 0.001* | ||||
| Yes | 189 (13.7) | 116 (61.4) | 73 (38.6) | ||
| No/missing | 1193 (86.3) | 286 (24.0) | 907 (76.0) | ||
| General attitudes toward deprescribing | Mean (SD) | ||||
| Motivation to deprescribe (missing n = 23) | 2.82 (0.82) | 2.91 (0.83) | 2.79 (0.82) | 0.010* | |
| PCP relationship (missing n = 21) | 4.13 (0.78) | 4.23 (0.72) | 4.08 (0.80) | 0.002* | |
| Navigation (missing n = 46) | 3.87 (0.59) | 3.94 (0.54) | 3.85 (0.61) | 0.009* | |
| Ownership (missing n = 46) | 3.94 (0.66) | 4.04 (0.58) | 3.90 (0.69) | < 0.001* | |
Abbreviations: PPI, proton pump inhibitor; DM-HR, diabetes-hypoglycemia risk; PCP, primary care provider; SD, standard deviation
*Statistically significant
Table 2.
Characteristics of Survey Respondents compared to Non-respondents
|
Respondents (N = 1382) n (%) |
Non-Respondents (N = 1824) n (%) |
p | |
| Medication cohort | 0.003* | ||
| PPI | 1116 (80.8) | 1508 (82.7) | |
| Gabapentin | 40 (2.9) | 81 (4.4) | |
| DM-HR | 226 (16.4) | 235 (12.9) | |
| Race (missing n = 355) | < 0 .001* | ||
| White | 1039 (83.4) | 1248 (77.8) | |
| Black | 162 (13.0) | 297 (18.5) | |
| Other/multiracial | 45 (3.6) | 60 (3.7) | |
| Hispanic (missing n = 227) | 82 (6.4) | 176 (10.4) | < 0.001* |
|
Service-connected disability ≥ 50% (missing n = 1171) |
627 (71.9) | 878 (75.5) | 0.068 |
| Female | 67 (4.9) | 128 (7.0) | 0.011* |
| Married (missing n = 20) | 734 (53.5) | 799 (44.1) | < 0.0001* |
| Mean (SD) | Mean (SD) | ||
| Age | 73.5 (9.7) | 69.4 (13.2) | < 0.001* |
| Risk of in-hospital mortality (missing n = 23) | − 1.7 (7.5) | − 1.6 (9.0) | 0.807 |
| Risk of 30-day, all-cause readmission (missing n = 23)† | 3.3 (5.2) | 3.9 (6.1) | 0.002* |
PPI, proton pump inhibitor; DM-HR, diabetes-hypoglycemia risk; PCP, primary care provider
*Statistically significant
†Based upon the Elixhauser comorbidity index38
Brochure Engagement and Reactions
Most respondents (71%) reported engaging with the brochure in some way; 29% read the brochure, an additional 33% read and completed brochure activities, and another 9% contacted their PCP about the brochure ahead of their appointment. Further, 26% reported discussing the brochure with family members or friends and 13% discussed it with a healthcare provider other than their PCP. Slightly more than half (53%) indicated that the brochure increased their concern about their medications. Nearly half (45%) felt that discussing deprescribing was moderately/very/most important for their upcoming PCP visit, and 14% indicated that they intended to deprescribe their medication at their upcoming visit (see Table 1, first column).
Deprescribing Discussions
Nearly one in three respondents indicated having a deprescribing discussion with their PCP (29%) at their brochure follow-up appointment. Of those who had a deprescribing discussion, the majority reported they initiated the discussion (78%), and more than half indicated that they explicitly discussed the brochure (57%).
Predicting Deprescribing Discussions: Bivariate Associations
Results from all bivariate analyses are reported in Table 1.
