To our knowledge this is the first study describing the living conditions, the health and nutritional status of a large representative sample of elderly Lebanese living in rural areas. This study evidenced the low socioeconomic status, the high frequency of poor nutritional and health status, and large gender disparities.
Regarding nutritional status, we found a high prevalence of malnutrition (8.0%) and risk of malnutrition (29.1%), especially in women. Our results are close to those of Kaiser et al. , who published pooled results from studies in five countries (Japan, South Africa, Suede, France, Switzerland) including 964 either healthy or frail home living elderly subjects (malnutrition: 5.8%, risk of malnutrition: 31.9%). Yet, the proportion of malnourished Lebanese elderly matched those reported by Guigoz among 3119 frail elderly individuals from several, mostly developed countries . In a cross-sectional study conducted in Iran among community dwelling elderly people, the authors found a prevalence of 12% of malnutrition and 45.3% of risk of malnutrition . In this sample, females were more often malnourished than men. In Turkey, a neighbor country of Lebanon, the risk of malnutrition, based on the MNA short from, was 28% among subjects admitted to an outpatient clinic . Beside problems of malnutrition, we observed a high level of obesity and overweight, which was present in more than half of the study sample. This is typically observed in low and middle income countries, undergoing nutrition transition, which is characterized by cultural and lifestyle changes, such as decreased physical activity, shift toward more unhealthy diet patterns, modern food processing and rapid growing urbanization . The coexistence of both, under – and over nutrition with a tendency toward non communicable diseases may be due to a high degree of socioeconomic inequities in these transitional countries : as described by Mendez et al.  in a study reporting pooled data from 36 developing countries on the prevalence of over- and underweight among women, two indicators of socioeconomic development, GNI (gross national income) and urbanization, were associated with the prevalence of overweight and inversely associated with the prevalence of underweight. This double burden may result in higher prevalence of disabilities among the elderly population . In our study, obesity was more prevalent among women compared to men and mean BMI was significantly higher in women despite the higher level of malnutrition. According to other authors, this may emphasize that BMI alone is not a reliable tool to assess nutritional status . Our results are close to the findings from a national population-based study published by Sibai et al. , who reported that 27.9% of the Lebanese elderly were obese. However, this study did not provide estimates for malnutrition. Another important finding of our study is the fact that women were highly disadvantaged regarding their socioeconomic status and health. Indeed, women were significantly more often illiterate and had a lower income than men. Data from the Lebanese household survey 2007  showed that among rural and urban elderly individuals, 56.0% of females versus 27.0% of males were illiterate, proportions that are very similar to those observed in our sample. Illiteracy is very common among Arabic countries , especially in rural areas . Moreover, among females, poor nutritional status, chronic disease, frailty, functional disability and cognitive impairment were common and significantly more frequent than in men. Similar findings were reported by other authors in international and regional settings. A study conducted by Chemaitelly et al.  in underprivileged communities of Beirut revealed that women were less educated, reported less subjective health, more chronic diseases and functional disability. These gender differences were also described by Kabir et al.  in a study conducted in Bangladesh, where 80% of women reported having four or more health problems compared to only 42% of their male counterparts. These findings may be partially due to educational and cultural influences; in fact, it seems that women report more symptoms of psychological distress, anxiety, and depression than men . However, women were also more disadvantaged on objective indicators such as number of drugs taken and frequency of falls. Regarding functional ability, most participants were living independently. Yet, we found that females were significantly more dependent in both, basic ADL and IADL activities. Our findings are consistent with results of previous studies conducted among older adults in Beirut . These facts may be explained by the higher level of comorbidity and frailty in females. Our study also revealed that women had more balance disorders and reported significantly more falls during last year than men. Similar gender differences were present in cognitive assessment. In both versions of the MMS, women showed a significantly higher degree of cognitive impairment than men. This means that independently of literacy, women had a worse cognitive function compared to males. Higher comorbidity, depressive disorders and a higher level of loneliness  may explain these observations. On the other hand, it may also reflect the lack of cognitive stimulation and exercise.
Important differences were found when comparing participants according to their cognitive health (MMS < 24 and MMS ≥ 24). Thus cognitive decline was more frequent among both socially disadvantaged groups (females, illiterate and those reporting the lowest income) and subjects with lower health and functional status. These findings confirm results from previous studies reporting an association between cognitive status and several socio-demographic factors such as age , female gender [55, 56] educational attainments [55, 57, 58] and having never been married . Moreover, poor nutritional status  and functional decline  are both common findings in patients suffering from reduced cognitive status.
Major strengths of our study are its sample size and the rural setting. In fact, few studies focused on the specific characteristics of this population and literature shows conflicting results [62–65]. Several studies mentioned health disparities between rural and urban residents :in a study conducted in rural areas of the USA, the authors observed a higher prevalence of cardiovascular disease compared to urban areas, after controlling for possible confounders . Other authors believe that these disparities are not primarily due to influence of residence but mostly to age, gender and socioeconomic status . In fact, the results of two longitudinal surveys in Canada showed a strong positive relationship between socioeconomic status and health . However, not only socioeconomic conditions but also a lower level of awareness may explain the rural–urban disparities. Unfortunately, in Lebanon, no comparison is possible as similar data from urban settings are not available.
Several limitations have to be considered in this survey. First, the cross-sectional design, which does not allow drawing causal relationship. Second, although our random sample can be considered as representative of the rural elderly, we could not do weighting to provide estimates of the prevalence of malnutrition for the whole Lebanese population due to missing population data in rural areas. Beside this, our sample size calculation was based on data which should not be representative for malnutrition among elderly people in developing countries. As mentioned above, this was due to lack of representative data. Furthermore, due to cognitive disorders, lack of memory and educational disparities, some information may be less accurate. In addition, some issues may affect private sphere and responses to these questions may suffer from less reliability. Moreover, most of the health related information was self-reported. Finally, several instruments were initially developed in western culture and may therefore not be culturally sensitive to Lebanon. For example, the MNA has not been validated previously in our population. Thus, it may be of great importance to undergo validation studies of this important screening tool as well as several other specific geriatric assessment tools which are not yet validated among the Lebanese population. The difficulties related to cross cultural adaptation are highlighted by el Rhazi K et al.  in an example of a quality of life questionnaire translated from English to Moroccan Arabic language. In the same way, several attempts were done to adapt the MMS to Arabic speaking cultures [16, 68], but until now validation studies are lacking.
In terms of health and social policy, this study has several implications. First of all, rural public health programs should be implemented stressing on the importance of wellbeing of elderly people with a special focus on women’s health. Furthermore, more general practitioners, nurses and social workers have to be educated regarding health and the specific needs of the elderly population. Health care centers and home care services should be implemented and screening should focus on frail elderly and those at risk of malnutrition who may benefit from early interventions. Moreover, as socio-economic status is associated with poorer health, it is urgent to ensure pensions for elderly and to guarantee overall health insurance. Finally, a special effort should be done to equalize men and women in terms of salary and educational achievement; the latter is of great importance especially in underprivileged rural areas where illiteracy remains high and where a special effort has to be done to improve education of future generations.