- Email: [email protected]
Evaluation of instruments to assess health literacy in Arabic language among Iraqis
Research in Social and Administrative Pharmacy 11 (2015) 803–813
Original Research
Evaluation of instruments to assess health literacy in Arabic language among Iraqis Ali Azeez Al-Jumaili, M.S.a,b,*, Mohammed Dakhil Al-Rekabi, Ph.D.c, Bernard Sorofman, Ph.D.a a
University of Iowa, College of Pharmacy, Iowa City, IA, USA b University of Baghdad, College of Pharmacy, Baghdad, Iraq c University of Kufa, Faculty of Pharmacy, Al-Najaf, Iraq
Abstract Background: Low health literacy is associated with lack of medical information, less use of preventive measures, low medication adherence rates, high health care costs and high risk of hospitalization. Objective: The aims were to compare the results of the three health literacy tests, to measure for the first time the health literacy level of Iraqis, to describe the use of standardized health literacy tests, to evaluate reliability and validity of the Arabic versions of these tests, and to investigate whether there is relationship between the participant characteristics and the health literacy level. Methods: A convenience sample of 95 subjects was studied in five community pharmacies in Al-Najaf and Babylon governorates, Iraq. Three health literacy tests, the Single Item Literacy Screener (SILS), the New Vital Sign (NVS) and the Short version of the Test of Functional Health Literacy in Adults (S-TOFHLA), were translated in the Arabic language and administered to the pharmacy customers. Results: There were no statistically significant associations between age, gender, education and current education status and NVS score, but there were significant positive associations between the level of education and each one of SILS, New SILS, and S-TOFHLA scores. Conclusions: SILS has one subjective, possibly culturally biased question. Since Iraqis are generally not exposed to reading product labels, the NVS test might be not an accurate measure for them. S-TOFHLA was the most comprehensive test and gave equitable results. The Arabic version of S-TOFHLA can be used to measure health literacy in 22 Arabic speaking countries. Ó 2015 Elsevier Inc. All rights reserved. Keywords: Health literacy; SILS; NVS; S-TOFHLA; Iraq; Pharmacy
Introduction According to the U.S. National Assessment of Adult Literacy (NAAL), literacy is “the ability to
use printed and written information to function in society and to achieve one’s goals”.1 Health literacy is the ability of a person to read, compute,
The authors Ali Azeez Al-Jumaili and Mohammed Dakhil Al-Rekabi are joint first authors. * Corresponding author. University of Iowa, College of Pharmacy, PHAR S532, Iowa City, IA, 52242, USA. Tel.: þ1 319 936 4354 (mobile); fax: þ1 319 353 5646. E-mail address: [email protected] (A.A. Al-Jumaili). 1551-7411/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.sapharm.2015.02.002
804
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
and understand health-related information such as in a physician appointment slip, medication labels and pamphlets.2 Health literacy is not only basic literacy, but also knowledge of health-related topics. A limited literacy does not mean illiteracy, which is inability to read or write.3 The NAAL (2003) survey reported that 14% of the U.S. adults had below basic health literacy, only 12% had proficient health literacy, more than half (53%) had intermediate health literacy and 22% had basic health literacy.1 Elderly people age 65 and older had lower health literacy than younger adults.1 Literacy and health are linked. The average reading skills of American adults is between the seventh and eighth-grade levels.4 Health education materials require a reading level higher than the average reading skills of U.S. adults.5 Weiss and colleagues concluded that populations with lower literacy rates have poorer health status especially in non-industrialized countries.6 Low health literacy is associated with poor health-related outcomes, such as high risk of hospitalization.7 Low health literacy is associated with lack of medical information, impaired health care knowledge, less use of preventive measures, lower medication compliance rates and higher health care costs.8 Similarly, Howard and colleagues concluded that the medical costs of patients with inadequate health literacy were significantly higher than that for patients with adequate health literacy.9 There are several measures of health literacy; one is the Single Item Literacy Screener (SILS) developed by Morris and colleagues and asking about the amount of help that is needed to read and understand doctor or pharmacy-written instructions.10 The Short version of the Test of Functional Health Literacy in Adults (STOFHLA) developed by Baker and colleagues11 shortened the time required by the TOFHLA.12 S-TOFHLA is a standard for assessment of health literacy and takes 12 min.10,11 In 2009, a review article included 42 literacy studies and concluded that S-TOFHLA was considered as a standard test for the comparison of other health literacy tests.13 The New Vital Sign (NVS) test developed by Weiss and colleagues6 measures both reading and numeracy skills of subjects through reading an ice cream label contents and answering six related questions.6 The NVS is shorter but less comprehensive than the S-TOFHLA. SILS, NVS and S-TOFHLA are the most popular health literacy tests. Four of the health literacy studies that used TOFHLA, S-TOFHLA, NVS or NVS-UK stratified the health literacy
results according to the gender, age, and years of school education.8,11,14,15 A recent SILS study in Iran classified the health literacy results by age, education degree and socioeconomic status.16 Many studies investigated using one test (STOFHLA,8,12,16–19 NVS,20 or SILS16), and others compared two health literacy tests (NVS vs STOFHLA,6,11,14,21,22 NVS vs SILS15 or SILS vs S-TOFHLA10), but comparing the three most common tests (SILS, NVS and S-TOFHLA) was the major contribution of the present study. Instruments; standard measures of health literacy The Single Item Literacy Screener (SILS) SILS is a primary indication that participants have inadequate reading abilities and may request help to read health-related information.10 SILS was administered in written format: “How often do you ask someone for help to read the instructions and leaflets from a doctor or pharmacy?” A participant could choose one of the following answers (5-point Likert scale): 5-never, 4-rarely, 3-sometimes, 2-often or 1-always. If a participant chooses sometimes, often or always, that indicates a difficulty with reading of health materials. Never and rarely indicate adequate reading ability.10 Short version of the Test of Functional Health Literacy in Adults (S-TOFHLA) S-TOFHLA evaluates both the numeracy and reading skills of participants. The reading part has two prose passages while the numeracy section includes four questions that evaluate understanding of glucose monitoring, prescription labels and appointment slips.2,11 The reading sections of the S-TOFHLA test included this statement: “fill in the blanks using a word from a list which best completes the sentence grammatically and contextually.” The passages are related to preparation for an upper gastrointestinal X-ray and Medicaid rights & responsibilities.11 When the S-TOFHLA was translated to Arabic, one of the 36 cloze items was dropped from the passages because it did not make cultural sense after translation to Arabic. The dropped item was the third one in the X-Ray passage with four choices (is, am, if, it). “Medicaid” was translated as health care assistance for needy people. The two prose passages have a total of 35 cloze items totaling 70 points (two points per each item). The cloze process includes deleting every fifth to seventh word in a passage. The participant should choose the most appropriate answer to fill the blank from a list of four words.11 The score of
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
the four numeracy items is 30 points (7.5 points per each one). Thus, the total score for the whole S-TOFHLA test (both readings and numeracy items) would be 100 points.11 The score could be classified into one of three health literacy levels: 0–53 refers to inadequate functional health literacy; 54–66 indicates marginal health literacy, while 67–100 refers to adequate functional health literacy.11 The Newest Vital Sign (NVS) Participants read an ice cream label and then the interviewer read the six questions to the participant. The column of the content percentages on the ice cream label was deleted in this study as Rowlands and colleagues (2013) did in their study with the NVS-U.K14 because these percentages were not necessary to answer the questions. One point was given for each correct answer of the NVS test. The scores between zero and one indicate “probably inadequate” (limited) health literacy, two to three indicate “possibility inadequate” literacy, while four to six refer to “adequate” health literacy.6 This is a study comparing health literacy measures in an Iraqi population during 2013– 2014. In 2011, Iraq had a population of 32,665,000 with a life expectancy at birth of 69 years.23 The Iraqi Household Socio-Economic Survey of 2007 indicates that approximately 23% (5,193,682) of Iraqis were illiterate.24 According to the UNESCO Institute for Statistics, the adult (15 and older) literacy rate in Iraq was 78.5% (86.0% for male and 71.2% for female) in 2011.23 In 2010, Iraq had 6,635,771 students at different educational levels.24 More than three quarters (77%) of Iraqi people are native Arabic speakers.25 English is taught in elementary, middle, high schools and some colleges in Iraq. However, Iraqis do not use English in their daily communication except among medical staff and some academic faculty. Consequently, the majority of Iraqi people are not skilled English speakers. Most Iraqis prefer reading the Arabic version of drug labels and leaflets because they are more understandable to them. Therefore, the Iraqi Ministry of Health (MOH) requires that medication labels and leaflets are written in both Arabic and English. For the same reason, pharmacists and physicians use the Arabic language to counsel patients. Although many health literacy studies have been conducted in the U.S. and Western Countries for 20 years, no such study has been conducted in Iraq.
805
The aims of this study were to compare the results of the three common health literacy tests, to measure for the first time the health literacy level of Iraqis, to describe the use of standardized health literacy tests for Iraqis, to evaluate reliability and validity of the Arabic versions of these tests, and to investigate whether there is relationship between Iraqi participant characteristics and the reported health literacy scores on common measures developed in Western culture.