Patient Characteristics
Patients were significantly more likely to report deprescribing discussions if they were White, and had greater education, better health, and adequate health literacy. When comparing between medication cohorts, patients in the DM-HR cohort were significantly less likely to have deprescribing discussions compared to those in the PPI and gabapentin cohorts. All p values < 0.01. In post hoc analyses, DM-HR patients prescribed sulfonylureas only were more likely to report a deprescribing discussion compared to DM-HR patients prescribed insulin only (22% vs. 11%, p = 0.03).
General Attitudes About Deprescribing
Compared to patients who did not have a deprescribing discussion, patients with deprescribing discussions reported significantly higher motivation to deprescribe, better PCP relationships, greater ability to navigate the medical care system, and greater ownership over their medical treatment. All p values < 0.01.
Brochure Engagement and Reactions
Patients were significantly more likely to have deprescribing discussions if they read and/or engaged more with the brochure and if they discussed the brochure with family/friends or a healthcare provider. They also were more likely to have deprescribing discussions if they expressed that the brochure made them concerned about their medication, if discussing deprescribing was moderately/very/most important for their upcoming PCP visit, and if they intended to deprescribe their medication at their upcoming visit. All p values < 0.001.
Predicting Deprescribing Discussions: Multivariable Models
Coefficient estimates from multivariable models are in Table 3.
Table 3.
Multivariable Models Predicting Report of a Deprescribing Discussion
| Model 1: Patient characteristics | Model 2: Patient characteristics and general attitudes |
Model 3: Patient characteristics, general attitudes, and brochure engagement |
|||||
|---|---|---|---|---|---|---|---|
| Percent concordant | 63.8% | 66.5% | 86.7% | ||||
| Wald chi-square test |
Χ2(14) = 66.34, p < 0.0001 |
Χ2(18) = 84.94, p < 0.0001 |
Χ2(26) = 335.29, p < 0.0001 |
||||
| OR | 95% CI | OR | 95% CI | OR | 95% CI | ||
| Medication cohort (ref = PPI) | – | – | – | – | – | – | |
| DM – HR | 0.47 | [0.32, 0.69] | 0.46 | [0.31, 0.67] | 0.26 | [0.16, 0.40] | |
| Gabapentin | 0.87 | [0.43, 1.76] | 0.83 | [0.41, 1.70] | 0.61 | [0.24, 1.55] | |
| Race/ethnicity (ref = White, non-Hispanic) | – | – | – | – | – | – | |
| Black, non-Hispanic | 0.47 | [0.29, 0.78] | 0.41 | [0.25, 0.69] | 0.31 | [0.17, 0.57] | |
| Hispanic | 0.80 | [0.51, 1.26] | 0.78 | [0.49, 1.23] | 0.65 | [0.38, 1.13] | |
| other/multiracial | 0.73 | [0.43, 1.24] | 0.66 | [0.39, 1.13] | 0.45 | [0.23, 0.86] | |
| Age (ref = ≤ 65 years) | – | – | – | – | – | – | |
| 66–75 | 0.82 | [0.58, 1.16] | 0.82 | [0.58, 1.16] | 0.76 | [0.49, 1.18] | |
| 76–85 | 0.92 | [0.63, 1.34] | 0.92 | [0.63, 1.34] | 0.84 | [0.52, 1.35] | |
| ≥ 86 | 0.98 | [0.59, 1.62] | 1.00 | [0.60, 1.66] | 0.84 | [0.44, 1.58] | |
| Education (ref = high school or less) | – | – | – | – | – | – | |
| Some college or vocational school | 1.74 | [1.29, 2.35] | 1.72 | [1.27, 2.33] | 1.88 | [1.29, 2.73] | |
| College degree | 1.95 | [1.30, 2.91] | 1.90 | [1.26, 2.86] | 2.11 | [1.27, 3.52] | |
| Graduate/advanced degree | 2.39 | [1.53, 3.73] | 2.27 | [1.45, 3.57] | 2.40 | [1.35, 4.27] | |
| Female (ref = male) | 1.25 | [0.73, 2.14] | 1.24 | [0.72, 2.14] | 1.04 | [0.53, 2.03] | |
|
Good/very good/excellent health (ref = fair/poor) |
1.16 | [0.89, 1.50] | 1.19 | [0.91, 1.55] | 0.98 | [0.70, 1.36] | |
|
Adequate health literacy (ref = inadequate) |
1.84 | [1.16, 2.92] | 1.75 | [1.09, 2.80] | 1.50 | [0.85, 2.64] | |
| Motivation to deprescribe | 1.32 | [1.14, 1.54] | 0.93 | [0.76, 1.13] | |||
