Prevalence of Limited Health Literacy Among Irish Adults

Abstract

The authors conducted 2 health literacy investigations in Cork, Ireland. Study 1 was undertaken in 5 community pharmacies and the outpatient department of 2 urban hospitals and assessed patients’ health literacy skills using the Rapid Estimate of Adult Literacy in Medicine (REALM) and the Newest Vital Sign (NVS). Study 2 took place in 1 outpatient department and evaluated health literacy using the REALM and the short form of the Test of Functional Health Literacy.in Adults (S-TOFHLA). The authors collected data relating to age, gender, ethnicity, and educational level achieved. All subjects were asked items relating to their ability to use health-specific materials. A total of 1,759 people (61.2% female) completed either Study 1 (n = 1,339) or Study 2 (n = 420). Limited health literacy ranged from 18.4% (REALM) and 57.2% (NVS) in Study 1, and 21.9% (REALM) and 14.1% (S-TOFHLA) in Study 2 and was associated with increased age and lower educational attainment across all three tools (p < .001). Patients with limited health literacy were significantly more likely to report problems with using health materials received from a doctor or pharmacist. At minimum, 1 in 7 Irish adults were found to have limited health literacy, which may affect their ability to promote, protect, and manage health. As in the United States and in the United Kingdom, improving health literacy should be a public health objective for Ireland.

The Organisation for Economic Co-operation and Development's International Adult Literacy Survey found that one in four adults surveyed in Ireland in 1994 did not display the literacy skills and confidence needed to take part effectively in society (National Adult Literacy Agency, 2002). Furthermore, research predominantly in the United States has shown that people with inadequate literacy find health-related documents such as appointment cards, consent forms and prescriptions difficult to read and understand. Health literacy is defined as “the degree to which individuals can obtain, process and understand the basic health information and services they need to make appropriate health decisions” (Institute of Medicine, 2004). Recent findings in health literacy research suggest that there is a direct link between individual health literacy and health outcomes (Institute of Medicine, 2004). It has been hypothesized that inadequate literacy may have a direct, negative effect on health (Dewalt & Pignone, 2005). Patients with limited health literacy may have less health knowledge (Gazmararian, Williams, Peel, & Baker, 2003), worse self-management skills (Schillinger et al., 2002), lower use of preventative services (Miller, Brownlee, McCoy, & Pignone, 2007), higher hospitalization rates (Baker, Parker, Williams, & Clark, 1998), worse self-rated health (Baker, Parker, Williams, Clark, & Nurss, 1997), and higher rates of mortality (Sudore et al., 2006). In addition, those with literacy problems may also have poorer medication adherence (Gazmararian et al., 2006) and greater difficulty understanding prescription drug labels (Davis et al., 2006). Overall, individual literacy skills have repeatedly demonstrated to be a stronger predictor of health status than age, income, employment status, education level, and racial or ethnic group (Wolf et al., 2007).

Research in health literacy is for the most part in its infancy outside of North America, while there is increasing interest throughout the European Union, Australia, and elsewhere abroad. The need for greater research on the problem, particularly in Ireland, is highlighted in a NALA report from 2002 titled, “Health Literacy, Policy and Strategy” (National Adult Literacy Agency, 2002). The World Health Organization has highlighted the importance of improving health literacy as a public health goal, although the nature and extent of the problem should first be articulated locally within each country to optimize the development of any public health response. The aim of the present study was to estimate population health literacy levels, by administering the most commonly used measures of health literacy among a sample of adults in Cork, Ireland, and to determine any associations with age, gender, education, as well as one's ability to use health materials. Our findings are intended to provide an initial prevalence of the scope of limited health literacy to inform the development of future interventions.

Method

Ethical approval was granted for the project from the Clinical Research Ethics Committee of the Cork Teaching Hospitals.

