Health Literacy and Socioeconomic Characteristics among Older People in Transitional Kosovo

Aims: Health literacy among older people has received little attention in transitional countries of Southeast Europe. Our aim was to assess the level and socioeconomic correlates of health literacy among older people in Kosovo, a post-war country in the Western Balkans. Study Design: Cross-sectional study. Place and Duration of Study: Kosovo, between January-March 2011. Methods: This nationwide survey, conducted in Kosovo in 2011, included 1753 individuals aged ≥65 years (886 men, 867 women; mean age 73.4±6.3 years; response rate: 77%). Participants were asked to assess, on a scale from 1 to 5, their level of difficulty with regard to access, understanding, appraisal, and application of health information. Subscale scores and an overall health literacy score were calculated for each participant. Information on socioeconomic characteristics was also collected. Results: Subscale scores of health literacy were strongly correlated with each-other Research Article British Journal of Medicine & Medical Research, 3(4): 1646-1658, 2013 1647 (range of Spearman’s rho: 0.8-0.9). Mean values of the overall health literacy scores were significantly higher in men, urban residents, married individuals, the highly educated, and the better off participants. Conclusions: This may be the first report from the Western Balkans addressing health literacy in a population-based sample. Future studies in Kosovo and other settings in the region should provide further insight into the magnitude and socioeconomic determinants of health literacy which is an under-researched topic in countries of Southeast Europe.


INTRODUCTION
Access to better information is required to support people's participation and enable them making their own health choices [1]. The decision-making process is impacted by people's health competencies, which is linked to literacy, and entails the knowledge, motivation and competence to access, understand, appraise and apply information to make decisions in everyday life in terms of healthcare, disease prevention, and health promotion during the course of life. Various personal characteristics, demographic and social factors may as well have an impact on health literacy [2]. There are indications that low literacy leads to marked variation in an individual's ability to obtain relevant health information, and in their opportunity and capability to apply the information in interactions with health professionals and health care services [3,4]. Consequently, low health literacy may lead to worse health outcomes, ranging from worse self-rated health status, longer hospitalization and higher use of healthcare services resulting in higher healthcare costs [2,5], difficulties to follow medical instructions [6][7], impaired ability to navigate the health system [8] and lower participation in screening programs [9].
Health literacy and its association with socio-demographic and socioeconomic factors have been mainly studied in USA and Canada and more recently in Australia, Asia and Europe [2,10].
On the other hand, data on health literacy in former communist countries of the Western Balkans including Kosovo are scarce. Kosovo is the newest state in Europe struggling to establish a functional democracy after the breakdown of former Yugoslavia and the subsequent war in the region. In the framework of a population-based survey, our aim was to assess the level and socioeconomic correlates of health literacy among older people in Kosovo in terms of accessing, understanding, appraising and applying the information related to health care, disease prevention and health promotion.

Study Population
A nation-wide cross-sectional study among individuals aged 65 years or older was conducted in Kosovo in 2011. A population-based sample of 2400 individuals aged ≥65 years was drawn based on the 2010 lists (sampling frame) available from the Kosovo Ministry of Labour and Social Welfare [11]. Twelve strata were established (based on sexstratification [men vs. women], place of residence [urban vs. rural areas] and agestratification [65-74 years, 75-84 years and ≥85years]). A simple random sample of 200 individuals in each of the twelve strata was drawn [11]. Of the initial 2400 individuals targeted for inclusion, 135 participants were ineligible and further 375 individuals refused to participate, leading to 1890 study participants [11]. Of these, 137 participants were excluded from the current analysis due to incomplete information regarding health literacy. Therefore, this report is based on 1753 individuals, with an overall response rate of 77.4% (1753/2265).