Methods Procedure The study was cross-sectional and descriptive. It surveyed a convenience sample of 95 Iraqi pharmacy customers from September 2013 to February 2014. Self-administered testing was conducted in five community pharmacies, three of them located in Al-Najaf and two pharmacies in Babylon governorates of Iraq. The pharmacists used Arabic translated SILS, S-TOFHLA and NVS tests. S-TOFHLA and NVS included both reading and quantitative sections which measured the ability to understand the numbers on an ice cream label (NVS), and appointment slips, prescription vials and blood glucose level (STOFHLA). The participants were pharmacy customers who came to pick prescriptions up or ask for over the counter medications for themselves or for a family member. Exclusion criteria were subjects (children) less than 15 years-old, health care professionals, subjects who had mental illness and subjects with vision impairment. All participants had some degree of literacy that enabled them to read the written literacy tests and answer the questions. All participants were Arabic native speakers. Each test required translation into formal Arabic. The translation used formal methodology by Wild (2005).26,27 Formal Arabic language can be commonly understood in 22 Arabic-speaking countries (in the Middle-east and North Africa). A pilot study with 25 subjects was conducted to assess any ambiguous statements or questions and to make sure the Arabic forms were understandable by pharmacy customers since these were the first Arabic versions of the three health literacy tests. The translation of NVS and STOFHLA was refined after the feedback from the pilot study. Pharmacists approached customers to ask them whether they would like to participate in
806
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
health literacy tests. After receiving verbal consent, the three tests were administered. The questions were given in the same order for all participants (SILS, NVS then S-TOFHLA). All the tests were answered at one time in the pharmacies. Basic demographic information (i.e., age, gender, educational level) was collected. No compensation was provided. Reliability analyses were conducted for the Arabic versions of S-TOFHLA and NVS tests using Cronbach’s alpha, which represents the internal consistency of the items, as performed by Baker and Weiss.6,11 Construct validity was also conducted by computing bivariate correlations (Pearson’s coefficient) between NVS and S-TOFHLA and among STOFHLA sections. The proposal was Institutional Review Board (IRB) exempt at the University of Iowa, USA and Ethical Committee approved in Iraq.
Statistical analyses Most analyses were performed using Statistical Analysis System (SAS, version 9.3, SAS Inc., Cary, North Carolina, USA). Various methods were used. Means, range and standard deviations (SD), frequencies and percentages of participant characteristics such as participants’ age, gender and educational degrees were measured. Because a cultural concern arose regarding the use of “sometimes,” sensitivity analysis was performed to compare the SILS results after moving each “sometimes” choice to the “adequate” literacy group, identified as New-SILS (NSILS). Fisher’s Exact test was performed due to small cell counts to compare the proportions of gender and education into the SILS, NSILS, NVS and STOFHLA test levels (Tables 1 and 2). For the Fisher’s Exact test, the five SILS and NSILS responses were classified into two levels (limited
Table 1 The results of SILS and New-SILS for each participant group A. Actual SILS results Participant characteristics
Gender Female Male Education (years) Less than middle school (!9) Middle school (9) High school (12) Some colleges (14–18) Total: N (total %)
Reading ability
Total N (row %)
Limited (Always, often, sometimes) N (row %)
Adequate (Rarely, never) N (row %)
30 (71.43) 42 (79.25)
12 (28.57) 11 (20.75)
42 (100) 53 (100)
6 15 11 40 72
0 3 6 14 23
6 18 17 54 95
(100) (83.33) (64.71) (74.07) (75.97)
(0.00) (16.67) (35.29) (25.93) (24.21)
(100) (100) (100) (100) (100)
B. Sensitivity analysis: New-SILS results after switching every “sometimes” answer to adequate category Participant characteristics
Gender Female Male a Education (years) Less than middle school (!9) Middle school (9) High school (12) Some colleges (14–18) Total: N (total %)
Limited (Always, often) N (row %)
Adequate (Sometimes, rarely, Never) N (row %)
Total N (row %)
17 (40.48) 21 (39.62)
25 (59.52) 32 (60.38)
42 (100) 53 (100)
6 8 5 19 38
0 10 12 35 57
6 18 17 54 95
(100) (44.44) (29.41) (35.19) (40.00)
(0) (55.56) (70.59) (64.81) (60.00)
(100) (100) (100) (100) (100)
Fisher’s Exact tests showed significant (P ! 0.05) relationship between level of education and NSILS result categories. a
807
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813 Table 2 The results of NVS and S-TOFHLA according to gender and education levels of the participants A. The results of NVS Participant characteristics
Health literacy level Probably inadequate (0–1 correct answer) N (row %)
Gender Female Male Education (years) Less than middle school (!9) Middle school (9) High school (12) Some colleges (14–18) Total: N (total %)
8 (19.51) 6 (13.33) 2 4 2 6 14
(40.00) (25.00) (11.76) (12.5) (16.28)
Total N (row %)
Possibly inadequate (2–3 correct answers) N (row %)
Adequate (4–6 correct answers) N (row %)
18 (43.90) 22 (48.89)
15 (36.59) 17 (37.78)
41 (100) 45 (100)
1 5 9 25 40
2 7 6 17 32
5 16 17 48 86
(20.00) (31.25) (52.94) (52.08) (46.51)
(40.00) (43.75) (35.29) (35.42) (37.21)
(100) (100) (100) (100) (100)
B. The results of S-TOFHLA Participant characteristics
Gender Female Male a Education (years) Less than middle school (!9) Middle school (9) High school (12) Some colleges (14–18) Total: N (total %)
Inadequate (Scores 0–53) N (row %)
Marginal (Scores 54–66) N (row %)
Adequate (Scores 67–100) N (row %)
Total N (row %)
4 (9.52) 4 (7.84)
4 (9.52) 4 (7.84)
34 (80.95) 43 (84.31)
42 (100) 51 (100)
2 1 2 3 8
2 3 1 2 8
2 14 14 47 77
6 18 17 52 93
(33.33) (5.56) (11.76) (5.77) (8.60)
(33.33) (16.67) (5.88) (3.85) (8.6)
(33.33) (77.78) (82.35) (90.38) (82.80)
(100) (100) (100) (100) (100)
a Fisher Exact test showed significant relationship (P ! 0.05) between education levels and S-TOFHLA score categories.
and adequate) (Table 1). On the other hand, each of NVS and S-TOFHLA had three result levels (inadequate, marginal and adequate) (Table 2). The significance level was 0.05 for Fisher tests and regression tests and 0.01 for bivariate correlation test. Ordinal logistic regression28 estimated the association between the SILS, NSILS and NVS results with the four independent variables: gender, age, education and current student status (whether participant was a current student or not). For the logistic regression analyses, the NVS had six categories (scores: 1–6) while SILS and NSILS had five categories. A generalized linear model (GLM) procedure was run for the continuous S-TOFHLA scores to measure the relationship between the explanatory variables: age (continuous), gender (categorical), educational degrees (categorical) and current student status (categorical) with the outcome variable, S-TOFHLA score.
Statistical Package for Social Sciences (SPSS) (IBM SPSS Statistics, Armonk, NY, USA) was used to measure the Pearson correlation (r) among the three health literacy tests, between numeric and reading sections of S-TOFHLA and between X-Ray and Medicaid-related cloze passages of S-TOFHLA. SPSS was also used to calculate the reliability analysis (Cronbach’s alpha) of both reading passages and numeric section of S-TOFHLA, and NVS questions. Two participants skipped the numeric sections of S-TOFHLA and they have not been included in the regression and Fisher test analyses. Other nine participants skipped NVS test and they have not been included in the NVS-related analyses.
Results Ninety-five subjects participated in the study, 42 females and 53 males. The average age of the
808
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
participants was 27.47 (9.58) with a range of (15–60) years old. The participants were classified into four groups according to their education degree, less than middle school (6.32%), completed middle school (18.95%), completed high school (17.89%), and two-year college, bachelor or master (56.84%). The results of SILS were categorized into either limited or adequate reading ability. More than three quarters (75.79%) of participants showed limited reading ability with a SILS answers: “always, often or sometimes.” The remaining (24.21%) had adequate reading ability since they answered “never or rarely” to the SILS question (Table 1A). Sensitivity analysis for SILS was conducted to examine the SILS results after switching every “sometimes” answer to the “adequate” group. The NSILS results showed that 40.0% of the participants had limited reading ability while 60.0% of them had adequate reading ability (Table 1B). Eighty-six participants had an NVS score average of 3.02 (1.74) out of six. The NVS test results showed that 16.3% of participants had a high probably of inadequate health literacy, 46.5% had possibly inadequate health literacy and 37.2% had adequate health literacy (Table 2A). In contrast, according to S-TOFHLA test, 8.6% had low literacy, 8.6% had marginal literacy and 82.8% of participants had adequate health literacy (Table 2B). The participants answered the prose passage (82%) more often correctly than the numeracy (75.4%) questions of the S-TOFHLA (Table 3). The X-Ray passage had higher literacy responses (85.9%) than the Medicaid equivalent rights and responsibilities passage (78.6%). The doctor appointment slipquestion had the lowest correct response (63.8%) while the blood sugar-level question had the highest correct response (79.7%) among the four numeracy S-TOFHLA questions. In this sample of Iraqi adults, the Fisher’s Exact test showed no significant associations between gender and education categories, and the NVS and SILS results (Tables 1 and 2). However, there were significant relationship (P ! 0.05) between education levels and NSILS and STOFHLA result categories according to the same test. The relationship between the four independent variables and each of SILS and NSILS responses and NVS scores were measured by using ordinal logistic regression. The logistic regression analysis indicated there were no significant associations
Table 3 Percentage of correct answers for S-TOFHLA portions
A 1
2
B
S-TOFHLA test items
Correct answers (%)
Total S-TOFHLA score Reading passages (N ¼ 95) Preparations for X-Ray 0–5 items (Preparing for X-ray instructions) 6–11 items (day before X-ray) 12–16 items (At X-ray day) Medicaid-related passage 1–7 items 8–13 items 14–19 items Numeracy items (N ¼ 93) Take medication every 6 h Blood sugar-level Doctor appointment slip Taking medication before lunch
79.12 81.99 85.9 88.8 94.6 90.2 78.6 80.0 70.2 86.0 75.4 76.49 79.7 63.8 73.8
The table format adapted from Baker.11
between the four independent variables and NVS scores (1–6). The logistic regression analysis indicated that there was significant positive relationship (P ! 0.05) between participant education level and the SILS and NSILS responses after controlling the other three explanatory variables. The participants with less than middle school had 11.1 times higher odds of choosing the lowest SILS response (always) compared to those with college education after controlling other variables (Table 4A).28 Similarly, Table 4B shows that participants with a less than middle school compared to those with college education had 12.8 times higher odds to choose the lowest NSILS response after controlling other variables.28 That means the participants with lowest education level (less than middle school) were 11.1 (for SILS) and 12.8 (for NSILS) times more likely to select the lowest SILS choice compared to those with the highest education level (some colleges) after controlling age and gender. Hence, there were positive relationships between education level and each of SILS and NSILS results while there were no significant relationships between the test responses and age, gender and current student status (Table 4). Likewise, multiple linear regression analysis showed that participants with some college education and middle school education had an average of 23.2 and 16.6 higher points (scores) in S-TOFHLA test respectively compared to
809
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813 Table 4 Odds ratios, 95% confidence limits & P-value of the SILS and NSILS logistic regression model results Effect (parameter)
Point 95% estimate Confidence (odds ratio) limits
A. SILS odds ratio estimates Age 0.975 Gender M vs F 1.384 Education 1 vs 4 11.094 Education 2 vs 4 1.247 Education 3 vs 4 0.732 B. NSILS odds ratio estimates Age 0.977 Gender M vs F 1.117 Education 1 vs 4 12.777 Education 2 vs 4 1.407 Education 3 vs 4 0.493
Pvalue
0.935 1.016 0.233 0.646 2.966 0.403 1.669 73.759 0.013a 0.460 3.378 0.665 0.271 1.972 0.537 0.936 1.020 0.298 0.511 2.444 0.781 1.904 85.733 0.009a 0.507 3.900 0.512 0.174 1.394 0.182
Each of the outcome variable (SILS or NSILS) had five levels. Education 1 ¼ less than middle school, Education 2 ¼ middle school, Education 3 ¼ high school, Education 4 ¼ some college degrees (high diploma, bachelor or master). Female and Education 4 were references for these analyses. a Ordinal Logistic regression analysis showed that significant positive association between choosing the lowest SILS & NSILS results (always) and lowest education level (P ! 0.05).