| PCP relationship | 1.28 | [1.06, 1.54] | 1.20 | [0.96, 1.51] | |||
| Navigation | 0.94 | [0.71, 1.25] | 0.78 | [0.55, 1.10] | |||
| Ownership | 1.18 | [0.93, 1.50] | 1.27 | [0.94, 1.72] | |||
|
Brochure engagement (ref = none/missing) |
– | – | |||||
| Read only | 1.52 | [0.94, 2.45] | |||||
| Read and engage activities | 2.23 | [1.39, 3.59] | |||||
| Contact PCP | 2.47 | [1.34, 4.58] | |||||
|
Discussed with family/friends (ref = no/missing) |
1.72 | [1.22, 2.41] | |||||
|
Discussed with HCP (ref = no/missing) |
3.18 | [2.08, 4.85] | |||||
|
Elevated medication concern (ref = not at all/missing) |
1.50 | [1.03, 2.18] | |||||
|
Discussion is important (ref = low/not at all/missing) |
9.11 | [6.52, 12.73] | |||||
|
Intention to deprescribe (ref = no/missing) |
1.94 | [1.28, 2.94] | |||||
Ref, reference; PPI, proton pump inhibitor, DM-HR, diabetes-hypoglycemia risk; PCP, primary care provider
Patient Characteristics (Model 1)
Findings from multivariable Model 1 were consistent with bivariate results; significant patient characteristics included being in the DM-HR cohort, Black race, and higher level of education. These associations remained significant in the full model (model 3). While adequate health literacy was associated with a higher likelihood of having discussions in model 1, this association was no longer significant in the full model (model 3). Unlike bivariate analyses, patient health status was not a significant predictor of deprescribing discussions when controlling for other patient characteristics.
General Attitudes About Deprescribing (Model 2)
Like bivariate results, findings from multivariable model 2, controlling for patient characteristics and the four attitudinal scales, showed increased deprescribing discussions if participants reported greater motivation to deprescribe or better PCP relationships. However, these associations were no longer significant when further controlling for brochure engagement and reactions (model 3). Unlike bivariate analyses, there was no significant association between the two ACE subscales (navigation and ownership) and having deprescribing discussions in models 2 and 3.
Brochure Engagement and Reactions (Model 3)
Similar to bivariate analyses, when controlling for patient characteristics and general deprescribing attitudes, all measures of brochure engagement and reactions remained significant predictors of having deprescribing discussions.
DISCUSSION
Distributing medication-specific EMPOWER brochures promotes patient engagement and discussion between patients and their care teams. The likelihood of having a deprescribing discussion varied by patient characteristics, such as education level, race, and medication cohort. Consistent with other studies, patients with greater education were more likely to have deprescribing discussions with their provider.30,31 This suggests that a higher educational level can support engagement with clinicians and health care SDM by providing knowledge and empowerment.
Survey respondents who identified as non-Hispanic Black were less likely to have deprescribing discussions compared to non-Hispanic White respondents. While literature was limited,32 there may be several explanations for this finding. Race is a social construct and may act as a proxy for effects of racism from social, environmental, and structural factors (e.g., differential access to education).33 Educational brochures may not be culturally tailored and mitigation of deprescribing disparities may require use of other strategies, such as storytelling, family member involvement, and inclusion of non-Hispanic Black people in designing patient-directed materials.34 Future efforts should harness community engagement methods to address disparities.
