**3. Results**

The demographic and health related information of our participants in the hierarchical mobility disability groups are summarized in Table 1. All variables showed significant differences among the four hierarchical mobility disability groups and were used as covariates in the following Cox regression analysis. The "mobility able" group was younger; had larger

There were five items in the instrumental activities of daily living (IADL) domain (shopping, finance, transportation, light housework, and telephone). Older adults were classified as having IADL disability if they reported any degree of difficulty or inability to

Covariates were included in the Cox regression models according to their values in 1999. Age was categorical data, whereas sex, current working status, smoking habit, and alcohol ingestion were dichotomous data. Education was recorded as illiteracy or no formal education, elementary school, primary or senior high school, and college or beyond. Spouse status was recorded as either living with spouse (married or living together) or living

The exercise habit was divided into inactive (less than three times a week) and active (at least three times a week). Self-rated health was divided into healthier and worsening. The number of co-morbidities was counted from the list of hypertension, diabetes, cardiac disease, stroke, cancer, and arthritis. Cognitive function was measured with the modified Taiwan version of SPMSQ using nine items (range 0-9), correct answers were coded 1, whereas errors were coded 0, thus, higher SPMSQ score means better cognition performance, individuals were categorized as having normal (6 and above) or abnormal (5 and below) cognitive function (Yen et al., 2010). Symptoms of depression were assessed with the CESD, on which each of the 10 items is scored from 0 to 3. Individuals were categorized

The group means difference among the four hierarchical mobility disability groups was determined by analysis of variance (ANOVA) for the continuous variables and chi-square test for discrete variables. Median age difference among the four groups were investigated by Brown-Mood test. The group comparison on the depression score among the four groups was determined by using the ANOVA on the log transformed CESD score. Cox proportional hazard regression analysis was used to determine if the hierarchical mobility disability stage was a significant predictor of future IADL disability four years later (in 2003) and eight

The median age onset of IADL disability of each hierarchical mobility disability group across eight years of follow up was determined by survival analysis. The survival time for 50% of participants in each hierarchical status of mobility disability to develop IADL disability (the median survival time) was determined for the whole group and separately for

The demographic and health related information of our participants in the hierarchical mobility disability groups are summarized in Table 1. All variables showed significant differences among the four hierarchical mobility disability groups and were used as covariates in the following Cox regression analysis. The "mobility able" group was younger; had larger

without spouse (never married, divorced, separated, or widowed).

as not having (9 and below) or having (10 and above) depressive symptoms.

**2.1.2 IADL disability status** 

perform on at least one item.

**2.2 Statistical analysis** 

men and women.

**3. Results** 

years later (in 2007) and to report its hazard ratio.

**2.1.3 Covariates in Cox regressions** 


percentages of men; had higher educational levels, larger percentages currently working, and spouses; smoked, drank alcohol, and exercised; and had a lower number of co-morbidities, a better perceived health status, better cognition, and lower depression symptom scores.

¥: median (q1,q3), a significant differences were found between "mobility able group and all other (1 item, 2 items, and 3 items) disabled" groups; b significant differences were found between "mobility able and 2 items disabled"; c significant differences were found between "mobility able and 3 items disabled" group; d significant differences was found between "1 item disabled and 3 items disabled groups"; e significant differences was found between "1 item disabled and 2 items disabled groups"; f no statistical significance between item 2 and item 3.

Table 1. Demographic and health-related information at baseline (year of 1999) (n=2127).

The Hierarchical Status of Mobility Disability

more covariates relating to health status.

number of co-morbidities, and cognition.

disability.

Men

Women

All

: not defined

**4. Discussion** 

of follow up, 1999-2007) (n=2127).

Predicts Future IADL Disability: A Longitudinal Study on Ageing in Taiwan 91

In the final model, significant risk factors for developing IADL disability were mobility disability stage, educational level, alcohol consumption, number of co-morbidities, and depression symptom score. It should be noted that age, sex, and working status were significant risk factors in Models 2 & 3, but became insignificant in the final Model 4, with

Similar to those from the four-year follow-up data, the hazard ratios for developing IADL disability after eight years were also hierarchical, but they had smaller values: 1.96, 2.64, and 2.88 for the groups of "1 item disabled", "2 items disabled", and "3 items disabled", respectively (Table 3). The final model in the eight-year follow-up, when compared to the four-year follow-up data, had more covariates: sex, educational level, cigarette smoking,

The median age at onset for "mobility able", "1 item disabled", 2 items disabled", and "3 items disabled" group was 82, 80, 77, and 76 years of age, respectively. The median survival time is reported for each hierarchical stage of mobility disability and as follows: greater than 8 years for "mobility able", 6 years for "1 item disabled", 6 years for "2 items disabled", and 2 years for "3 items disabled" (Table 4). Inspection of the data in the four mobility disability groups revealed that it took longer period for men than for women to develop IADL

Mobility disability at baseline N Median survival time (years)

Mobility Able 997 (6, ) 1 Item Disabled 173 6 (2, ) 2 Items Disabled 58 6 (2, ) 3 Items Disabled 29 2 (2, 7)

Mobility Able 534 7 (4, ) 1 Item Disabled 186 6 (2, ) 2 Items Disabled 110 2 (2, 6) 3 Items Disabled 40 2 (2, )

Mobility Able 1531 8 (6, ) 1 Item Disabled 359 6 (2, ) 2 Items Disabled 168 4 (2, ) 3 Items Disabled 69 2 (2, )

Table 3. The median survival time for each hierarchical status of mobility disability (8 years

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

Tables 2 and 3 show the results of the Cox regression models, with significant effects of the hierarchy of mobility on developing IADL disability. As shown in Table 2, the unadjusted hazard ratios for developing IADL disability after four years, with "mobility able" as the reference group, were as follows: 2.15 for "1 item disabled", 3.09 for "2 items disabled", and 3.63 for "3 items disabled". After adjustment for potential risk factors, the hierarchical structure of hazard ratios of the four-level mobility status remained the same, though diminished in value (1.55, 1.85, and 2.19), yet they were still the strongest among the significant risk factors.