Setting and Study Participants

This research comprised two cross-sectional substudies conducted in Cork, Ireland, one in January and July of 2009 (Study 1) and the other in January of 2011 (Study 2). Study 1, comprising 1,339 participants 18 years of age and older, was conducted in the outpatient department of two urban teaching hospitals and in a community sample of five pharmacies. Study 2 was also conducted in the outpatient department of one urban teaching hospital. In the hospital setting, patients and/or family members attending the outpatient department were approached by trained research assistants. In the community setting, any person entering the pharmacy was approached by the research assistant. In both cases, the purpose of the study was explained and verbal consent was obtained. Exclusion criteria for the study were any significant visual or hearing impairment based on self-report, or if English was not a patient's first language. The format of the interviews involved the collection of demographic and educational information followed by literacy assessments. The order of health literary assessments always commenced with the REALM screening tool followed by either the NVS or the S-TOFHLA depending on the study.

Health Literacy Measures

Rapid Estimate of Adult Literacy in Medicine (REALM)

The REALM is a word-recognition test comprising 66 health-related words that are arranged in order of increasing difficulty (Davis et al., 1993). Patients are asked to read aloud as many words as they can. The REALM tests the ability to recognize and pronounce words and uses the correct pronunciation as a surrogate measure of understanding. Scores are based on the total number of words pronounced correctly, with and Irish standard for pronunciation for scoring and interpreted as limited (0–60) or adequate (61–66).

The Test of Functional Health Literacy in Adults (TOFHLA)

We measured functional health literacy and numeracy with the shortened version of the TOFHLA. This test uses health-related materials such as prescription bottle labels and appointment slips in a 12-min test, with 36 reading comprehension items in two passages and four numeracy items (Baker, Williams, Parker, Gazmararian, & Nurss, 1999). The passages use a modified Cloze procedure where every fifth to seventh word is omitted and subjects select the correct word from among a set of four options. The S-TOFHLA was used for this research and collected in Study 2. A few minor changes were made to the S-TOFHLA in the terminology used so as to reflect Irish vernacular (e.g., “jelly” replaced with “jam”). Each of the 4 numeracy items was assigned a score of 7 points (28 possible points for the numeracy section) and 2 points each for the 36 Cloze items (72 possible points for reading comprehension). Total scores range from 0 to 100 and are classified into inadequate (0–53), marginal (54–66), and adequate (67–100) categories. For the purposes of this study, inadequate and marginal categories were combined resulting in a score of 0–66 for limited and a score of 67–100 as adequate health literacy.

The Newest Vital Sign (NVS)

The NVS consists of a nutrition label from an ice-cream container which is presented to the patient and to which they refer to during the test (Weiss et al., 2005). The patient is asked 6 questions in relation to the information on the label and scores range from 0 to 6. Total scores can be classified into limited (0–1), possibility of limited (2–3), and adequate (4–6) literacy. For the purposes of this study, a score of 0–3 was said to represent limited health literacy while a score of 4–6 was viewed as adequate health literacy (see Table 1).

Table 1. Health literacy categories according to three screening tools

	Health literacy category
Health literacy assessment	Limited	Adequate
REALM	0–60	61–66
S-TOFHLA	0–66	67–100
NVS	0–3	4–6

Difficulty with Health Materials

Participants were asked to answer the following four questions:

1.	Are medical forms difficult to fill out?
2.	Are medical booklets or leaflets difficult to understand?
3.	Are instructions on tablets from the chemist hard to understand?
4.	Do you ever need help to read information or filling in forms from doctors, nurses, or hospitals? (Ibrahim et al., 2008)

Response options included never, sometimes, often, and always, which were dichotomized into those who positively endorsed the questions (sometimes or often or always) compared with those who answered never.

Analysis Plan

Demographic characteristics were compared by study and setting (community vs. hospital) using chi-square tests. Spearman correlations were used to assess the associations between total scores on the three literacy measures. Percentages and exact 95% CIs of those with limited literacy were calculated for each measure. The proportion that positively endorsed (sometimes, often, or always) each screening question by literacy group was calculated and compared using chi-square tests. Data manipulation and analyses were conducted using Stata/MP version 11.0 (StataCorp, College Station, TX).

Results

Overall, a total of 2,361 people were approached to participate in these studies. Of these, 314 declined and 288 were ineligible to participate according to the exclusion criteria. Excluding those who were ineligible resulted in a response rate of 84.9% (1759/2073) across the two studies. The characteristics of the study sample are shown in Table 2. Those in the community setting tended to be younger, more highly educated, and had more adequate health literacy than did those recruited from the hospitals (all ps<.001).