Veterans in the DM-HR cohort were less likely to have deprescribing discussions compared to those in the PPI and gabapentin cohorts. This may be related to the specific DM-HR medications, as patients may perceive insulin as more appropriate to continue than sulfonylureas.35 Our study combined patients taking insulin and sulfonylureas, possibly precluding our ability to identify an impact for those taking sulfonylureas. However, in post hoc bivariate analyses, patients were less likely to report deprescribing discussions if prescribed insulin only compared to sulfonylurea only. Efforts to address hypoglycemia risk in patients prescribed insulin may need different or additional components than the EMPOWER brochures.
Patients’ general attitudes toward deprescribing were associated with the likelihood of having a deprescribing discussion, even when controlling for patient characteristics. However, general attitudes were no longer significant once controlling for brochure-related behaviors. Although many patients often have favorable attitudes toward deprescribing in general,36 such beliefs do not consistently result in deprescribing actions, highlighting the importance of tailored, medication-specific interventions.
By sending medication-specific brochures directly to patients, we intended to promote discussions that fostered SDM and patient-centered care. The odds of having a discussion with their care team were more than nine times higher among Veterans who indicated that having this discussion was important after reading the brochure compared to those who did not indicate discussions were important. Notably, 78% of those who had a deprescribing discussion reported initiating that discussion. This is especially relevant because it suggests that patients raise the topic of deprescribing during healthcare visits, potentially alleviating some provider-level deprescribing barriers (e.g., perceptions that patients are not interested in deprescribing).8 Regardless of brochure receipt or other intervention, the obligation still remains for clinicians to address deprescribing with their patients. Interventions that are multi-faceted and directly address provider-level barriers may augment the effect of patient-centered approaches.
Limitations
Our sample included only Veteran users of the VA who were predominately male; while reflective of the VA population,37 these findings may not generalize to non-VA settings or to Veterans who do not utilize VA services. Future studies in other settings are warranted. Patient characteristics differed between respondents and non-respondents, so results may not be generalizable to all individuals taking potentially inappropriate medication. The gabapentin cohort was small, making it difficult to compare this medication group to the other two groups, yet we were still able to identify factors associated with deprescribing discussions. Because we mailed the brochures, we were unable to determine at which point patients reviewed the brochure information (e.g., prior to their PCP appointment or upon survey completion). Finally, there may have been inaccurate recall of events that occurred during the visit; however, patient perspectives of encounters likely contribute to their subsequent health behaviors.
CONCLUSION
Medication-specific brochures are an effective strategy to promote SDM for medication deprescribing. Decisions should be individualized and consider each patient’s health conditions, values, and goals. Patient educational brochures are a low-cost mechanism to promote deprescribing discussions between patients and their primary care teams. Future research should explore different strategies to promote enhance greater deprescribing reach and adoption among patients with lower health literacy and reduce race-based deprescribing disparities, such as by including community participatory methods when designing interventions.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary file1 (200 KB)
Author Contributions:
We would like to thank Lara Lobrutto and Risette Z. Maclaren for their important administrative contributions to this work.
Funding
This work was supported by the Department of Veterans Affairs, Veteran Health Administration, Health Service Research and Development (IIR 18–228; 1101HX002798-01A1).
Data Availability
The datasets analyzed during the current study are available from the last author on reasonable request.
Declarations:
Conflict of Interest:
Dr. Miller reported receiving grants from US Department of Veterans Affairs (VA) during the conduct of the study. Dr. Bokhour reported receiving grants from VA Health Services Research and development during the conduct of this study. Dr. Linsky reported receiving personal fees from the National Institute on Aging for presenting at the US Deprescribing Research Network Annual Meeting, and from the Agency for Healthcare Research and Quality of authoring a chapter on deprescribing for Making Healthcare Sager IV outside the submitted work. There are no other disclosures reported.
