**4. Discussion**

The purposes of this study were to ascertain the longitudinal relationship of developing the mobility disability and IADL disability and to report the hazard ratio, the median age onset, and the median survival time to the onset of IADL disability in each hierarchical stage of

The Hierarchical Status of Mobility Disability

consideration these task-specific impairments.

should be taken into consideration.

status.

disabled.

**5. Conclusion** 

Predicts Future IADL Disability: A Longitudinal Study on Ageing in Taiwan 93

suggested that heavy housework requires more muscle power, whole-body strength, balance and coordination, while climbing stairs and walking on a level surface require standing balance and lower extremity strength and velocity (Bean et al., 2008; Chen et al., 2010). Early intervention for mobility performance and stability needs to take into

The results of this study confirm the time window that health care providers have in order to reverse mobility disability and to prevent IADL disability. Our findings suggest that people with two or more items disabled in mobility develop IADL disability in 2-4 years, whereas people with one or less items disabled in mobility develop IADL disability in 6 years or longer. The survival analyses further suggested that men and women have different disablement patterns. In general, the interval for 50% of participants to develop IADL disability (median survival time) was shorter for women than that for men. Therefore, in health promotion or early intervention, the different time windows for men and women

In this Taiwan Longitudinal Study on Ageing (TLSA) dataset, individuals were followed up every four year. The data of the year of 1999, 2003, and 2007 were used in the current study that included two follow-ups at 4 years and 8 years later. The 3rd quartile was not reported from the statistics output of the survival analysis, it is probably because the maximal follow up duration was 8-year and by that time not yet 75% of the individuals developed that certain hierarchical disability level. This study was also limited by the long interval of follow up (every four years). However, a shorter follow-up period consumes more resources. The need for a balance between the large cost and the additional information that could be gathered by a shorter follow-up period needs to be carefully considered. In addition, the population in our study was free of IADL disability at baseline, with 977 individuals excluded due to initial IADL disability. Our results should not be generalized to people who have both mobility and IADL disability. Furthermore, mobility disability is a changing condition, but our prediction of future IADL disability was based only on baseline mobility

The hierarchical status of mobility disability is the strongest predictor of IADL disability even after adjustments for the significant risk factors of demographics, health behaviors, and health status. Very different results of IADL disability development were found between the groups with two or more items disabled and those with one or less items disabled in mobility, which provides support for the value of hierarchical stages of mobility categorization, as opposed to the previous dichotomous definition, with any one item

People who have more disabled mobility items but are free of IADL disability initially are at higher risk of developing IADL disability than those with one or less item disabled, and the time to development is only 2-4 years. We suggest that health care providers focus on people who have two or more items disabled in mobility and that they intervene within the time window of 2-4 years in order to reverse mobility disability or to prevent IADL

disability, both of which are situated in the earlier stages of the disablement process.

mobility disability for use in the development of early intervention programs. In this study, we defined a four-level hierarchy of disability severity in the mobility domain by the summed number of items labeled as difficult among three items: heavy housework, climbing stairs, and walking on a level surface. Our results indicate that the hierarchy of mobility disability used in this study can significantly identify people with different demographics, health behaviors, and health status. Furthermore, this hierarchical mobility status also has a hierarchical structure in terms of the hazard ratio, the median age onset, and the median survival time to development of IADL disability.

Assessing a hierarchy of mobility disability based on the numbers of items disabled can discriminate between older adults with different levels of physical performance (Wang et al., 2005). The results of this study further substantiate the predictive validity of this hierarchy of mobility disability for future IADL disability at four years and eight years later. Individuals with more severe levels of mobility disability were at greater risk of developing IADL disability, even after adjusting for other risk factors of demographics, health behaviors, and health status. From the magnitude of hazard ratio in the final Cox model, it could be seen that the level of mobility disability appeared to be the strongest predictor of future IADL disability. To the best of our knowledge, this is the first study to examine the predictive validity of the item-wise hierarchy of mobility disability for future IADL disability based on longitudinal follow up on a nationally representative sample of the Taiwanese community-dwelling elderly.

Consistent with previous studies, demographics such as age, sex, and working status were significant risk factors in the initial Cox models (Jette & Branch, 1981; Pinsky et al., 1987; Guralnik & Kaplan, 1989), but they became insignificant in the final model due to the addition of covariates related to health status. Health status, namely number of comorbidities, cognition, and depression symptom score, were significant risk factors in the final model. However, from the perspective of health promotion, demographics such as age and gender are non-modifiable and hence are not the focus of discussion in this study. Education and the health behaviors that were found to be significant risk factors of IADL disability in the current study, such as health status, alcohol consumption, and cigarette smoking, are in agreement with the literature (Jette & Branch, 1981; Pinsky et al., 1987; Guralnik & Kaplan, 1989) and are valuable in guiding health promotion policy or programs for people at younger ages. For example, policy for extending the years for obligatory education could help people get higher education, and that might in turn lead to better socioeconomic status and policy for health education for the publics could facilitate better health behaviors. Heightened socioeconomic status and better health behaviors could lessen the numbers of comorbidities people will develop during the process of ageing and decrease the negative impact that comorbidities might have on cognitive function and emotional health.

Surprisingly, habitual exercise was not significant in either the four or eight-year follow-up, and we propose two possibilities for this insignificance. First, our dichotomous cutting point was based on frequency of exercise per week, which did not consider exercise intensity and may fail to reflect the health benefits of exercises. Second, general exercise may be insufficient for people who already have some mobility disability, and specific training for specific impairments may be necessary, such as intervention for joint range of motion and lower leg eccentric contraction to improve the ability to climb stairs. Previous research has suggested that heavy housework requires more muscle power, whole-body strength, balance and coordination, while climbing stairs and walking on a level surface require standing balance and lower extremity strength and velocity (Bean et al., 2008; Chen et al., 2010). Early intervention for mobility performance and stability needs to take into consideration these task-specific impairments.

The results of this study confirm the time window that health care providers have in order to reverse mobility disability and to prevent IADL disability. Our findings suggest that people with two or more items disabled in mobility develop IADL disability in 2-4 years, whereas people with one or less items disabled in mobility develop IADL disability in 6 years or longer. The survival analyses further suggested that men and women have different disablement patterns. In general, the interval for 50% of participants to develop IADL disability (median survival time) was shorter for women than that for men. Therefore, in health promotion or early intervention, the different time windows for men and women should be taken into consideration.

In this Taiwan Longitudinal Study on Ageing (TLSA) dataset, individuals were followed up every four year. The data of the year of 1999, 2003, and 2007 were used in the current study that included two follow-ups at 4 years and 8 years later. The 3rd quartile was not reported from the statistics output of the survival analysis, it is probably because the maximal follow up duration was 8-year and by that time not yet 75% of the individuals developed that certain hierarchical disability level. This study was also limited by the long interval of follow up (every four years). However, a shorter follow-up period consumes more resources. The need for a balance between the large cost and the additional information that could be gathered by a shorter follow-up period needs to be carefully considered. In addition, the population in our study was free of IADL disability at baseline, with 977 individuals excluded due to initial IADL disability. Our results should not be generalized to people who have both mobility and IADL disability. Furthermore, mobility disability is a changing condition, but our prediction of future IADL disability was based only on baseline mobility status.
