Table 2. Interrupted time series analysis of yearly smoking prevalence during the 2006–2018 economic crisis (recession) among Brazilians aged 18 years and older residing in one of 27 capital cities, according to age group.

	18–30 years old	31–44 years old	45–64 years old	65 and older
Yearly average percentage variation of smoking prevalence in the 2006–2014 period	-0.61^a	-0.75^a	-0.81^a	-0.22^a
	[-0.75 to -0.48]	[-0.82 to -0.69]	[-1.05 to -0.58]	[-0.34 to -0.11]
Yearly average percentage variation of smoking prevalence in the 2015–2018 period	-0.34^a	0.07	-0.50^a	-0.70^a
	[-0.45 to -0.23]	[-0.31 to 0.46]	[-0.87 to -0.14]	[-0.84 to -0.56]
Level change in 2015 vs predicted level without economic crisis	0.43	-0.12	0.03	1.25^a
	[-0.28 to 1.16]	[-.06 to 0.81]	[-.36 to 1.43]	[0.30 to 2.20]
Difference in trends (period 2015-2018; period 2006–2014)	0.27^a	0.83^a	0.30	-0.47^a
Difference in trends (period 2015-2018; period 2006–2014)	[0.12 to 0.42]	[0.45 to 1.21]	[-0.16 to 0.78]	[-0.61 to -0.34]

95% confidence intervals in brackets.

^a p < 0.05.

Models considered autocorrelation and used robust standard errors to account for repeated measures.