Table 2.
Multivariable generalized linear models for annual medical service utilization and expenditure stratified by high and low SMMIa (n = 1337)
| Total (n = 1337) | Men (n = 689) | Women (n = 648) | ||||
|---|---|---|---|---|---|---|
| High risk | Low risk | High risk | Low risk | High risk | Low risk | |
| n (%) | 330 (24.7) | 1007 (75.3) | 171 (24.8) | 518 (75.2) | 159 (24.5) | 489 (75.5) |
| Deceased (%) | 165 (50.0) | 305 (30.3) | 90 (52.6) | 183 (35.3) | 75 (47.2) | 122 (25.0) |
| Medical service utilization | ||||||
| Ambulatory care visits, timesb | ||||||
| Outpatient services | 25.1 (21.9)*** | 28.7 (24.6) | 25.4 (22.3) | 26.8 (23.7) | 24.7 (21.7)*** | 30.7 (25.3) |
| Preventive care | 1.00 (0.93) | 1.27 (1.16) | 0.84 (0.82) | 1.02 (0.98) | 1.17 (1.08) | 1.53 (1.46) |
| Dental services | 0.65 (0.14)*** | 1.03 (0.42) | 0.74 (0.14)*** | 1.11 (0.46) | 0.56 (0.14) | 0.94 (0.30) |
| Emergency | 0.81 (0.31)* | 0.57 (0.28) | 0.83 (0.39)* | 0.62 (0.29) | 0.79 (0.29) | 0.51 (0.27) |
| Hospitalization, daysb | 13.6 (2.15)*** | 7.13 (1.32) | 16.0 (2.58)*** | 9.03 (1.45) | 10.9 (1.52)*** | 5.12 (1.18) |
| Medical service expenditure, 1000 NT$c | ||||||
| Ambulatory cared | ||||||
| Outpatient services | 20.8 (16.4)** | 25.9 (18.4) | 23.2 (17.1) | 26.0 (18.1) | 18.3 (15.9)* | 25.7 (18.8) |
| Preventive care | 0.21 (0.18)* | 0.28 (0.24) | 0.18 (0.15) | 0.23 (0.19) | 0.25 (0.20)* | 0.34 (0.31) |
| Dental services | 0.54 (0.07)** | 0.88 (0.29) | 0.63 (0.06)* | 0.98 (0.36) | 0.44 (0.07) | 0.77 (0.24) |
| Emergency | 2.80 (0.76)** | 1.70 (0.51) | 2.74 (0.80)** | 1.77 (0.59) | 2.87 (0.72) | 1.62 (0.43) |
| Hospitalizationd | 77.5 (10.6)*** | 38.7 (6.89) | 79.6 (12.2)** | 49.8 (8.03) | 75.3 (6.50)*** | 26.9 (5.86) |
| Total medical expenditured | 102 (34.4)*** | 67.4 (33.4) | 106 (39.6) | 78.8 (34.8) | 97.2 (31.7)** | 55.4 (32.2) |
aThe high-risk group comprises participants with a lower SMMI (<11.4 kg/m2 for men and 8.50 kg/m2 for women). The low-risk group comprises participants with a normal SMMI and serves as a reference group
bMean (median). Outcome was assessed using GLMs with Poisson distribution
cNT$, with an exchange rate of approximately NT$31 to US$1 in 2017
dMean (median). Outcomes was assessed using GLMs with gamma distribution
Models were adjusted for age, region of residence, ethnicity, education level, living status, self-reported financial status, household income (NT$/mo), perceived health status, ADL, CCI, energy (kcal/d), protein (g/d), DDS (≤4 or > 4), and physical activity (<1.5 MET/day or ≥ 1.5 MET/day)
* P < 0.05, ** P < 0.01, *** P < 0.001