participants with less than middle school education. In other words, there was a positive significant (P ! 0.05) association between the educational level and the continuous S-TOFHLA scores. Additionally, the multiple linear regression analysis showed the not-current students had 9.87 fewer points than the S-TOFHLA scores of current students. The current student status had significant positive relationship with S-TOFHLA scores. Coefficient of determination of the regression model (R2) was 0.20, explaining 20% of the outcome (S-TOFHLA score) variance (Table 5). The time required to answer each test was recorded. The average time for the NVS Arabic version was 10.8 (7.13) with an outlier (participant who took 30 min) and 8.0 (6.73) minutes without this outlier. The English, Spanish and Turkish versions for NVS required average time 3.0 1.2, 3.4 1.2 and 6.3 1.27 min, respectively.6,29 On the other hand, the Arabic version of S-TOFHLA test required 11.1 (8.67) minutes in average which was almost the same as reported the English version time (12.0 min). SILS required less than a minute to be answered.
Table 5 Multiple regression model results of continuous STOFHLA scores with age, gender, education level and current student status Parameter
Estimate Standard t Pr O jtj error value
Intercept Age Gender M Gender F Education 2 (middle school) Education 3 (high school) Education 4 (some college) Education 1 (less than middle school) Not-current student Current student R2
72.588 0.194 1.482 0.000 16.615
7.902 0.209 3.413 7.977
2.08
0.0402a
11.826
8.024
1.47
0.1441
23.171
7.238
3.20
0.0019a
4.806
2.05
0.0431a
9.19 !0.0001 0.93 0.3574 0.43 0.6652
0.000
9.866 0.000 0.203
a Significant (P ! 0.05) positive association between the educational level and the continuous S-TOFHLA scores. The analysis references were female for gender, education-1 for education level and current student for current student status.
Cronbach’s alpha of the S-TOFHLA reading section was 0.89. However, Cronbach’s alpha of the S-TOFHLA numeric questions was 0.615. The Cronbach’s alpha of NVS was 0.69. Regarding validity, the numeric and reading sections of S-TOFHLA had significant positive correlation (r ¼ 0.34). Furthermore, there was a significant high positive correlation (0.724) between X-Ray and Medicaid cloze passages of STOFHLA. There was a significant (P ! 0.01) positive correlation (r ¼ 0.513) between NVS and total STOFHLA scores. Both numeric (r ¼ 0.414) and reading (r ¼ 0.351) section scores had significant positive correlation with NVS scores. There were weak positive correlations between SILS and STOFHLA (r ¼ 0.112) and very weak positive correlation between SILS and NVS (r ¼ 0.089 t).
Discussion This study used for the first time, to our knowledge, three common health literacy tests (SILS, NVS & S-TOFHLA) simultaneously and also for the first measured health literacy for Iraqis.
810
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
Exact Fisher tests showed there were statistically significant differences between each of the NSILS and S-TOFHLA result levels according to educational categories while non-significance difference in NVS and SILS levels according to both gender and educational categories. The sample size might not be large enough to detect differences that actually exist between the participant categories. Similarly, Ickes and Cottrel concluded there were no significant relationships between gender, college-year (juniors vs seniors) and TOFHLA scores after using Analysis of Variance (ANOVA) for analysis.8 Morris and colleagues reported that SILS is practical because it is very brief. However, it cannot detect people with limited reading ability (i.e., false negative result). SILS results in the present study indicated that majority of participants had limited reading ability (chose sometimes, often or always). An Iranian study used SILS and found that 30.3% of participants had inadequate health literacy.16 Again, Iraqis usually choose “sometimes” as a positive cultural preference. Many Iraqi participants (34.7%) chose “sometimes,” which culturally may be seen as a literate response, but actually it reflected limited reading ability on the test. The sensitivity analysis changed SILS results (Table 1B) showed that more than half the participants had adequate reading ability after switching “sometimes” to the “adequate” group. Hence, the SILS can be largely influenced by the cultural way of thinking. Both SILS and the NSILS results had a significant positive association with education level (P ! 0.05) (Table 4). Likewise, the logistic regression analysis of an Iranian study reported significant positive relationship between educational level and the SILS response.16 The logistic regression analysis showed there was no significant association between NVS scores and the four participants’ characteristics. The results of the present study matched the results of a Portuguese study that also found no significant association between NVS results and participants’ gender and education level.15 Similarly, the developers of NVS test indicated no significant difference in NVS scores according to gender (independent t-test) for both English and Spanish versions.6 In contrast the developers of the NVS-UK concluded weak correlations between NVS-UK scores and educational level (Pearson’s r ¼ 0.22).14 The NVS test results inferred that most of the participants had limited health literacy. Stagliano
and colleagues (2013) found that 17.8% of the American participants had inadequate literacy.20 According to the same study the mean of the NVS scores for American adults was 3.7 2.0 while the mean for the Iraqi adults was lower (3.02 1.74). Another study in the U.S. showed that mean NVS scores was 3.0 (1.9).21 The difference between Iraqi and American NVS scores might be because it is uncommon for Iraqis to read food labels. The mean of NVS score of this study (3.02 1.74) was relatively higher than the results (2.60 0.08) of a study in Turkey with a Turkish version of NVS.29 Salgado and colleagues used a Portuguese version of NVS and found that 95% of Portuguese elderly participants had limited health literacy with poor NVS average scores (0.81 0.1).15 They mentioned that Portugal has recent and limited experience with food labels.15 Additionally, they disclosed a weakness in the fifth question (“is it safe for you to eat this ice cream?”) structure. They found that question number-five was the question answered correctly most often. They argued that the answer to this dichotomous question depends mainly on the chance although the participants were unaware of the question because most those who answered this question correctly (not safe) failed to know the reason in the next question.15 The present study results agreed with VanGeest and colleagues who indicated that NVS was a sensitive but not specific test for low health literacy and that means NVS has the possibility of giving false positive results.30 NVS can show people have low health literacy, but in fact they have adequate health literacy.30 Rothman and colleagues concluded that even people with high literacy may misinterpret the content of nutrition labels.31 According to S-TOFHLA, 82.8% of Iraqi participants had adequate health literacy level. On the other hand, both developers of the English version of TOFHLA and S-TOFHLA found that slightly more than 50% of American participants had adequate health literacy according to those two tests.11,12 Thirty one percent taking the Spanish version of S-TOFHLA had adequate health literacy.12 Multivariate linear regression analysis showed a significant positive association between level of education, current student status and S-TOFHLA scores (Table 5). That indicated people with higher education usually have better health literacy. The multiple linear regression analysis also
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
showed the current student participants had STOFHLA scores 9.87 higher than those notcurrent students. Similarly, another study revealed that statistically difference in S-TOFHLA scores according to years of education.9 The coefficient of determination of the multiple linear regression analysis was relatively low (R2 ¼ 0.2). However, only two related studies conducted multivariate regression analyses for STOFHLA, and they did not report an R2-value to compare the results.9,11 Use of regression analysis was preferred over the bivariate correlation (Pearson or Spearman correlation coefficient) analyses used in another study,14 because bivariate correlation analysis does not control the effect of other covariates on the outcome variable (STOFHLA score). Salgado and colleagues found that there was an inverse relationship between age and NVS scores.15 The answer time for each test was reported in this study. The SILS question was answered within a very short time (a minute). The S-TOFHLA average time of response was similar to English form. However, the NVS test needed longer time than that reported in the literature because participants took the NVS test in Arabic and were unfamiliar with reading food label contents in their daily life. Because higher education in public universities is free in Iraq, most Iraqis pursue their college/ university study. Thus, more than half (56.8%) of the participants had some college degrees. The majority of the study participants were young (average 27.5 years). The main reason may be because those Iraqi elderly patients usually send a young family member to pharmacies to pick up medications since those young adults are more educated and can understand pharmacist counseling better. Another reason is that young adults usually serve their elderly relatives as a form of traditional norms and courtesy. Regarding to the instrument reliability, Cronbach’s alpha for the S-TOFHLA for Arabic version compared well to the English version.11 The Cronbach’s alpha for the Arabic version NVS was similar to the Spanish version and less than that of the English version.6 That means the Arabic version of S-TOFHLA readings and NVS had good internal consistency (reliability), but this was not the case of numeric section of S-TOFHLA. The validity was assessed by the correlations within and between the health literacy tests. The numeric section of S-TOFHLA had a significant
811
positive correlation with the reading section. Moreover, the Arabic version of S-TOFHLA had positive significant correlations between its reading sections and also between total STOFHLA scores and NVS scores. The significant positive Pearson correlations among the STOFHLA sections indicate that all the sections measure health literacy in the same direction. According to Rowlands et al, correlations (between S-TOFHLA and NVS significantly) larger than 0.3 are considered acceptable.14 Because the Arabic versions of S-TOFHLA and NVS had significant correlation higher than 0.3, they had acceptable validity. Limitations This was the first study in health literacy in Iraq, and that added challenges to conduct the research. The selection of participants was a convenience sample rather than randomized sample. Hence, the results are not generalizable and may not apply for all pharmacy customers in Iraq. The participation was voluntary and without compensation. Pharmacy customers who refused to participate might have low health literacy, so they avoided taking tests that could embarrass them. Random sampling would be recommended for future studies.