Table 2. Characteristics of study sample, by study and site

		Study 1
Characteristic	All participants (N = 1,759)	Community (n = 261)	Hospital (n = 1,078)	Study 2 Hospital (n = 420)
Age, %
18–30 years	18.3	36.4	17.4	9.1
31–45 years	27.8	29.5	30.1	21.0
46–60 years	31.0	23.4	28.6	42.1
>60 years	22.9	10.7	23.9	27.9
Female, %	61.2	64.0	60.4	61.7
Education, %
Primary/during secondary	19.0	8.4	20.5	21.7
Junior/intermediate certificate	19.3	10.7	19.3	24.8
Leaving certificate/postcourse	35.8	25.7	37.9	36.4
Third level (degree/postgrad)	25.9	55.2	22.3	17.1

Total scores on the REALM were moderately correlated with total scores on the S-TOFHLA (r = 0.51) and the NVS (r = 0.49). Depending on the health literacy tool used, the proportion of participants with limited health literacy varied: 18.4% (95% CI [16.4, 20.6]) for the REALM and 57.2% (95% CI [54.6, 59.9]) for the NVS in Study 1 and 21.9% (95% CI [18.0, 26.2]) for the REALM and 14.1% (95% CI [10.7, 17.4]) for the S-TOFHLA in Study 2. Limited health literacy was associated with increased age and lower educational attainment across all three tools used (p < .001). Male gender was associated with limited health literacy for the REALM tool only (p < .001).

Responses to the four questions on participants’ experiences with health materials, according to health literacy status, are shown in Table 3. For the REALM and NVS measures in Study 1, those with limited literacy were significantly more likely to report difficulty in all four scenarios. In Study 2, participants defined as having limited literacy with the REALM were significantly more likely to report difficulty in three of the four scenarios; participants defined as having limited literacy with the S-TOFHLA were significantly more likely to report difficulty in one of the four scenarios.

Table 3. Screening questions, by literacy level

	Study 1 (n = 1,339)						Study 2 (n = 420)
	NVS			REALM			REALM			S-TOFHLA
	Limited 57.2% ^a	Adequate 42.8% ^a	p	Limited 18.4% ^a	Adequate 81.5% ^a	p	Limited 21.9% ^a	Adequate 78.1% ^a	p	Limited 14.1% ^a	Adequate 85.9% ^a	p
Are medical forms difficult to fill out?	45.5	29.3	<.001	55.3	34.7	<.001	58.7	42.1	.01	45.8	45.7	.99
Are medical booklets or leaflets difficult to understand?	33.8	20.3	<.001	42.7	24.7	<.001	48.9	31.7	.002	35.6	35.5	.98
Are instructions on tablets from the chemist hard to understand?	13.1	9.4	.04	15.9	10.5	.02	14.1	11.6	.51	13.6	11.9	.72
Do you ever need help to read information or filling in forms from doctors, nurses or hospitals?	25.2	13.8	<.001	41.7	15.5	<.001	38.0	19.2	<.001	33.9	21.6	.04

Discussion

Prevalence

The prevalence of limited health literacy, in this population sample, varied, dependent on the tool used from 18.4% for REALM to 57.2% for NVS in Study 1 and 14.1% for S-TOFHLA to 21.9% for REALM in Study 2. This contrasts with findings in Australia, which found 10.6% for REALM, 6.8% for TOFHLA, and 26.0% for NVS (Barber et al., 2009). However, it should be noted that this Australian study used the full TOFHLA, so the findings are not exactly comparable. Pooled analysis of data from a systematic review of U.S. health literacy prevalence studies reported a weighted prevalence of low health literacy of 26% (95% CI [22%, 29%]) and marginal health literacy of 20% (95% CI [16%, 23%]; Paasche-Orlow, Parker, Gazmararian, Nielsen-Bohlman, & Rudd, 2005). Most studies in this review used either REALM or versions of the TOFHLA. UK data, in a coronary heart disease population, using the REALM tool, found a prevalence of 19% of limited health literacy (Ibrahim et al., 2008). However, the threshold cutoff used in that study was lower than what is recommended for the REALM (defined as 59 and below vs. 60 and below).