Footnotes
Prior Presentations
This work was presented at the Society of General Internal Medicine Annual Meeting on May 17, 2024.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Masnoon N, Shakib S, Kalisch-Ellett L, Caughey GE. What is polypharmacy? A systematic review of definitions. BMC Geriatrics. 2017;17:1-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Khezrian M, McNeil CJ, Murray AD, Myint PK. An overview of prevalence, determinants and health outcomes of polypharmacy. Therapeut Adv Drug Saf. 2020;11:2042098620933741. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Maher RL, Hanlon J, Hajjar ER. Clinical consequences of polypharmacy in elderly. Expert Opin Drug Saf. 2014;13(1):57-65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Scott IA, Hilmer SN, Reeve E, et al. Reducing inappropriate polypharmacy: the process of deprescribing. JAMA Int Med. 2015;175(5):827-834. [DOI] [PubMed] [Google Scholar]
- 5.Moen J, Bohm A, Tillenius T, Antonov K, Nilsson JLG, Ring L. “I don’t know how many of these [medicines] are necessary..”—A focus group study among elderly users of multiple medicines. Patient Educ Couns. 2009;74(2):135–141. [DOI] [PubMed]
- 6.Linsky A, Simon SR, Bokhour B. Patient perceptions of proactive medication discontinuation. Patient Educ Couns. 2015;98(2):220-225. [DOI] [PubMed] [Google Scholar]
- 7.Linsky A, Zimmerman KM. Provider and system-level barriers to deprescribing: interconnected problems and solutions. Public Policy Aging Rep. 2018;28(4):129-133. [Google Scholar]
- 8.Linsky A, Simon SR, Marcello TB, Bokhour B. Clinical provider perceptions of proactive medication discontinuation. Am J Manag Care. 2015;21(4):277-283. [PubMed] [Google Scholar]
- 9.Hung A, Sloan CE, Boyd C, Bayliss EA, Hastings SN, Maciejewski ML. Deprescribing medications: Do out-of-pocket costs have a role? J Am Geriatr Soc. 2022;70(11):3334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Le Bosquet K, Barnett N, Minshull J. Deprescribing: Practical ways to support person-centred, evidence-based deprescribing. Pharmacy. 2019;7(3):129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Alrawiai S. Deprescribing, shared decision-making, and older people: perspectives in primary care. J Pharm Policy Pract. 2023;16(1):153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dang S. Shared Decision Making-The Pinnacle of Patient-Centered Care. J Indian Acad Geriatr. 2018;14.
- 13.Makoul G, Clayman ML. An integrative model of shared decision making in medical encounters. Patient Educ Couns. 2006;60(3):301-312. [DOI] [PubMed] [Google Scholar]
- 14.Fraenkel L, McGraw S. What are the essential elements to enable patient participation in medical decision making? J Gen Int Med. 2007;22:614-619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Jansen J, Naganathan V, Carter SM, et al. Too much medicine in older people? Deprescribing through shared decision making. Bmj. 2016;353. [DOI] [PubMed]
- 16.Reeve E, Shakib S, Hendrix I, Roberts MS, Wiese MD. Review of deprescribing processes and development of an evidence‐based, patient‐centred deprescribing process. British J Clin Pharmacol. 2014;78(4):738-747. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Martin P, Tamblyn R, Ahmed S, Tannenbaum C. An educational intervention to reduce the use of potentially inappropriate medications among older adults (EMPOWER study): protocol for a cluster randomized trial. Trials. 2013;14:1-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Fajardo MA, Weir KR, Bonner C, Gnjidic D, Jansen J. Availability and readability of patient education materials for deprescribing: an environmental scan. British J Clin Pharmacol. 2019;85(7):1396-1406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Mack A. Examination of the evidence for off-label use of gabapentin. J Manag Care Pharm. 2003;9(6):559-568. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Yan PZ, Butler PM, Kurowski D, Perloff MD. Beyond neuropathic pain: gabapentin use in cancer pain and perioperative pain. Clin J Pain. 2014;30(7):613-629. [DOI] [PubMed] [Google Scholar]