Conclusion The study participants scored higher on STOFHLA compared to SILS and NVS tests. Cultural differences could influence the scores of SILS and NVS tests. The SILS is a brief, but subjective test and may not reflect the accurate level of literacy because it has a subjective and possibly culturally biased question. The NVS might be not applicable as a health literacy test for people who are not used to reading product labels in their daily lives such as Iraqis. The S-TOFHLA was the most convincing test because it was neither subjective nor dependent on the familiarity with product labels, and almost all participants (n ¼ 93) completed this test. Hence, if we would conduct a national health literacy survey for Iraqis in the future, we would choose the S-TOFHLA test. In this sample of Iraqi participants, there were no significant differences in NVS according to age, gender, education and current student status. However, there were significant positive associations
812
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
between the level of education and each of the SILS, NSILS and S-TOFHLA results indicating higher educated people might have higher health literacy level. In general, the Arabic version of S-TOFHLA and NVS had acceptable reliability (Cronbach’s alpha) and validity. Since the Arabic version of STOFHLA test gave equitable results in Iraq, it can be used to measure health literacy in 22 Arabic speaking countries.
10.
11.
12.
Acknowledgments We would like to thank Iraqi pharmacists who helped us to conduct the literacy tests in the five community pharmacies were Zaid Ali, Abdelfadhel J, Ali Shlah, Karrar M, and Ali Mohammed. We also appreciate Jacob Simmering for his advice for part of the analysis. We would like to acknowledge Iraqi Higher Committee for Educational Development (HCED) for supporting one of the authors (Ali Al-Jumaili) with a scholarship.
13.
References
16.
1. U.S. Department of Education IOES, National center for Education Statistics. The Health Literacy of America’s Adults, Results from the 2003 National Assessment of Adult Literacy. 2006 ed., http://nces. ed.gov/pubs2006/2006483.pdf; 2006. Accessed 04. 07.14. 2. Tkacz VL, Metzger A, Pruchnicki MC. Health literacy in pharmacy. Am J Health Syst Pharm 2008;65: 974–981. 3. Adult Literacy in America. Vol Third Edition. 2002 ed. U.S Department of Education-Office of Educational Research and Improvement- National Center for Educational Statistics, https://nces.ed.gov/pubs93/ 93275.pdf; 2002. Accessed 02.09.14. 4. Heilman CB. Readability of patient education materials. World Neurosurg 2013;80:e109–110. 5. Wallace LS, Keenum AJ, Roskos SE, Blake GH, Colwell ST, Weiss BD. Suitability and readability of consumer medical information accompanying prescription medication samples. Patient Educ Couns 2008;70:420–425. 6. Weiss BD, Mays MZ, Martz W, et al. Quick assessment of literacy in primary care: the newest vital sign. Ann Fam Med 2005;3:514–522. 7. von Wagner C, Steptoe A, Wolf MS, Wardle J. Health literacy and health actions: a review and a framework from health psychology. Health Educ Behav 2009;36:860–877. 8. Ickes MJ, Cottrell R. Health literacy in college students. J Am Coll Health 2010;58:491–498. 9. Howard DH, Gazmararian J, Parker RM. The impact of low health literacy on the medical costs
14.
15.
17.
18.
19.
20.
21.
22.
23.
24.
of Medicare managed care enrollees. Am J Med 2005;118:371–377. Morris NS, MacLean CD, Chew LD, Littenberg B. The Single Item Literacy Screener: evaluation of a brief instrument to identify limited reading ability. BMC Fam Pract 2006;7:21. Baker DW, Williams MV, Parker RM, Gazmararian JA, Nurss J. Development of a brief test to measure functional health literacy. Patient Educ Couns 1999;38:33–42. Parker RM, Baker DW, Williams MV, Nurss JR. The test of functional health literacy in adults: a new instrument for measuring patients’ literacy skills. J Gen Intern Med 1995;10:537–541. Mancuso JM. Assessment and measurement of health literacy: an integrative review of the literature. Nurs Health Sci 2009;11:77–89. Rowlands G, Khazaezadeh N, Oteng-Ntim E, Seed P, Barr S, Weiss BD. Development and validation of a measure of health literacy in the UK: the newest vital sign. BMC Public Health 2013;13:116. Salgado TM, Ramos SB, Sobreira C, et al. Newest Vital Sign as a proxy for medication adherence in older adults. J Am Pharm Assoc 2003;2013(53): 611–617. Peiravian F, Rasekh HR, Jahani Hashemi H, Mohammadi N, Jafari N, Fardi K. Drug literacy in Iran: the experience of using “The Single Item Health Literacy Screening (SILS) tool”. Iran J Pharm Res 2014;13:217–224. Aguirre AC, Ebrahim N, Shea JA. Performance of the English and Spanish S-TOFHLA among publicly insured Medicaid and Medicare patients. Patient Educ Couns 2005;56:332–339. Quinlan P, Price KO, Magid SK, Lyman S, Mandl LA, Stone PW. The relationship among health literacy, health knowledge, and adherence to treatment in patients with rheumatoid arthritis. HSS J 2013;9:42–49. Chen AM, Yehle KS, Albert NM, et al. Relationships between health literacy and heart failure knowledge, self-efficacy, and self-care adherence. Res Soc Adm Pharm 2014;10:378–386. Stagliano V, Wallace LS. Brief health literacy screening items predict newest vital sign scores. J Am Board Fam Med 2013;26:558–565. Patel PJ, Joel S, Rovena G, et al. Testing the utility of the newest vital sign (NVS) health literacy assessment tool in older African-American patients. Patient Educ Couns 2011;85:505–507. Osborn CY, Weiss BD, Davis TC, et al. Measuring adult literacy in health care: performance of the newest vital sign. Am J Health Behav 2007;31(suppl 1): S36–S46. UNESCO Institute for Statistics. UIS Statistics in Brief: Education (all levels) profile – Iraq (in 2011), 2013. UNESCO National Education Support StrategyRepublic of Iraq j 2010–2014, UNESCO Iraq Office,
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813 UNESCO National Education Support Strategy http:// www.unesco.org/new/fileadmin/MULTIMEDIA/ FIELD/Iraq/pdf/Publications/UNESS_2011 English (compressed).pdf; Accessed 11.05.14 25. KWINTESSENTIAL. Iraq – Language, Culture, Customs and Etiquette, http://www.kwintessential. co.uk/resources/global-etiquette/iraq.html; 2013. Accessed 15.07.14. 26. Wild D, Grove A, Martin M, et al. Principles of good practice for the translation and cultural adaptation process for Patient-Reported Outcomes (PRO) measures: report of the ISPOR Task Force for translation and cultural adaptation. Value Health 2005;8:94– 104. 27. Davis E, Davern M, Waters E, et al. CPQOL Translation Guidelines, Translating the CP QOL-Child &
28.
29.
30.
31.
813
CP QOL-Teen into languages other than English; 2013. p. 9–11. FORUM SG. Beyond binary outcomes: PROC logistic to model ordinal and nominal dependent variables. In: Elkin E, ed. Statistics and Data Analysis, http:// support.sas.com/resources/papers/proceedings12/ 427-2012.pdf; 2012. Accessed 09.10.14. Ozdemir H, Alper Z, Uncu Y, Bilgel N. Health literacy among adults: a study from Turkey. Health Educ Res 2010;25:464–477. VanGeest JB, Welch VL, Weiner SJ. Patients’ perceptions of screening for health literacy: reactions to the newest vital sign. J Health Commun 2010;15:402–412. Rothman RL, Housam R, Weiss H, et al. Patient understanding of food labels: the role of literacy and numeracy. Am J Prev Med 2006;31:391–398.