Demographic Associations

The present study demonstrated associations between limited health literacy and educational level and age, irrespective of which tool was used. This is in line with findings reported in the literature (Schillinger et al., 2002). Female gender was associated with adequate health literacy only when measured by the REALM tool for our study, while this in line with findings from Barber and colleagues (2009), it contrasts with other studies that have found no association between limited health literacy and gender (Paasche-Orlow et al., 2005).

Implications for Health Care Professionals and Policymakers

This is the first study to examine the prevalence of limited health literacy in an Irish setting. The conservative estimate of 1 in 7 participants having limited health literacy highlights the significant challenges facing health care professionals and policymakers. This is made all the more worrying as we evolve into a society where greater emphasis is placed upon the patient assuming a more active role in his or her self-care. In doing so, it is assumed that patients will have a minimum standard of health literacy, an assumption which is not supported by our findings. Previous studies have shown that limited health literacy has a strong, independent association with mortality even after adjusting for an extensive set of covariates, including sociodemographic characteristics, chronic conditions, and detailed measures of baseline physical and mental health (Baker et al., 2007). Physicians commonly overestimate patients’ literacy levels, and only rarely consider limited literacy skills in their assessment of patient understanding (Bass, Wilson, Griffith, & Barnett, 2002). This lack of a health literacy focus in the patient consultation is all the more surprising when one considers the implications of limited health literacy on patient outcomes. The responsibility for recognizing limited health literacy lies with every health care professional and it is only when we are aware of these issues, that we can develop strategies to address the problem. To achieve this goal, appropriate teaching methods with an emphasis on the importance of interactive communication should be embedded in the undergraduate courses for all health care professionals.

Measurement of Health Literacy

The three most common health literacy screening tools were used, all of which have been widely documented in the literature. The prevalence of limited health literacy varied considerably depending on the tool used. Because the REALM and NVS differ fundamentally in their format and content, it is not surprising that patients obtain greatly differing scores, a finding that has been reported elsewhere (Osborn et al., 2007). The REALM, as a word recognition test, does not consider the more complex cognitive functions required for the successful interpretation of the NVS. Similar to the NVS, the S-TOFHLA, tests for numeracy and comprehension skills in a health care context. Nonetheless, the REALM appears to have been the most clearly predictive of the self assessment items used in this project.

Strengths and Limitations

A response rate of nearly 85% was achieved, which is higher than that of similar studies conducted in this area (Barber et al., 2009; Paasche-Orlow et al., 2005); we attribute this to the face-to-face nature of the participant recruitment. Those who declined may possess lower literacy levels (Baker et al., 1996), and thus, the prevalence found may underestimate actual prevalence; however, we do not believe that this would have contributed significantly because of the high response rate achieved.

The large sample size of this study, in comparison with others reported in the literature (Barber et al., 2009; Paasche-Orlow et al., 2005) allows for increased confidence in the precision of estimates of limited health literacy for each screening tool used. However, a population-based sampling method would clearly provide a more accurate estimate. The sample size and diversity of settings have helped to increase the generalizability of the findings to the wider population despite the fact that recruitment was confined to one geographical area of Ireland. In view of the mismatch between the REALM, TOFHLA, and NVS observed in previous studies (Barber et al., 2009), we included all three in our study design to capture the range of estimates of limited health literacy. In addition, although it seems unlikely, the nonrandomized selection of participants raises the possibility of selection bias.

Conclusions

At minimum, 1 in 7 Irish adults were found to have limited health literacy, which may affect their ability to promote, protect, and manage health. As in the United States and in the United Kingdom, improving health literacy should be a public health objective for Ireland.

Acknowledgments

The authors thank the research assistants Hilary Gallwey, Elaine Hanley, Zoe Mannix, Peter McNally, Odhran O'Donoghue, Michelle O' Driscoll, and Susan Spillane, staff at the research sites and participants of the surveys.

Notes

Note. NVS = Newest Vital Sign; REALM = Rapid Estimate of Adult Literacy in Medicine; S-TOFHLA = short form of the Test of Functional Health Literacy in Adults.

Note. NVS = Newest Vital Sign; REALM = Rapid Estimate of Adult Literacy in Medicine; S-TOFHLA = short form of the Test of Functional Health Literacy in Adults.

^a Percentage of respondents who reported any difficulty (sometimes, often, or always).