- 21.Smith BH, Higgins C, Baldacchino A, Kidd B, Bannister J. Substance misuse of gabapentin. Br J Gen Pract. 2012;62(601):406-407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hypoglycemia A. Defining and reporting hypoglycemia in diabetes. Diabetes Care. 2005;28(5):1245. [DOI] [PubMed] [Google Scholar]
- 23.Tannenbaum C, Martin P, Tamblyn R, Benedetti A, Ahmed S. Reduction of inappropriate benzodiazepine prescriptions among older adults through direct patient education: the EMPOWER cluster randomized trial. JAMA Int Med. 2014;174(6):890-898. [DOI] [PubMed] [Google Scholar]
- 24.Martin P, Tamblyn R, Ahmed S, Tannenbaum C. A drug education tool developed for older adults changes knowledge, beliefs and risk perceptions about inappropriate benzodiazepine prescriptions in the elderly. Patient Educ Couns. 2013;92(1):81-87. [DOI] [PubMed] [Google Scholar]
- 25.Martin P, Tamblyn R, Ahmed S, Benedetti A, Tannenbaum C. A consumer-targeted, pharmacist-led, educational intervention to reduce inappropriate medication use in community older adults (D-PRESCRIBE trial): study protocol for a cluster randomized controlled trial. Trials. 2015;16:1-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Epstein RM, Street Jr RL. Patient-centered communication in cancer care: promoting healing and reducing suffering. 2007.
- 27.Duke CC, Lynch WD, Smith B, Winstanley J. Validity of a new patient engagement measure: the Altarum Consumer Engagement (ACE) Measure™. Patient-Patient-Centered Outcome Res. 2015;8:559-568. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Linsky AM, Stolzmann K, Meterko M. The patient perceptions of deprescribing (PpoD) survey: short-form development. Drugs Aging. 2020;37(12):909-916. [DOI] [PubMed] [Google Scholar]
- 29.Hoddinott SN, Bass MJ. The dillman total design survey method. Can Fam Physician. 1986;32:2366. [PMC free article] [PubMed] [Google Scholar]
- 30.Nallapeta N, Reynolds JL, Bakhai S. Deprescribing proton pump inhibitors in an academic, primary care clinic: quality improvement project. J Clinical Gastroenterol. 2020;54(10):864-870. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Lukacena KM, Keck JW, Freeman PR, Harrington NG, Huffmyer MJ, Moga DC. Patients’ attitudes toward deprescribing and their experiences communicating with clinicians and pharmacists. Ther Adv Drug Saf. 2022;13:20420986221116465. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Green AR, Boyd CM, Gleason KS, et al. Designing a primary care–based deprescribing intervention for patients with dementia and multiple chronic conditions: a qualitative study. J Gen Int Medicine. 2020;35:3556-3563. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Boyd RW, Lindo EG, Weeks LD, McLemore MR. On racism: a new standard for publishing on racial health inequities. Health Aff Forefront. 2020;
- 34.Zisman-Ilani Y, Khaikin S, Savoy ML, et al. Disparities in shared decision-making research and practice: the case for black American patients. Ann Fam Med. 2023;21(2):112-118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Crutzen S, Abou J, Smits SE, et al. Older people’s attitudes towards deprescribing cardiometabolic medication. BMC geriatr. 2021;21(1):366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Reeve E, Wolff JL, Skehan M, Bayliss EA, Hilmer SN, Boyd CM. Assessment of attitudes toward deprescribing in older Medicare beneficiaries in the United States. JAMA Int Med. 2018;178(12):1673-1680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.National Center for Veterans Analysis and Statistics. Profile of Veterans 2017; 2019.
- 38.Mehta HB, Li S, An H, Goodwin JS, Alexander GC, Segal JB. Development and validation of the summary Elixhauser Comorbidity Score for use with ICD-10-CM–coded data among older adults. Ann Int Med. 2022;175(10):1423-1430. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary file1 (200 KB)
Data Availability Statement
The datasets analyzed during the current study are available from the last author on reasonable request.