Original Research
Evaluation of instruments to assess health literacy in Arabic language among Iraqis Ali Azeez Al-Jumaili, M.S.a,b,*, Mohammed Dakhil Al-Rekabi, Ph.D.c, Bernard Sorofman, Ph.D.a a
University of Iowa, College of Pharmacy, Iowa City, IA, USA b University of Baghdad, College of Pharmacy, Baghdad, Iraq c University of Kufa, Faculty of Pharmacy, Al-Najaf, Iraq
Abstract Background: Low health literacy is associated with lack of medical information, less use of preventive measures, low medication adherence rates, high health care costs and high risk of hospitalization. Objective: The aims were to compare the results of the three health literacy tests, to measure for the first time the health literacy level of Iraqis, to describe the use of standardized health literacy tests, to evaluate reliability and validity of the Arabic versions of these tests, and to investigate whether there is relationship between the participant characteristics and the health literacy level. Methods: A convenience sample of 95 subjects was studied in five community pharmacies in Al-Najaf and Babylon governorates, Iraq. Three health literacy tests, the Single Item Literacy Screener (SILS), the New Vital Sign (NVS) and the Short version of the Test of Functional Health Literacy in Adults (S-TOFHLA), were translated in the Arabic language and administered to the pharmacy customers. Results: There were no statistically significant associations between age, gender, education and current education status and NVS score, but there were significant positive associations between the level of education and each one of SILS, New SILS, and S-TOFHLA scores. Conclusions: SILS has one subjective, possibly culturally biased question. Since Iraqis are generally not exposed to reading product labels, the NVS test might be not an accurate measure for them. S-TOFHLA was the most comprehensive test and gave equitable results. The Arabic version of S-TOFHLA can be used to measure health literacy in 22 Arabic speaking countries. Ó 2015 Elsevier Inc. All rights reserved. Keywords: Health literacy; SILS; NVS; S-TOFHLA; Iraq; Pharmacy
Introduction According to the U.S. National Assessment of Adult Literacy (NAAL), literacy is “the ability to
use printed and written information to function in society and to achieve one’s goals”.1 Health literacy is the ability of a person to read, compute,
The authors Ali Azeez Al-Jumaili and Mohammed Dakhil Al-Rekabi are joint first authors. * Corresponding author. University of Iowa, College of Pharmacy, PHAR S532, Iowa City, IA, 52242, USA. Tel.: þ1 319 936 4354 (mobile); fax: þ1 319 353 5646. E-mail address: [email protected] (A.A. Al-Jumaili). 1551-7411/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.sapharm.2015.02.002
804
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
and understand health-related information such as in a physician appointment slip, medication labels and pamphlets.2 Health literacy is not only basic literacy, but also knowledge of health-related topics. A limited literacy does not mean illiteracy, which is inability to read or write.3 The NAAL (2003) survey reported that 14% of the U.S. adults had below basic health literacy, only 12% had proficient health literacy, more than half (53%) had intermediate health literacy and 22% had basic health literacy.1 Elderly people age 65 and older had lower health literacy than younger adults.1 Literacy and health are linked. The average reading skills of American adults is between the seventh and eighth-grade levels.4 Health education materials require a reading level higher than the average reading skills of U.S. adults.5 Weiss and colleagues concluded that populations with lower literacy rates have poorer health status especially in non-industrialized countries.6 Low health literacy is associated with poor health-related outcomes, such as high risk of hospitalization.7 Low health literacy is associated with lack of medical information, impaired health care knowledge, less use of preventive measures, lower medication compliance rates and higher health care costs.8 Similarly, Howard and colleagues concluded that the medical costs of patients with inadequate health literacy were significantly higher than that for patients with adequate health literacy.9 There are several measures of health literacy; one is the Single Item Literacy Screener (SILS) developed by Morris and colleagues and asking about the amount of help that is needed to read and understand doctor or pharmacy-written instructions.10 The Short version of the Test of Functional Health Literacy in Adults (STOFHLA) developed by Baker and colleagues11 shortened the time required by the TOFHLA.12 S-TOFHLA is a standard for assessment of health literacy and takes 12 min.10,11 In 2009, a review article included 42 literacy studies and concluded that S-TOFHLA was considered as a standard test for the comparison of other health literacy tests.13 The New Vital Sign (NVS) test developed by Weiss and colleagues6 measures both reading and numeracy skills of subjects through reading an ice cream label contents and answering six related questions.6 The NVS is shorter but less comprehensive than the S-TOFHLA. SILS, NVS and S-TOFHLA are the most popular health literacy tests. Four of the health literacy studies that used TOFHLA, S-TOFHLA, NVS or NVS-UK stratified the health literacy
results according to the gender, age, and years of school education.8,11,14,15 A recent SILS study in Iran classified the health literacy results by age, education degree and socioeconomic status.16 Many studies investigated using one test (STOFHLA,8,12,16–19 NVS,20 or SILS16), and others compared two health literacy tests (NVS vs STOFHLA,6,11,14,21,22 NVS vs SILS15 or SILS vs S-TOFHLA10), but comparing the three most common tests (SILS, NVS and S-TOFHLA) was the major contribution of the present study. Instruments; standard measures of health literacy The Single Item Literacy Screener (SILS) SILS is a primary indication that participants have inadequate reading abilities and may request help to read health-related information.10 SILS was administered in written format: “How often do you ask someone for help to read the instructions and leaflets from a doctor or pharmacy?” A participant could choose one of the following answers (5-point Likert scale): 5-never, 4-rarely, 3-sometimes, 2-often or 1-always. If a participant chooses sometimes, often or always, that indicates a difficulty with reading of health materials. Never and rarely indicate adequate reading ability.10 Short version of the Test of Functional Health Literacy in Adults (S-TOFHLA) S-TOFHLA evaluates both the numeracy and reading skills of participants. The reading part has two prose passages while the numeracy section includes four questions that evaluate understanding of glucose monitoring, prescription labels and appointment slips.2,11 The reading sections of the S-TOFHLA test included this statement: “fill in the blanks using a word from a list which best completes the sentence grammatically and contextually.” The passages are related to preparation for an upper gastrointestinal X-ray and Medicaid rights & responsibilities.11 When the S-TOFHLA was translated to Arabic, one of the 36 cloze items was dropped from the passages because it did not make cultural sense after translation to Arabic. The dropped item was the third one in the X-Ray passage with four choices (is, am, if, it). “Medicaid” was translated as health care assistance for needy people. The two prose passages have a total of 35 cloze items totaling 70 points (two points per each item). The cloze process includes deleting every fifth to seventh word in a passage. The participant should choose the most appropriate answer to fill the blank from a list of four words.11 The score of
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
the four numeracy items is 30 points (7.5 points per each one). Thus, the total score for the whole S-TOFHLA test (both readings and numeracy items) would be 100 points.11 The score could be classified into one of three health literacy levels: 0–53 refers to inadequate functional health literacy; 54–66 indicates marginal health literacy, while 67–100 refers to adequate functional health literacy.11 The Newest Vital Sign (NVS) Participants read an ice cream label and then the interviewer read the six questions to the participant. The column of the content percentages on the ice cream label was deleted in this study as Rowlands and colleagues (2013) did in their study with the NVS-U.K14 because these percentages were not necessary to answer the questions. One point was given for each correct answer of the NVS test. The scores between zero and one indicate “probably inadequate” (limited) health literacy, two to three indicate “possibility inadequate” literacy, while four to six refer to “adequate” health literacy.6 This is a study comparing health literacy measures in an Iraqi population during 2013– 2014. In 2011, Iraq had a population of 32,665,000 with a life expectancy at birth of 69 years.23 The Iraqi Household Socio-Economic Survey of 2007 indicates that approximately 23% (5,193,682) of Iraqis were illiterate.24 According to the UNESCO Institute for Statistics, the adult (15 and older) literacy rate in Iraq was 78.5% (86.0% for male and 71.2% for female) in 2011.23 In 2010, Iraq had 6,635,771 students at different educational levels.24 More than three quarters (77%) of Iraqi people are native Arabic speakers.25 English is taught in elementary, middle, high schools and some colleges in Iraq. However, Iraqis do not use English in their daily communication except among medical staff and some academic faculty. Consequently, the majority of Iraqi people are not skilled English speakers. Most Iraqis prefer reading the Arabic version of drug labels and leaflets because they are more understandable to them. Therefore, the Iraqi Ministry of Health (MOH) requires that medication labels and leaflets are written in both Arabic and English. For the same reason, pharmacists and physicians use the Arabic language to counsel patients. Although many health literacy studies have been conducted in the U.S. and Western Countries for 20 years, no such study has been conducted in Iraq.
805
The aims of this study were to compare the results of the three common health literacy tests, to measure for the first time the health literacy level of Iraqis, to describe the use of standardized health literacy tests for Iraqis, to evaluate reliability and validity of the Arabic versions of these tests, and to investigate whether there is relationship between Iraqi participant characteristics and the reported health literacy scores on common measures developed in Western culture.
