Table 3.
Lag period for SPMb | (A) AD-affected days |
(B) AD-free days |
P-value for (A) – (B) | ||||
---|---|---|---|---|---|---|---|
RR | 95% CI | RR | 95% CI | ||||
Lag 0 | 0.972 | 0.931 | 1.015 | 1.004 | 0.983 | 1.025 | 0.19 |
Lag 1 | 1.000 | 0.963 | 1.039 | 1.005 | 0.986 | 1.025 | 0.84 |
Lag 2 | 1.032 | 0.996 | 1.070 | 1.017 | 0.998 | 1.037 | 0.49 |
Lag 3 | 1.028 | 0.988 | 1.071 | 1.038 | 1.018 | 1.059 | 0.69 |
Lag 4 | 1.038 | 1.001 | 1.076 | 1.026 | 1.010 | 1.042 | 0.58 |
Lag 5 | 1.060 | 1.011 | 1.111 | 1.023 | 1.007 | 1.038 | 0.15 |
Lag 6 | 0.999 | 0.942 | 1.060 | 1.014 | 0.993 | 1.035 | 0.66 |
Lags 0–5 | |||||||
All clinics | 1.066 | 1.007 | 1.127 | 1.062 | 1.022 | 1.104 | 0.94 |
Except Clinic IV | 1.071 | 1.011 | 1.135 | 1.055 | 1.014 | 1.097 | 0.68 |
AD Asian Dust, SPM suspended particulate matter, RR relative risk, CI confidence interval
aUsing single-lag model with adjustment for the day of the week, public holidays, month, year, the natural cubic splines of daily mean temperature, relative humidity and pollen (lag 0–5) with 3° of freedom. An interaction term between SPM and pollen concentration was included. Refer to Additional file 2: Table S2 for further information
bDelayed effects longer than 1 day were constrained as average