Data Collection
A structured interviewer-administered questionnaire (including 25 items) was used to assess four dimensions of health literacy: access (5 items), understanding (7 items), appraisal (8 items) and application (5 items) of health information in three different situations/domains: health promotion, disease prevention and cure of disease. The health literacy instrument employed in the current study was developed in the framework of a large EU supported project [2].
Participants were asked to assess, in a scale ranging from 1 (unable -implying least health literacy score) to 5 (without any difficulty -maximal health literacy score), their level of difficulty with regard to access/understanding/appraisal/application of health information.
The health literacy instrument was pre-tested in a sample of older people (N=38) attending primary health care services in Kosovo and Albania before conducting the current survey.
A full version of the 25-item instrument used for the assessment of health literacy in our study is presented in Appendix 1.
An overall health literacy score (overall index) was calculated for each participant ranging from 25 (least health literacy score) to 125 (maximal health literacy score). In addition, four subscale scores (domain indexes) were calculated in line with the four domains explored namely: access (range: 5-25), understanding (range: 7-35), appraisal (range: 8-40) and application (range: 5-25) of health information.
In addition, we standardized the overall health literacy index in our sample with the overall health literacy score pertinent to the Test of Functional Health Literacy in Adults (TOFHLA) in order to compare our findings with previous studies [12]. Information on demographic factors (age and sex) and socioeconomic characteristics [place of residence (urban areas vs. rural areas), marital status (dichotomized into: married vs. not married), educational level (years of completed formal schooling), and self-perceived poverty (dichotomized into: not poor vs. poor)] was also collected.

Statistical Analyses
Age-sex and place-of-residence standardized/weighted percentages and their respective 95% confidence intervals (95%CIs) were calculated for the socioeconomic characteristics of study participants.
Cronbach's alpha, used to assess the internal consistency of the health literacy instrument, ranged from 0.90 to 0.94 for the subscale scores and the overall health literacy score.
Mann-Whitney test was used to compare mean values of health literacy scores by different categories of demographic and socioeconomic characteristics.
Spearman's correlation coefficient was used to assess the linear association between health literacy indexes (subscale scores).
General linear model was used to assess the association between the overall health literacy index and socio-demographic and socioeconomic factors. Age-adjusted and multivariableadjusted mean values and their respective 95%CIs of the overall health literacy score according to different categories of the socioeconomic characteristics were calculated. SPSS, version 15.0 was used for all the statistical analyses.

RESULTS AND DISCUSSION
Mean age of participants (54% women) was 73.4±6.3 years. On average, participants had 4.5 years of formal education, 62% resided in rural areas, and 48% regarded themselves as poor (Table1).
Mean overall and subscale health literacy scores were all significantly higher in men, urban residents, married individuals, among those who had at least one year of formal schooling and the better off participants (P<0.001 for all) [ Table 2].
Scores of health literacy domains/indexes were highly and significantly correlated with eachother (Spearman's rho ranged from 0.8 to 0.9) [ Table 3].
Age, sex, place of residence, education level, and self-perceived poverty, except marital status, were significant "predictors" of the overall health literacy score in unadjusted and multivariable-adjusted general linear models (Table 4). In multivariable-adjusted analysis, men and the "younger" participants reported a significantly higher mean health literacy score compared, respectively, to women (85.4 vs. 80.3, respectively) and the older participants (90.1 vs. 73.8, respectively). Furthermore, urban residents had a significantly higher mean overall health literacy score compared to rural counterparts (86.2 vs. 79.5, respectively). Education was strongly and linearly associated with health literacy score: individuals with ≥9 years of education had a (multivariable-adjusted) mean score of 101.5 compared to 80.1 among those with 1-8 years of education and 66.9 among individuals without any formal schooling. Furthermore, wealthier participants had a significantly higher mean health literacy score compared to their poorer counterparts (85.6 vs. 80.1, respectively) [ Table 4].
Our study provides novel and important information regarding the socio-demographic and socioeconomic factors associated with health literacy level among the older population in Kosovo. We found significant associations of health literacy with sex, age, education, place of residence and self-perceived poverty.   As a potential tool for improving decision making on health, health literacy could be of particular importance among older persons which are often regarded as a disadvantaged population group. Furthermore, health literacy deteriorates with age, as demonstrated in a study where the score of functional health literacy declined by 0.9 for every year of increase in age, controlling for a number of socio-demographic variables [13]. Conversely, another report indicated that older persons with lower health literacy levels had significantly higher rates of chronic conditions and worse physical health compared to people with adequate health literacy [14]. Also, a study conducted in the USA reported that older individuals had a lower average health literacy compared to younger adults [15].