Methods Procedure The study was cross-sectional and descriptive. It surveyed a convenience sample of 95 Iraqi pharmacy customers from September 2013 to February 2014. Self-administered testing was conducted in five community pharmacies, three of them located in Al-Najaf and two pharmacies in Babylon governorates of Iraq. The pharmacists used Arabic translated SILS, S-TOFHLA and NVS tests. S-TOFHLA and NVS included both reading and quantitative sections which measured the ability to understand the numbers on an ice cream label (NVS), and appointment slips, prescription vials and blood glucose level (STOFHLA). The participants were pharmacy customers who came to pick prescriptions up or ask for over the counter medications for themselves or for a family member. Exclusion criteria were subjects (children) less than 15 years-old, health care professionals, subjects who had mental illness and subjects with vision impairment. All participants had some degree of literacy that enabled them to read the written literacy tests and answer the questions. All participants were Arabic native speakers. Each test required translation into formal Arabic. The translation used formal methodology by Wild (2005).26,27 Formal Arabic language can be commonly understood in 22 Arabic-speaking countries (in the Middle-east and North Africa). A pilot study with 25 subjects was conducted to assess any ambiguous statements or questions and to make sure the Arabic forms were understandable by pharmacy customers since these were the first Arabic versions of the three health literacy tests. The translation of NVS and STOFHLA was refined after the feedback from the pilot study. Pharmacists approached customers to ask them whether they would like to participate in
806
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
health literacy tests. After receiving verbal consent, the three tests were administered. The questions were given in the same order for all participants (SILS, NVS then S-TOFHLA). All the tests were answered at one time in the pharmacies. Basic demographic information (i.e., age, gender, educational level) was collected. No compensation was provided. Reliability analyses were conducted for the Arabic versions of S-TOFHLA and NVS tests using Cronbach’s alpha, which represents the internal consistency of the items, as performed by Baker and Weiss.6,11 Construct validity was also conducted by computing bivariate correlations (Pearson’s coefficient) between NVS and S-TOFHLA and among STOFHLA sections. The proposal was Institutional Review Board (IRB) exempt at the University of Iowa, USA and Ethical Committee approved in Iraq.
Statistical analyses Most analyses were performed using Statistical Analysis System (SAS, version 9.3, SAS Inc., Cary, North Carolina, USA). Various methods were used. Means, range and standard deviations (SD), frequencies and percentages of participant characteristics such as participants’ age, gender and educational degrees were measured. Because a cultural concern arose regarding the use of “sometimes,” sensitivity analysis was performed to compare the SILS results after moving each “sometimes” choice to the “adequate” literacy group, identified as New-SILS (NSILS). Fisher’s Exact test was performed due to small cell counts to compare the proportions of gender and education into the SILS, NSILS, NVS and STOFHLA test levels (Tables 1 and 2). For the Fisher’s Exact test, the five SILS and NSILS responses were classified into two levels (limited
Table 1 The results of SILS and New-SILS for each participant group A. Actual SILS results Participant characteristics
Gender Female Male Education (years) Less than middle school (!9) Middle school (9) High school (12) Some colleges (14–18) Total: N (total %)
Reading ability
Total N (row %)
Limited (Always, often, sometimes) N (row %)
Adequate (Rarely, never) N (row %)
30 (71.43) 42 (79.25)
12 (28.57) 11 (20.75)
42 (100) 53 (100)
6 15 11 40 72
0 3 6 14 23
6 18 17 54 95
(100) (83.33) (64.71) (74.07) (75.97)
(0.00) (16.67) (35.29) (25.93) (24.21)
(100) (100) (100) (100) (100)
B. Sensitivity analysis: New-SILS results after switching every “sometimes” answer to adequate category Participant characteristics
Gender Female Male a Education (years) Less than middle school (!9) Middle school (9) High school (12) Some colleges (14–18) Total: N (total %)
Limited (Always, often) N (row %)
Adequate (Sometimes, rarely, Never) N (row %)
Total N (row %)
17 (40.48) 21 (39.62)
25 (59.52) 32 (60.38)
42 (100) 53 (100)
6 8 5 19 38
0 10 12 35 57
6 18 17 54 95
(100) (44.44) (29.41) (35.19) (40.00)
(0) (55.56) (70.59) (64.81) (60.00)
(100) (100) (100) (100) (100)
Fisher’s Exact tests showed significant (P ! 0.05) relationship between level of education and NSILS result categories. a
807
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813 Table 2 The results of NVS and S-TOFHLA according to gender and education levels of the participants A. The results of NVS Participant characteristics
Health literacy level Probably inadequate (0–1 correct answer) N (row %)
Gender Female Male Education (years) Less than middle school (!9) Middle school (9) High school (12) Some colleges (14–18) Total: N (total %)
8 (19.51) 6 (13.33) 2 4 2 6 14
(40.00) (25.00) (11.76) (12.5) (16.28)
Total N (row %)
Possibly inadequate (2–3 correct answers) N (row %)
Adequate (4–6 correct answers) N (row %)
18 (43.90) 22 (48.89)
15 (36.59) 17 (37.78)
41 (100) 45 (100)
1 5 9 25 40
2 7 6 17 32
5 16 17 48 86
(20.00) (31.25) (52.94) (52.08) (46.51)
(40.00) (43.75) (35.29) (35.42) (37.21)
(100) (100) (100) (100) (100)
B. The results of S-TOFHLA Participant characteristics
Gender Female Male a Education (years) Less than middle school (!9) Middle school (9) High school (12) Some colleges (14–18) Total: N (total %)
Inadequate (Scores 0–53) N (row %)
Marginal (Scores 54–66) N (row %)
Adequate (Scores 67–100) N (row %)
Total N (row %)
4 (9.52) 4 (7.84)
4 (9.52) 4 (7.84)
34 (80.95) 43 (84.31)
42 (100) 51 (100)
2 1 2 3 8
2 3 1 2 8
2 14 14 47 77
6 18 17 52 93
(33.33) (5.56) (11.76) (5.77) (8.60)
(33.33) (16.67) (5.88) (3.85) (8.6)
(33.33) (77.78) (82.35) (90.38) (82.80)
(100) (100) (100) (100) (100)
a Fisher Exact test showed significant relationship (P ! 0.05) between education levels and S-TOFHLA score categories.
and adequate) (Table 1). On the other hand, each of NVS and S-TOFHLA had three result levels (inadequate, marginal and adequate) (Table 2). The significance level was 0.05 for Fisher tests and regression tests and 0.01 for bivariate correlation test. Ordinal logistic regression28 estimated the association between the SILS, NSILS and NVS results with the four independent variables: gender, age, education and current student status (whether participant was a current student or not). For the logistic regression analyses, the NVS had six categories (scores: 1–6) while SILS and NSILS had five categories. A generalized linear model (GLM) procedure was run for the continuous S-TOFHLA scores to measure the relationship between the explanatory variables: age (continuous), gender (categorical), educational degrees (categorical) and current student status (categorical) with the outcome variable, S-TOFHLA score.
Statistical Package for Social Sciences (SPSS) (IBM SPSS Statistics, Armonk, NY, USA) was used to measure the Pearson correlation (r) among the three health literacy tests, between numeric and reading sections of S-TOFHLA and between X-Ray and Medicaid-related cloze passages of S-TOFHLA. SPSS was also used to calculate the reliability analysis (Cronbach’s alpha) of both reading passages and numeric section of S-TOFHLA, and NVS questions. Two participants skipped the numeric sections of S-TOFHLA and they have not been included in the regression and Fisher test analyses. Other nine participants skipped NVS test and they have not been included in the NVS-related analyses.
Results Ninety-five subjects participated in the study, 42 females and 53 males. The average age of the
808
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
participants was 27.47 (9.58) with a range of (15–60) years old. The participants were classified into four groups according to their education degree, less than middle school (6.32%), completed middle school (18.95%), completed high school (17.89%), and two-year college, bachelor or master (56.84%). The results of SILS were categorized into either limited or adequate reading ability. More than three quarters (75.79%) of participants showed limited reading ability with a SILS answers: “always, often or sometimes.” The remaining (24.21%) had adequate reading ability since they answered “never or rarely” to the SILS question (Table 1A). Sensitivity analysis for SILS was conducted to examine the SILS results after switching every “sometimes” answer to the “adequate” group. The NSILS results showed that 40.0% of the participants had limited reading ability while 60.0% of them had adequate reading ability (Table 1B). Eighty-six participants had an NVS score average of 3.02 (1.74) out of six. The NVS test results showed that 16.3% of participants had a high probably of inadequate health literacy, 46.5% had possibly inadequate health literacy and 37.2% had adequate health literacy (Table 2A). In contrast, according to S-TOFHLA test, 8.6% had low literacy, 8.6% had marginal literacy and 82.8% of participants had adequate health literacy (Table 2B). The participants answered the prose passage (82%) more often correctly than the numeracy (75.4%) questions of the S-TOFHLA (Table 3). The X-Ray passage had higher literacy responses (85.9%) than the Medicaid equivalent rights and responsibilities passage (78.6%). The doctor appointment slipquestion had the lowest correct response (63.8%) while the blood sugar-level question had the highest correct response (79.7%) among the four numeracy S-TOFHLA questions. In this sample of Iraqi adults, the Fisher’s Exact test showed no significant associations between gender and education categories, and the NVS and SILS results (Tables 1 and 2). However, there were significant relationship (P ! 0.05) between education levels and NSILS and STOFHLA result categories according to the same test. The relationship between the four independent variables and each of SILS and NSILS responses and NVS scores were measured by using ordinal logistic regression. The logistic regression analysis indicated there were no significant associations
Table 3 Percentage of correct answers for S-TOFHLA portions
A 1
2
B
S-TOFHLA test items
Correct answers (%)
Total S-TOFHLA score Reading passages (N ¼ 95) Preparations for X-Ray 0–5 items (Preparing for X-ray instructions) 6–11 items (day before X-ray) 12–16 items (At X-ray day) Medicaid-related passage 1–7 items 8–13 items 14–19 items Numeracy items (N ¼ 93) Take medication every 6 h Blood sugar-level Doctor appointment slip Taking medication before lunch
79.12 81.99 85.9 88.8 94.6 90.2 78.6 80.0 70.2 86.0 75.4 76.49 79.7 63.8 73.8
The table format adapted from Baker.11
between the four independent variables and NVS scores (1–6). The logistic regression analysis indicated that there was significant positive relationship (P ! 0.05) between participant education level and the SILS and NSILS responses after controlling the other three explanatory variables. The participants with less than middle school had 11.1 times higher odds of choosing the lowest SILS response (always) compared to those with college education after controlling other variables (Table 4A).28 Similarly, Table 4B shows that participants with a less than middle school compared to those with college education had 12.8 times higher odds to choose the lowest NSILS response after controlling other variables.28 That means the participants with lowest education level (less than middle school) were 11.1 (for SILS) and 12.8 (for NSILS) times more likely to select the lowest SILS choice compared to those with the highest education level (some colleges) after controlling age and gender. Hence, there were positive relationships between education level and each of SILS and NSILS results while there were no significant relationships between the test responses and age, gender and current student status (Table 4). Likewise, multiple linear regression analysis showed that participants with some college education and middle school education had an average of 23.2 and 16.6 higher points (scores) in S-TOFHLA test respectively compared to
809
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813 Table 4 Odds ratios, 95% confidence limits & P-value of the SILS and NSILS logistic regression model results Effect (parameter)
Point 95% estimate Confidence (odds ratio) limits
A. SILS odds ratio estimates Age 0.975 Gender M vs F 1.384 Education 1 vs 4 11.094 Education 2 vs 4 1.247 Education 3 vs 4 0.732 B. NSILS odds ratio estimates Age 0.977 Gender M vs F 1.117 Education 1 vs 4 12.777 Education 2 vs 4 1.407 Education 3 vs 4 0.493
Pvalue
0.935 1.016 0.233 0.646 2.966 0.403 1.669 73.759 0.013a 0.460 3.378 0.665 0.271 1.972 0.537 0.936 1.020 0.298 0.511 2.444 0.781 1.904 85.733 0.009a 0.507 3.900 0.512 0.174 1.394 0.182
Each of the outcome variable (SILS or NSILS) had five levels. Education 1 ¼ less than middle school, Education 2 ¼ middle school, Education 3 ¼ high school, Education 4 ¼ some college degrees (high diploma, bachelor or master). Female and Education 4 were references for these analyses. a Ordinal Logistic regression analysis showed that significant positive association between choosing the lowest SILS & NSILS results (always) and lowest education level (P ! 0.05).
