Abstract
Health literacy is an important priority in health care delivery, but its effect on clinical outcomes remains incompletely elucidated. This observational cohort study examined the association of health literacy with medication knowledge, adherence, and adverse drug events among cognitively intact veterans older than 65 years old who were taking 5 or more medications and who were enrolled in a Veterans Administration primary care clinic. Health literacy was determined by the Rapid Estimate of Adult Literacy in Medicine. Medication knowledge and adherence were assessed by clinical pharmacist interview and refill data. Adverse drug events were determined by interview and chart review at 3 and 12 months. The 310 subjects had a mean age of 74 years, 99% were White, and 97% were male. Percentage of medications known was 29% for the low health literacy group versus 49% (marginal) and 56% (adequate), p < .001. Known medication purposes were lower in the lower health literacy group (49% vs. 71% vs. 74%; p < .001). Health literacy was not associated with medication adherence: the low health literacy group took 84% of medications by label instructions compared with 80% (marginal) and 77% (adequate), p = .14; or with adverse drug events at 1 year (48% vs. 33% vs. 40%; p = .30). Patients with lower health literacy have poorer medication knowledge but not lower adherence or increased adverse drug events.
Health literacy is an important national priority for improving health care delivery and outcomes (Office of Disease Prevention and Health Promotion, 2010; Rudd, Citation2011). Low health literacy is associated with lower health-related knowledge, suboptimal use of preventive health services, and poorer chronic illness control (Berkman, Sheridan, Donahue, Halpern, & Crotty, 2011; Herndon, Chaney, & Carden, Citation2011), but the effect of low health literacy on certain clinical outcomes remains incompletely elucidated (Berkman et al., Citation2011; Easton, Entwistle, & Williams, Citation2010; Keller, Wright, & Pace, Citation2008). Better understanding of the relation between health literacy and clinical outcomes is fundamental to the design of initiatives aimed at improving health care delivery to patients with low health literacy.
Adverse drug events (ADEs) represent a clinically significant outcome that has received little study in the context of health literacy (Berkman et al., Citation2011). Low health literacy may lead to increased incidence of ADEs through misunderstanding of prescription instructions and warning labels (Davis et al., Citation2006), incorrect self-administration (Kripalani et al., Citation2006), inability to communicate symptoms or signs effectively, or barriers to accessing care for monitoring and evaluation (Murray et al., Citation2009). Health literacy may also influence outcomes through an effect on medication adherence (Pignone, DeWalt, Sheridan, Berkman, & Lohr, 2005), although studies have shown inconsistent and sometimes contradictory relations (Berkman et al., Citation2011).
Investigators have described a relation between low health literacy and hypoglycemic events (Sarkar et al., Citation2010) and the lack of an association between health literacy and anticoagulation control (Fang, Machtinger, Wang, & Schillinger, Citation2006). However, to our knowledge, no published studies have measured the association between health literacy and ADEs more broadly. Our objective was to examine the association between health literacy and ADEs in an elderly veteran population; we hypothesized that patients with lower health literacy would experience a higher incidence of ADEs than would patients with adequate health literacy. We also aimed to further examine the relation between health literacy and medication knowledge and adherence in this population.
Method
Patient Population
We recruited participants from the Iowa City VA Healthcare System primary care clinic from 2001 to 2003. Eligible subjects were older than 65 years of age, took five or more scheduled nontopical medications, were fluent in English, and were cognitively intact. Patients who were enrolled in a pharmacist-based anticoagulation clinic or had impaired cognitive function, defined as having four or more errors on the 10-item Pfeiffer Short Portable Mental Status Questionnaire (Pfeiffer, Citation1975) were excluded. The study sample consisted of the 310 individuals enrolled in the Enhanced Pharmacy Outpatient Clinic Study (Kaboli et al., Citation2004; n = 532) who completed a Rapid Estimate of Adult Literacy in Medicine (REALM) evaluation.
Data Elements
Demographic information included age, race, gender, and self-reported education level (grade school, high school, and beyond high school). Health literacy was assessed using the REALM (Davis et al., Citation1993) a brief test of word recognition and pronunciation using health-related vocabulary. Instrumental activities of daily living scores were obtained as part of the Enhanced Pharmacy Outpatient Clinic Study using the Modified Older American Resources and Services Questionnaire, in which scores range from 0 (total dependence) to 21 (total independence). One OARS question, assessing need for help in taking medicines in correct amounts at correct times, was included in the analysis.