Variable
Unadjusted Our results are generally in concordance with those reported by previous research conducted in the region and beyond, which have highlighted negative associations of health literacy with age and education [15][16][17][18][19][20]. The rate of inadequate or marginal health literacy was found in 81.8% of primary care patients aged ≥65 years in a study in Serbia [19], whereas 59% of adults aged 65 years or older in USA reported below basic or basic health literacy levels [15] compared to 73.6% in our study. Furthermore, health literacy level was reported to be significantly lower among women [19] and those below the poverty line or with a lower income [15,17,19]. The association of health literacy with sex is controversial since some population-based surveys have reported mean health literacy scores to be higher among women than men [15,17]. These sex discrepancies might be influenced by the distribution of gender education gap and educational attainment through the life course. For example, our survey included people aged ≥65 years whereas other studies have surveyed people aged ≥16 years [15] and 18-90 years [17]; usually females are overrepresented among tertiary education students and graduates [21] and they perform better compared to males [22] in developed countries. On the other hand, almost two-thirds of female participants in our survey had no formal schooling and this fact, giving the strong association between health literacy and education, might explain the different sex health literacy results between our study and those reported elsewhere.
The two most widely used tests for measuring health literacy are the Rapid Estimate of Adult Literacy in Medicine (REALM) and TOFHLA. The first one mainly tests the recognition of medical and health related terms [23], whereas TOFHLA assesses numeracy and comprehension skills thus determining whether subjects can read or understand a written prescription [12]. Upon a standardized measurement scale with TOFHLA, in our study, inadequate and marginal health literacy was found in 58.7% and 14.9% of participants, whereas the remaining 26.4% of individuals had an adequate level of health literacy.
Health literacy among old adults has been measured in different settings and using various health literacy tools [13][14]19,[24][25] whereas other studies have explored the health literacy in relation to health care, disease prevention and health system navigation [15,18].
We used a new instrument trying to capture the areas embedded in the current broader concept of health literacy which covers both personal abilities and health system characteristics determining one's ability for making sound health decisions. Our tool was a preliminary version of the HLS-EU instrument, developed by the European Health Literacy Consortium and discussed elsewhere [2].
It is important to study the socioeconomic correlates of health literacy as they can partly explain the pathway to unfavorable health outcomes. The personal socioeconomic and demographic characteristics of a person together with personal aspects such as vision and hearing skills, or verbal ability determine the level of health literacy at a point in time. This level of health literacy then determines the interactions of the individual with the health system in terms of access and utilization of health care, the quality of doctor-patient interaction and self-care, leading finally to various health outcomes [26]. Therefore, it is logical to assume that, the better the health literacy level, the better the health outcomes. In this context, the aim should always be toward improvement of the health literacy level of individuals and, to achieve this objective, the following potential routes are suggested: a) improve health literacy in the population; b) improve written and multimedia communication; c) improve oral communication in health care visits; and, d) alter the system of care by making the task or situation less demanding through, for instance, simplifying or making the system more "readable" [2,27]. Education seems to be vital for increasing the level of health literacy which consequently leads to behavioral change. Thus, it has been suggested that educating diabetic patients about disease self-management may result in higher engagement in healthy behaviors and preventive health care services [28]. Yet, changing behaviors is a complex process and different behavioral change theories have been suggested to explain the attitudes-to-behavior change transition, either through a series of attitude changes, or consequential behavioral change [29]. However, caution is needed about the education-age relationship and attitudes and behavior change.
Our study has several limitations in line with its cross-sectional design which is susceptible to biases of selection and information. Our study included a large population-based sample and the response rate was quite high. Furthermore, the instrument we used for assessment of health literacy was based on a vigorous research work conducted in the framework of a large EU supported project [2]. In addition, we pre-tested our health literacy tool in a sample of older people in Kosovo and Albania before conducting the current survey. Yet, we cannot dismiss the possibility of differential reporting among categories of older people differing in socioeconomic characteristics. Finally, findings from cross-sectional studies should be interpreted with caution.

CONCLUSION
This is probably the first report from the Western Balkans addressing health literacy in a population-based sample. Health literacy is an under-researched topic in countries of Southeast Europe and future prospective studies should be conducted in order to determine the magnitude and determinants of health literacy among the older population in Kosovo and other transitional settings.

CONSENT
All authors declare that 'written informed consent was obtained from the patient (or other approved parties) for publication of this case report and accompanying images.

ETHICAL APPROVAL
An ethical approval from the Professional Ethical Board of the Ministry of Health of Kosovo is available.