participants with less than middle school education. In other words, there was a positive significant (P ! 0.05) association between the educational level and the continuous S-TOFHLA scores. Additionally, the multiple linear regression analysis showed the not-current students had 9.87 fewer points than the S-TOFHLA scores of current students. The current student status had significant positive relationship with S-TOFHLA scores. Coefficient of determination of the regression model (R2) was 0.20, explaining 20% of the outcome (S-TOFHLA score) variance (Table 5). The time required to answer each test was recorded. The average time for the NVS Arabic version was 10.8 (7.13) with an outlier (participant who took 30 min) and 8.0 (6.73) minutes without this outlier. The English, Spanish and Turkish versions for NVS required average time 3.0 1.2, 3.4 1.2 and 6.3 1.27 min, respectively.6,29 On the other hand, the Arabic version of S-TOFHLA test required 11.1 (8.67) minutes in average which was almost the same as reported the English version time (12.0 min). SILS required less than a minute to be answered.
Table 5 Multiple regression model results of continuous STOFHLA scores with age, gender, education level and current student status Parameter
Estimate Standard t Pr O jtj error value
Intercept Age Gender M Gender F Education 2 (middle school) Education 3 (high school) Education 4 (some college) Education 1 (less than middle school) Not-current student Current student R2
72.588 0.194 1.482 0.000 16.615
7.902 0.209 3.413 7.977
2.08
0.0402a
11.826
8.024
1.47
0.1441
23.171
7.238
3.20
0.0019a
4.806
2.05
0.0431a
9.19 !0.0001 0.93 0.3574 0.43 0.6652
0.000
9.866 0.000 0.203
a Significant (P ! 0.05) positive association between the educational level and the continuous S-TOFHLA scores. The analysis references were female for gender, education-1 for education level and current student for current student status.
Cronbach’s alpha of the S-TOFHLA reading section was 0.89. However, Cronbach’s alpha of the S-TOFHLA numeric questions was 0.615. The Cronbach’s alpha of NVS was 0.69. Regarding validity, the numeric and reading sections of S-TOFHLA had significant positive correlation (r ¼ 0.34). Furthermore, there was a significant high positive correlation (0.724) between X-Ray and Medicaid cloze passages of STOFHLA. There was a significant (P ! 0.01) positive correlation (r ¼ 0.513) between NVS and total STOFHLA scores. Both numeric (r ¼ 0.414) and reading (r ¼ 0.351) section scores had significant positive correlation with NVS scores. There were weak positive correlations between SILS and STOFHLA (r ¼ 0.112) and very weak positive correlation between SILS and NVS (r ¼ 0.089 t).
Discussion This study used for the first time, to our knowledge, three common health literacy tests (SILS, NVS & S-TOFHLA) simultaneously and also for the first measured health literacy for Iraqis.
810
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
Exact Fisher tests showed there were statistically significant differences between each of the NSILS and S-TOFHLA result levels according to educational categories while non-significance difference in NVS and SILS levels according to both gender and educational categories. The sample size might not be large enough to detect differences that actually exist between the participant categories. Similarly, Ickes and Cottrel concluded there were no significant relationships between gender, college-year (juniors vs seniors) and TOFHLA scores after using Analysis of Variance (ANOVA) for analysis.8 Morris and colleagues reported that SILS is practical because it is very brief. However, it cannot detect people with limited reading ability (i.e., false negative result). SILS results in the present study indicated that majority of participants had limited reading ability (chose sometimes, often or always). An Iranian study used SILS and found that 30.3% of participants had inadequate health literacy.16 Again, Iraqis usually choose “sometimes” as a positive cultural preference. Many Iraqi participants (34.7%) chose “sometimes,” which culturally may be seen as a literate response, but actually it reflected limited reading ability on the test. The sensitivity analysis changed SILS results (Table 1B) showed that more than half the participants had adequate reading ability after switching “sometimes” to the “adequate” group. Hence, the SILS can be largely influenced by the cultural way of thinking. Both SILS and the NSILS results had a significant positive association with education level (P ! 0.05) (Table 4). Likewise, the logistic regression analysis of an Iranian study reported significant positive relationship between educational level and the SILS response.16 The logistic regression analysis showed there was no significant association between NVS scores and the four participants’ characteristics. The results of the present study matched the results of a Portuguese study that also found no significant association between NVS results and participants’ gender and education level.15 Similarly, the developers of NVS test indicated no significant difference in NVS scores according to gender (independent t-test) for both English and Spanish versions.6 In contrast the developers of the NVS-UK concluded weak correlations between NVS-UK scores and educational level (Pearson’s r ¼ 0.22).14 The NVS test results inferred that most of the participants had limited health literacy. Stagliano
and colleagues (2013) found that 17.8% of the American participants had inadequate literacy.20 According to the same study the mean of the NVS scores for American adults was 3.7 2.0 while the mean for the Iraqi adults was lower (3.02 1.74). Another study in the U.S. showed that mean NVS scores was 3.0 (1.9).21 The difference between Iraqi and American NVS scores might be because it is uncommon for Iraqis to read food labels. The mean of NVS score of this study (3.02 1.74) was relatively higher than the results (2.60 0.08) of a study in Turkey with a Turkish version of NVS.29 Salgado and colleagues used a Portuguese version of NVS and found that 95% of Portuguese elderly participants had limited health literacy with poor NVS average scores (0.81 0.1).15 They mentioned that Portugal has recent and limited experience with food labels.15 Additionally, they disclosed a weakness in the fifth question (“is it safe for you to eat this ice cream?”) structure. They found that question number-five was the question answered correctly most often. They argued that the answer to this dichotomous question depends mainly on the chance although the participants were unaware of the question because most those who answered this question correctly (not safe) failed to know the reason in the next question.15 The present study results agreed with VanGeest and colleagues who indicated that NVS was a sensitive but not specific test for low health literacy and that means NVS has the possibility of giving false positive results.30 NVS can show people have low health literacy, but in fact they have adequate health literacy.30 Rothman and colleagues concluded that even people with high literacy may misinterpret the content of nutrition labels.31 According to S-TOFHLA, 82.8% of Iraqi participants had adequate health literacy level. On the other hand, both developers of the English version of TOFHLA and S-TOFHLA found that slightly more than 50% of American participants had adequate health literacy according to those two tests.11,12 Thirty one percent taking the Spanish version of S-TOFHLA had adequate health literacy.12 Multivariate linear regression analysis showed a significant positive association between level of education, current student status and S-TOFHLA scores (Table 5). That indicated people with higher education usually have better health literacy. The multiple linear regression analysis also
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
showed the current student participants had STOFHLA scores 9.87 higher than those notcurrent students. Similarly, another study revealed that statistically difference in S-TOFHLA scores according to years of education.9 The coefficient of determination of the multiple linear regression analysis was relatively low (R2 ¼ 0.2). However, only two related studies conducted multivariate regression analyses for STOFHLA, and they did not report an R2-value to compare the results.9,11 Use of regression analysis was preferred over the bivariate correlation (Pearson or Spearman correlation coefficient) analyses used in another study,14 because bivariate correlation analysis does not control the effect of other covariates on the outcome variable (STOFHLA score). Salgado and colleagues found that there was an inverse relationship between age and NVS scores.15 The answer time for each test was reported in this study. The SILS question was answered within a very short time (a minute). The S-TOFHLA average time of response was similar to English form. However, the NVS test needed longer time than that reported in the literature because participants took the NVS test in Arabic and were unfamiliar with reading food label contents in their daily life. Because higher education in public universities is free in Iraq, most Iraqis pursue their college/ university study. Thus, more than half (56.8%) of the participants had some college degrees. The majority of the study participants were young (average 27.5 years). The main reason may be because those Iraqi elderly patients usually send a young family member to pharmacies to pick up medications since those young adults are more educated and can understand pharmacist counseling better. Another reason is that young adults usually serve their elderly relatives as a form of traditional norms and courtesy. Regarding to the instrument reliability, Cronbach’s alpha for the S-TOFHLA for Arabic version compared well to the English version.11 The Cronbach’s alpha for the Arabic version NVS was similar to the Spanish version and less than that of the English version.6 That means the Arabic version of S-TOFHLA readings and NVS had good internal consistency (reliability), but this was not the case of numeric section of S-TOFHLA. The validity was assessed by the correlations within and between the health literacy tests. The numeric section of S-TOFHLA had a significant
811
positive correlation with the reading section. Moreover, the Arabic version of S-TOFHLA had positive significant correlations between its reading sections and also between total STOFHLA scores and NVS scores. The significant positive Pearson correlations among the STOFHLA sections indicate that all the sections measure health literacy in the same direction. According to Rowlands et al, correlations (between S-TOFHLA and NVS significantly) larger than 0.3 are considered acceptable.14 Because the Arabic versions of S-TOFHLA and NVS had significant correlation higher than 0.3, they had acceptable validity. Limitations This was the first study in health literacy in Iraq, and that added challenges to conduct the research. The selection of participants was a convenience sample rather than randomized sample. Hence, the results are not generalizable and may not apply for all pharmacy customers in Iraq. The participation was voluntary and without compensation. Pharmacy customers who refused to participate might have low health literacy, so they avoided taking tests that could embarrass them. Random sampling would be recommended for future studies.