Medication knowledge and adherence were obtained at study baseline during a detailed face-to-face interview between the patient and a clinical pharmacist. Pharmacists followed a structured protocol (protocol available from the corresponding author) in which stepwise questioning was used to assess knowledge of medication name and indication. Patients who could recall medication names from memory were recorded as having correct understanding; indication was likewise judged.
Complete medication lists were then detailed by asking patients to recall strength, dose, and schedule. For patients unable to recall their medications, sequential prompting (showing medication to patient, having patient read from label, enlisting help from accompanying person, or referring to electronic medical record medication list) was used as needed to obtain a complete list of medication names, strengths, doses, and schedules. Patients were considered adherent to a medication if they reported taking the correct strength at the dose and schedule specified on the label.
Medication possession ratios were calculated as an additional, distinct measure of adherence. The medication possession ratio, a refill-based adherence measure that is based on administrative pharmacy refill data (Hess et al., Citation2006) was calculated for the year before and the year after baseline. Medication possession ratios greater than 0.80 are considered adherent, and 1.0 represents perfect adherence.
The study pharmacist assessed ADEs by interview at 3- and 12-month follow-up. The pharmacist first elicited symptoms associated with medications by asking open-ended questions. If a patient reported symptoms, the study pharmacist followed with close-ended questions to determine whether the symptoms and medication were potentially linked. The study pharmacist determined whether the symptoms constituted an ADE on the basis of reasonable causation. Similar methods of patient report with provider corroboration have been used elsewhere as reference standard in assessment of ADEs (Forster, Murff, Peterson, Gandhi, & Bates, 2005; Zhu et al., Citation2011). In addition, the investigators reviewed charts to identify documented ADEs.
Analyses
On the basis of their REALM scores, patients were stratified into three groups: ≤6th-grade equivalent (low health literacy), 7th–8th-grade equivalent (marginal), and ≥ 9th-grade equivalent (adequate). We compared patient characteristics across the three literacy level groups using two predetermined contrasts. The first contrast compared the low health literacy group to the remaining two groups, and the second contrast compared the adequate health literacy group to the remaining two groups. Contrasts of continuous variables were conducted using a generalized linear model approach and chi-squared tests were used for categorical variables. We further used multivariable logistic regression to determine whether health literacy and education were associated with risk for adverse drug events. Because health literacy and education are often strongly correlated, we conducted three separate models. The first included health literacy but not education, the second included education but not health literacy, and the third included both health literacy and education. All three models were adjusted for a number of factors potentially related to adverse drug event risk, including age, number of medications, and medication adherence.
We had full access to and take full responsibility for the integrity of the data. All analyses were conducted using SAS version 9.2 (Cary, NC). The study was approved by the University of Iowa Institutional Review Board and the Iowa City VA Medical Center Research and Development Committee.
Results
Patient Characteristics
The 310 patients we recruited had a mean age of 74 years; 99.0% were White and 97.1% were male. In terms of education, 94 patients (30.3%) had completed grade school, 128 (41.3%) completed high school, and 88 (28.4%) reported education beyond the high school level (Table ). REALM scores categorized patients into low (n = 27, 8.7%), marginal (n = 94, 30.3%), or adequate (n = 189, 61.0%) health literacy levels. Patients took on average 13.2 medications at baseline and 13.6 medications at 12 months. Mean Modified Older Americans Resources and Services activities of daily living score at baseline was 13.2 (SD = 1.7). In addition, 288 patients (92.9%) reported needing no help taking medication in the correct amounts and at the correct times; 22 patients (7.1%) reported needing some help.
Table 1. Patient characteristics and medication knowledge, adherence, and safety (N = 310)
Medication Knowledge
Overall, 56.4% of medication names were known to patients, and 78.5% of medication purposes were known to patients. Lower health literacy was associated with less knowledge of medication names and purposes (Table ). Patients with low health literacy knew 32.2% of medications by name, as compared with 54.6% of medication names for patients with marginal health literacy, and 60.8% for patients with adequate health literacy (p < .001). The lower literacy group knew the purpose of 61.8% of their medications, compared with 77.4% and 81.4% in the marginal and adequate literacy groups, respectively (p < .001).