\*\*\* <0.0001, \*\* <0.01, \*<0.05

Table 2. Hazard ratio of each hierarchical mobility disability group to develop IADL disability across 4 years follow-up (years of 1999-2003) (n=2073).

In the final model, significant risk factors for developing IADL disability were mobility disability stage, educational level, alcohol consumption, number of co-morbidities, and depression symptom score. It should be noted that age, sex, and working status were significant risk factors in Models 2 & 3, but became insignificant in the final Model 4, with more covariates relating to health status.

Similar to those from the four-year follow-up data, the hazard ratios for developing IADL disability after eight years were also hierarchical, but they had smaller values: 1.96, 2.64, and 2.88 for the groups of "1 item disabled", "2 items disabled", and "3 items disabled", respectively (Table 3). The final model in the eight-year follow-up, when compared to the four-year follow-up data, had more covariates: sex, educational level, cigarette smoking, number of co-morbidities, and cognition.

The median age at onset for "mobility able", "1 item disabled", 2 items disabled", and "3 items disabled" group was 82, 80, 77, and 76 years of age, respectively. The median survival time is reported for each hierarchical stage of mobility disability and as follows: greater than 8 years for "mobility able", 6 years for "1 item disabled", 6 years for "2 items disabled", and 2 years for "3 items disabled" (Table 4). Inspection of the data in the four mobility disability groups revealed that it took longer period for men than for women to develop IADL disability.


: not defined

90 Rehabilitation Medicine

Tables 2 and 3 show the results of the Cox regression models, with significant effects of the hierarchy of mobility on developing IADL disability. As shown in Table 2, the unadjusted hazard ratios for developing IADL disability after four years, with "mobility able" as the reference group, were as follows: 2.15 for "1 item disabled", 3.09 for "2 items disabled", and 3.63 for "3 items disabled". After adjustment for potential risk factors, the hierarchical structure of hazard ratios of the four-level mobility status remained the same, though diminished in value (1.55, 1.85, and 2.19), yet they were still the strongest among the

Mobility Able 1 1 1 1 1 Item Disabled 2.15 (1.75-2.68)\*\*\*1.59(1.27-2.00)\*\*\* 1.57 (1.25-1.98)\*\* 1.55 (1.21-1.99)\*\*

60-65 years 1 1 1 65-70 years 1.58 (1.09-2.29)\* 1.60 (1.10-2.33)\* 0.89 (0.12-6.61)

Illiterate 1 1 1 Elementary school 0.71 (0.58-0.88)\*\* 0.72 (0.59-0.90)\*\* 0.78 (0.62-0.99)\*

Men vs. Women 0.80 (0.65-0.98)\* 0.79 (0.63-1.00)\* 0.80 (0.62-1.03)

College and above 0.77 (0.43-1.07) 0.73 (0.46-1.17) 0.80 (0.49-1.31)

Yes vs. No 0.69 (0.50-0.95) \* 0.68 (0.50-0.94)\* 0.76 (0.53-1.09)

Yes vs. No 0.97 (0.79-1.18) 0.96 (0.79-1.17) 1.03 (0.83-1.28)

Yes vs. No 1.23 (0.96-1.57) 1.30 (1.00-1.69)

Yes vs. No 0.77 (0.60-1.00)\* 0.75 (0.57-0.99)\*

Yes vs. No 0.89 (0.73-1.08) 0.96 (0.78-1.18) No. of co-morbidities 1.12 (1.02-1.24)\*

Good vs. Poor 1.03 (0.82-1.29) SPMSQ 1.39 (0.90-2.17) CESD-10 1.31 (1.03-1.67)\*

Table 2. Hazard ratio of each hierarchical mobility disability group to develop IADL

disability across 4 years follow-up (years of 1999-2003) (n=2073).

2 Items Disabled 3.09 (2.40-3.98)\*\*\*2.17 (1.67-2.83)\*\*\*

3 Items Disabled 3.63 (2.59-5.10)\*\*\*2.61 (1.85-3.69)\*\*\*

70-75 years 2.14 (1.55-2.96)\*\*\*

Over 75 years 3.24 (2.33-4.49)\*\*\*

Model 1 Model 2 Model 3 Model 4

0.63 (0.46-0.85)\*\* 0.65 (0.48-0.88)\*\* 0.72 (0.52-0.99)\*

2.14 (1.64-2.79)\*\*\*1.85 (1.37-2.50)\*\*\*

2.48 (1.74-3.52)\*\*\*2.19 (1.46-3.28)\*\*

2.19 (1.58-3.03)\*\*\* 1.13 (0.15-8.29)

3.30 (2.37-4.59)\*\*\* 1.80 (0.25-13.20)

significant risk factors.

Age

Sex

Educational level

school

Work status

Spouse status

Exercise

Cigarette smoking

Self rated health

\*\*\* <0.0001, \*\* <0.01, \*<0.05

Alcohol consumption

Junior or senior

Table 3. The median survival time for each hierarchical status of mobility disability (8 years of follow up, 1999-2007) (n=2127).
