Table 4

Results of the ecological-longitudinal analysis for three cities, Barcelona, Valencia and Vigo, 1992–1994

City		Black smoke^4-150				Sulphur dioxide^4-151
City		β	95% CI			β	95% CI
Air pollutants introduced separately in two models
All three	0.00064	0.00021, 0.00106		0.00041	0.00011, 0.00072
Barcelona	0.00064			0.00041
Valencia	0.00063			0.00041
Vigo	0.00064			0.00041

Air pollutants introduced jointly in the same model
All three	0.00071	0.00034, 0.00108		0.00020	0.00009, 0.00031
Barcelona	0.00071			0.00020
Valencia	0.00071			0.00020
Vigo	0.00071			0.00020

↵4-150 One lag, introduced linearly.
↵4-151 Current value and one lag, introduced linearly.
Models also contained the following explanatory variables with parameters statistically significant (p<0.1):
intercept, ns(t,36), friday, saturday, sunday, holidays, ns(temperature_t-1,4), ns(temperature_t-2,4), ns(temperature_t-3,4), ns(temperature_t-4,4), ns(temperature_t-7,4), ns(temperature_t-2,4):humidity_t-2, ns(temperature_t-3,4):humidity_t-3, ns(temperature_t-4,4):humidity_t-4, humidity_t, ma(influenza,15), total_t-1, total_t-2, total_t-3, total_t-4, total_t-5.and total_t-6.
Where ns(x,k) denotes a natural cubic spline for x with k degrees of freedom; day of the week an indicator for the day of the week; : denotes interaction; ma(,15) a moving average of 15 terms and total, daily total mortality.