Table 2. Patient characteristics and medication knowledge, adherence, and safety, stratified by health literacy groupings
Medication Adherence
At baseline, patients reported taking 71.4% of medications as instructed on their prescription drug labels. Health literacy was not associated with self-reported medication adherence (Table ). At baseline, patients with low health literacy reported taking 73.8% of medications correctly, compared with 73.0% among patients with marginal health literacy and 70.4% among those with adequate health literacy.
Medication possession ratios were 0.92 in the year before enrollment and 0.89 in the year after enrollment. Medication refill adherence did not differ significantly between health literacy groups, using either data prior to baseline enrollment or during the year after enrollment (Table ).
The number of medications prescribed was inversely related to medication adherence. When analyzed by tertiles, patients on 9 or fewer medications were taking a greater percentage of these medications according to label (83%) than were patients prescribed more than 12 medications (76%; p = .02). Health literacy was not associated with the number of prescribed medications (Table ).
Adverse Drug Events
A total of 51 patients (16.5%) experienced an ADE within the first 3 months of the study, which increased to 119 patients (38.4%) over the full 12-month follow-up period. Patients with low health literacy appeared to have more ADEs at 3 and 12 months, but this increase was not statistically significant (Table ). At 3 months, 7 (25.9%) of patients with low health literacy had experienced an ADE, compared with 11 (11.7%) and 33 (17.5%) of the marginal and adequate literacy groups (p = .16 for low vs. marginal/adequate literacy; p = .55 for adequate vs. marginal/low literacy). At 12 months, 13 (48.1%) low, 31 (33.0%) marginal, and 75 (39.7%) adequate health literacy patients had one or more ADEs during the preceding 12 months.
In multivariable analysis, health literacy and education defined by grade level completed did not predict occurrence of ADEs at 12 months. (Table ). The health literacy measure and education did appear to interact distinctly with ADEs, with the lowest literacy group as measured by REALM having an odds ratio of 1.6 (0.65, 3.7; p = .32). When both education and measured health literacy were included, the low literacy group had odds ratio of 1.8 (0.69, 4.6; p = .23); when education alone was included, there was no trend towards increased risk for ADEs.
Table 3. Multivariable models of adverse drug event risk at 12 months: Health literacy versus education
Discussion
In this study of elderly veterans’ medication knowledge, adherence, and adverse drug events, we found a significant association between health literacy and knowledge of the medication name and purpose, as has been shown in other studies (e.g., Kripalani et al., Citation2006). However, we found no significant association between health literacy and adherence, or between health literacy and ADEs. Our mixed results highlight the complex and persistently undefined role of health literacy in medication adherence and patient safety.
The lack of an observed association between low health literacy and more distal outcomes such as adherence and ADEs could be related to several factors. First, the study may have been underpowered to show a difference in ADEs. The incidence of ADEs at 3 and 12 months appeared higher among patients with low health literacy, but this was not statistically significant. On the basis of the sample size and a 12-month ADE frequency of 38%, the study was sufficiently powered at the 0.80 level to find a two-group difference in ADE frequency of 30% versus 46%. If present, but not identified because of lack of power, any potential effect is to fall with these bounds.
Second, the REALM is a test of word recognition and pronunciation (Davis et al., Citation1993; Keller et al., Citation2008) it does not assess domains of health literacy that may affect patients’ medication management, such as numeracy or the ability to obtain, process, and act on health information. It is possible that, if health literacy were assessed using other instruments, such as the Short Test of Functional Health Literacy in Adults or Newest Vital Sign (Nurss, Parker, Williams, & Baker, Citation1998; Weiss, Mays, & Martz, Citation2005) results may differ.
Third, patients with low health literacy may have coping strategies developed over a lifetime of functioning, which allow them to take medications appropriately, despite having less medication knowledge. However, one published study that assessed whether medication management strategies might mediate the relation between health literacy and adherence did not find this to be the case (Kripalani, Gatti, & Jacobson, Citation2010). Still, the lack of a consistent relationship between health literacy and medication adherence suggests that other factors are at play (Berkman et al., Citation2011).