Conclusion The study participants scored higher on STOFHLA compared to SILS and NVS tests. Cultural differences could influence the scores of SILS and NVS tests. The SILS is a brief, but subjective test and may not reflect the accurate level of literacy because it has a subjective and possibly culturally biased question. The NVS might be not applicable as a health literacy test for people who are not used to reading product labels in their daily lives such as Iraqis. The S-TOFHLA was the most convincing test because it was neither subjective nor dependent on the familiarity with product labels, and almost all participants (n ¼ 93) completed this test. Hence, if we would conduct a national health literacy survey for Iraqis in the future, we would choose the S-TOFHLA test. In this sample of Iraqi participants, there were no significant differences in NVS according to age, gender, education and current student status. However, there were significant positive associations
812
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813
between the level of education and each of the SILS, NSILS and S-TOFHLA results indicating higher educated people might have higher health literacy level. In general, the Arabic version of S-TOFHLA and NVS had acceptable reliability (Cronbach’s alpha) and validity. Since the Arabic version of STOFHLA test gave equitable results in Iraq, it can be used to measure health literacy in 22 Arabic speaking countries.
10.
11.
12.
Acknowledgments We would like to thank Iraqi pharmacists who helped us to conduct the literacy tests in the five community pharmacies were Zaid Ali, Abdelfadhel J, Ali Shlah, Karrar M, and Ali Mohammed. We also appreciate Jacob Simmering for his advice for part of the analysis. We would like to acknowledge Iraqi Higher Committee for Educational Development (HCED) for supporting one of the authors (Ali Al-Jumaili) with a scholarship.
13.
References
16.
1. U.S. Department of Education IOES, National center for Education Statistics. The Health Literacy of America’s Adults, Results from the 2003 National Assessment of Adult Literacy. 2006 ed., http://nces. ed.gov/pubs2006/2006483.pdf; 2006. Accessed 04. 07.14. 2. Tkacz VL, Metzger A, Pruchnicki MC. Health literacy in pharmacy. Am J Health Syst Pharm 2008;65: 974–981. 3. Adult Literacy in America. Vol Third Edition. 2002 ed. U.S Department of Education-Office of Educational Research and Improvement- National Center for Educational Statistics, https://nces.ed.gov/pubs93/ 93275.pdf; 2002. Accessed 02.09.14. 4. Heilman CB. Readability of patient education materials. World Neurosurg 2013;80:e109–110. 5. Wallace LS, Keenum AJ, Roskos SE, Blake GH, Colwell ST, Weiss BD. Suitability and readability of consumer medical information accompanying prescription medication samples. Patient Educ Couns 2008;70:420–425. 6. Weiss BD, Mays MZ, Martz W, et al. Quick assessment of literacy in primary care: the newest vital sign. Ann Fam Med 2005;3:514–522. 7. von Wagner C, Steptoe A, Wolf MS, Wardle J. Health literacy and health actions: a review and a framework from health psychology. Health Educ Behav 2009;36:860–877. 8. Ickes MJ, Cottrell R. Health literacy in college students. J Am Coll Health 2010;58:491–498. 9. Howard DH, Gazmararian J, Parker RM. The impact of low health literacy on the medical costs
14.
15.
17.
18.
19.
20.
21.
22.
23.
24.
of Medicare managed care enrollees. Am J Med 2005;118:371–377. Morris NS, MacLean CD, Chew LD, Littenberg B. The Single Item Literacy Screener: evaluation of a brief instrument to identify limited reading ability. BMC Fam Pract 2006;7:21. Baker DW, Williams MV, Parker RM, Gazmararian JA, Nurss J. Development of a brief test to measure functional health literacy. Patient Educ Couns 1999;38:33–42. Parker RM, Baker DW, Williams MV, Nurss JR. The test of functional health literacy in adults: a new instrument for measuring patients’ literacy skills. J Gen Intern Med 1995;10:537–541. Mancuso JM. Assessment and measurement of health literacy: an integrative review of the literature. Nurs Health Sci 2009;11:77–89. Rowlands G, Khazaezadeh N, Oteng-Ntim E, Seed P, Barr S, Weiss BD. Development and validation of a measure of health literacy in the UK: the newest vital sign. BMC Public Health 2013;13:116. Salgado TM, Ramos SB, Sobreira C, et al. Newest Vital Sign as a proxy for medication adherence in older adults. J Am Pharm Assoc 2003;2013(53): 611–617. Peiravian F, Rasekh HR, Jahani Hashemi H, Mohammadi N, Jafari N, Fardi K. Drug literacy in Iran: the experience of using “The Single Item Health Literacy Screening (SILS) tool”. Iran J Pharm Res 2014;13:217–224. Aguirre AC, Ebrahim N, Shea JA. Performance of the English and Spanish S-TOFHLA among publicly insured Medicaid and Medicare patients. Patient Educ Couns 2005;56:332–339. Quinlan P, Price KO, Magid SK, Lyman S, Mandl LA, Stone PW. The relationship among health literacy, health knowledge, and adherence to treatment in patients with rheumatoid arthritis. HSS J 2013;9:42–49. Chen AM, Yehle KS, Albert NM, et al. Relationships between health literacy and heart failure knowledge, self-efficacy, and self-care adherence. Res Soc Adm Pharm 2014;10:378–386. Stagliano V, Wallace LS. Brief health literacy screening items predict newest vital sign scores. J Am Board Fam Med 2013;26:558–565. Patel PJ, Joel S, Rovena G, et al. Testing the utility of the newest vital sign (NVS) health literacy assessment tool in older African-American patients. Patient Educ Couns 2011;85:505–507. Osborn CY, Weiss BD, Davis TC, et al. Measuring adult literacy in health care: performance of the newest vital sign. Am J Health Behav 2007;31(suppl 1): S36–S46. UNESCO Institute for Statistics. UIS Statistics in Brief: Education (all levels) profile – Iraq (in 2011), 2013. UNESCO National Education Support StrategyRepublic of Iraq j 2010–2014, UNESCO Iraq Office,
Al-Jumaili et al. / Research in Social and Administrative Pharmacy 11 (2015) 803–813 UNESCO National Education Support Strategy http:// www.unesco.org/new/fileadmin/MULTIMEDIA/ FIELD/Iraq/pdf/Publications/UNESS_2011 English (compressed).pdf; Accessed 11.05.14 25. KWINTESSENTIAL. Iraq – Language, Culture, Customs and Etiquette, http://www.kwintessential. co.uk/resources/global-etiquette/iraq.html; 2013. Accessed 15.07.14. 26. Wild D, Grove A, Martin M, et al. Principles of good practice for the translation and cultural adaptation process for Patient-Reported Outcomes (PRO) measures: report of the ISPOR Task Force for translation and cultural adaptation. Value Health 2005;8:94– 104. 27. Davis E, Davern M, Waters E, et al. CPQOL Translation Guidelines, Translating the CP QOL-Child &
28.
29.
30.
31.
813
CP QOL-Teen into languages other than English; 2013. p. 9–11. FORUM SG. Beyond binary outcomes: PROC logistic to model ordinal and nominal dependent variables. In: Elkin E, ed. Statistics and Data Analysis, http:// support.sas.com/resources/papers/proceedings12/ 427-2012.pdf; 2012. Accessed 09.10.14. Ozdemir H, Alper Z, Uncu Y, Bilgel N. Health literacy among adults: a study from Turkey. Health Educ Res 2010;25:464–477. VanGeest JB, Welch VL, Weiner SJ. Patients’ perceptions of screening for health literacy: reactions to the newest vital sign. J Health Commun 2010;15:402–412. Rothman RL, Housam R, Weiss H, et al. Patient understanding of food labels: the role of literacy and numeracy. Am J Prev Med 2006;31:391–398.