Social support is one such factor that may mediate the relation between health literacy and outcomes. Older veterans with low health literacy may rely on others in taking medications correctly, and 7% of our study group reported needing some help to take medications at correct times, suggesting that social support may have tempered the effect of low health literacy. In addition to direct assistance, a social support system might provide reminders, prompting, assistance in obtaining refills, or observations of new signs or symptoms that are not reflected in a self-report of reliance on others.
One surprising conclusion from this study was that low health literacy patients do as well in following medications regimens with multiple drugs as do their counterparts with adequate health literacy; medication adherence did not vary significantly across the health literacy groups (Table ). Further work should be directed at identifying characteristics or compensatory strategies of patients with low health literacy but good outcomes. Understanding the strategies of “positive deviants” (Marsh, Schroeder, Dearden, Sternin, & Sternin, 2004); patients with low health literacy who are exceptionally successful in accessing care, adhering to medication regimen, achieving favorable markers of disease state control, and avoiding adverse drug events may inform the development of interventions.
Strengths and Limitations
The rates of medication adherence and ADEs are comparable to previously identified rates, lending validity to our study outcomes. The rate of ADEs, with 17% of patients experiencing one or more ADEs at 3 months, and 38% at 12 months is comparable to rates found using similar techniques to identify ADEs, and consistent with prior analyses that found higher rates of identified ADEs in studies using patient interview, and among patient populations taking greater numbers of medications (Gandhi et al., Citation2003; Thomsen, Winterstein, Sondergaard, Haugbolle, & Melander, 2007). Our measure of medication adherence was reasonably robust, comprising rigorous and formalized pharmacist interview; the adherence rates we found are comparable to overall rates in large population studies (Fischer et al., Citation2010; Siegel, Lopez, & Meier, Citation2007) using medication possession ratios.
This study has a number of potential limitations. First, as noted earlier, the study may have been underpowered to prove a difference; incidence of ADEs was higher in the low health literacy group, but the trend was not statistically significant. Second, though we did not formally distinguish new from established medications, given the clinical setting, we believe a small proportion of the medications in this study were newly prescribed. Adherence may be lower for newly prescribed medications (Fischer et al., Citation2010), and low health literacy may pose greater challenges for medications that are relatively unfamiliar to patients. Future studies should assess whether health literacy is associated with medication outcomes for newly prescribed medications.
Third, we excluded patients enrolled in the pharmacy based anticoagulation clinic. As anticoagulants have been implicated in a substantial proportion of ADEs (Gurwitz et al., Citation2003; Thomsen et al., Citation2007) exclusion of these patients may lead to underestimation of ADEs and possible attenuation of the effect of health literacy on ADEs. However, studies of ADEs among outpatients suggest that cardiovascular drugs (not including anticoagulants), analgesics, selective serotonin reuptake inhibitors, hypoglycemic agents, and anti-infectives are more common culprit agents (Gandhi et al., Citation2003; Long et al., Citation2010), with anticoagulants accounting for 8% of ADEs in one study of more than 30,000 older adults in an ambulatory setting (Long et al., Citation2010). Given that our study was not designed to account for medication type, it is possible that differential exposure to high-ADE risk drugs between the different health literacy levels explains the lack of significant effect.
Fourth, our study sampled patients who presented for a clinic appointment, and therefore selection may be biased to those patients who, regardless of health literacy level, demonstrated ability to obtain scheduled medical care. Although adherence is estimated to play a role in 21–37% of preventable ADEs, and patient education in 18%, inadequate monitoring is a factor in around 60% of events (Gurwitz et al., Citation2003; Thomsen et al., Citation2007). Low health literacy may exert effects though compromised adherence, reduced effectiveness of education provided, and missed monitoring if low literacy patients do not present for appropriate follow-up (Easton et al., Citation2010); our study population included only those patients who had already overcome potential literacy-related barriers to accessing regular care.
Last, the study population of elderly, mostly White, male veterans lends strengths as well as limitations to our conclusions. Factors such as race and health insurance status are potential confounders when looking at health outcomes. The homogeneity of our study patients and their ability to access health care in effect removes these confounders, though it also limits generalizability. Care delivery through the VA Healthcare System simplifies the process of receiving medications; patients who receive care in less integrated systems, or who use multiple pharmacies, may suffer greater consequences of low health literacy.
Design of study methodologies that can consistently distinguish remediable factors related specifically to health literacy from societal factors, which in themselves can affect health outcomes (Pollitt, Rose, & Kaufman, Citation2005), remains an ongoing challenge.
Conclusion
Our study demonstrated that lower health literacy is associated with poorer medication knowledge, but this did not translate into significantly lower adherence or a significantly greater incidence of ADEs in this study population. Researchers should assess what factors may attenuate the effect of health literacy on medication use and other outcomes. A better understanding of such factors may contribute to the development of additional interventions to improve patient safety.
Notes
All models were adjusted for age, number of medications, percentage of medication names known, percentage of medication purposes known, percentage of medications taken according to label directions and medication possession ratio. None of these adjustment variables were significant independent predictors of adverse drug event risk.
This article not subject to US copyright law.
Related Research Data
References
- Berkman , N. D. , Sheridan , S. L. , Donahue , K. E. , Halpern , D. J. , & Crotty , K. ( 2011 ). Low health literacy and health outcomes: An updated systematic review . Annals of Internal Medicine , 155 , 97 – 107 .
- Davis , T. , Long , S. W. , Jackson , R. H. , Mayeaux , E. J. , George , R. B. , Murphy , P. W. , … Crouch , M. A. ( 1993 ). Rapid estimate of adult literacy in medicine: A shortened screening instrument . Family Medicine , 25 , 391 – 395 .
- Davis , T. , Wolf , M. , Bass , P. , Middlebrooks , M. , Kennen , E. , Baker , D. , … Parker , R. M. ( 2006 ). Low literacy impairs comprehension of prescription drug warning labels . Journal of General Internal Medicine , 21 , 847 – 851 .
- Easton , P. , Entwistle , V. A. , & Williams , B. ( 2010 ). Health in the ‘hidden population’ of people with low literacy. A systematic review of the literature . BMC Public Health , 10 , 459 – 468 .
- Fang , M. , Machtinger , E. , Wang , F. , & Schillinger , D. ( 2006 ). Health literacy and anticoagulation-related outcomes among patients taking warfarin . Journal of General Internal Medicine , 21 , 841 – 846 .
- Fischer , M. , Stedman , M. , Lii , J. , Vogeli , C. , Shrank , W. , Brookhart , M. , … Weismann , J. S. ( 2010 ). Primary medication non-adherence: Analysis of 195,930 electronic prescriptions . Journal of General Internal Medicine , 25 , 284 – 290 .
- Forster , A. J. , Murff , H. J. , Peterson , J. F. , Gandhi , T. K. , & Bates , D. W. ( 2005 ). Adverse drug events occurring following hospital discharge . Journal of General Internal Medicine , 20 , 317 – 323 .
- Gandhi , T. K. , Weingart , S. N. , Borus , J. , Seger , A. C. , Peterson , J. , Burdick , E. , … Bates , D. W. ( 2003 ). Adverse drug events in ambulatory care . New England Journal of Medicine , 348 , 1556 – 1564 .
- Gurwitz , J. H. , Field , T. S. , Harrold , L. R. , Rothschild , J. , Debellis , K. , Seger , A. C. , … Bates , D. W. ( 2003 ). Incidence and preventability of adverse drug events among older persons in the ambulatory setting . JAMA , 289 , 1107 – 1116 .
- Herndon , J. B. , Chaney , M. , & Carden , D. ( 2011 ). Health literacy and emergency department outcomes: A systematic review . Annals of Emergency Medicine , 57 , 334 – 345 .
- Hess , L. , Raebel , M. , Connor , D. A. , & Malone , D. C. ( 2006 ). Measurement of adherence in pharmacy administrative databases: A proposal for standard definitions and preferred measures . Annals of Pharmacotherapy , 40 , 128 – 1288 .
- Kaboli , P. J. , Hoth , B. A. , Carter , B. L. , Chrischilles , E. , Shorr , R. I. , Bhattacharyya , A. , … Rosenthal , G. E. ( 2004 ). The VA Enhanced Pharmacy Outpatient Clinic (EPOC) study: A randomized-controlled pharmacist–physician intervention trial . Journal of General Internal Medicine , 19 ( Suppl 1 ), 227 .
- Keller , D. L. , Wright , J. , & Pace , H. A. ( 2008 ). Impact of health literacy on health outcomes in ambulatory care patients: A systematic review . Annals of Pharmacotherapy , 42 , 1272 – 1281 .
- Kripalani , S. , Gatti , M. E. , & Jacobson , T. A. (2010). Association of age, health literacy, and medication management strategies with cardiovascular medication adherence. Patient Education and Counseling , 81, 177–181.
- Kripalani , S. , Henderson , L. , Chiu , E. , Robertson , R. , Kolm , P. , & Jacobson , T. ( 2006 ). Predictors of medication self-management skill in a low-literacy population . Journal of General Internal Medicine , 21 , 852 – 856 .
- Long , A. , Bendz , L. , Horvath , M. , Cozart , H. , Eckstrand , J. , Whitehurst , J. , … Ferrante , J. ( 2010 ). Characteristics of ambulatory anticoagulant adverse drug events: A descriptive study . Thrombosis Journal , 8 ( 1 ), 5 – 12 .
- Marsh , D. R. , Schroeder , D. G. , Dearden , K. A. , Sternin , J. , & Sternin , M. ( 2004 ). The power of positive deviance . BMJ , 329 , 1177 – 1179 .
- Murray , M. D. , Tu , W. , Wu , J. , Morrow , D. , Smith , F. , & Brater , D. C. ( 2009 ). Factors associated with exacerbation of heart failure include treatment adherence and health literacy skills . Clinical Pharmacology and Therapeutics , 85 , 651 – 658 .
- Nurss , J. R. , Parker , R. M. , Williams , M. V. , & Baker , D. W. ( 1998 ). Short Test of Functional Health Literacy in Adults . Snow Camp , NC : Peppercorn Books and Press .
- Office of Disease Prevention, & Health Promotion. ( 2010 ). National action plan to improve health literacy . Washington , DC : Author . Retrieved from http://www.health.gov/communication/HLActionPlan
- Pfeiffer , E. ( 1975 ). A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients . Journal of the American Geriatrics Society , 23 , 433 – 441 .
- Pignone , M. , DeWalt , D. A. , Sheridan , S. , Berkman , N. , & Lohr , K. N. ( 2005 ). Interventions to improve health outcomes for patients with low literacy. A systematic review . Journal of General Internal Medicine , 20 , 185 – 192 .
- Pollitt , R. , Rose , K. , & Kaufman , J. ( 2005 ). Evaluating the evidence for models of life course socioeconomic factors and cardiovascular outcomes: A systematic review . BMC Public Health , 5 ( 1 ), 7 – 20 .
- Rudd , R. E. ( 2011 ). Improving Americans’ health literacy . New England Journal of Medicine , 363 , 2283 – 2285 .
- Sarkar , U. , Karter , A. J. , Liu , J. Y. , Moffet , H. H. , Adler , N. E. , & Schillinger , D. ( 2010 ). Hypoglycemia is more common among type 2 diabetes patients with limited health literacy: The Diabetes Study of Northern California (DISTANCE) . Journal of General Internal Medicine , 25 , 962 – 968 .
- Siegel , D. , Lopez , J. , & Meier , J. ( 2007 ). Antihypertensive medication adherence in the Department of Veterans Affairs . The American Journal of Medicine , 120 , 26 – 32 .
- Thomsen , L. A. , Winterstein , A. G. , Sondergaard , B. , Haugbolle , L. S. , & Melander , A. ( 2007 ). Systematic review of the incidence and characteristics of preventable adverse drug events in ambulatory care . Annals of Pharmacotherapy , 41 , 1411 – 1426 .
- Weiss , B. D. , Mays , M. Z. , & Martz , W. ( 2005 ). Quick assessment of literacy in primary care: The newest vital sign . Annals of Family Medicine , 3 , 514 – 522 .
- Zhu , J. , Stuver , S. O. , Epstein , A. M. , Schneider , E. C. , Weissman , J. S. , & Weingart , S. N. ( 2011 ). Can we rely on patients’ reports of adverse events? Medical Care , 49 , 947 – 955 .
- This article not subject to US copyright law.