Table 6.

Smoking-attributable fraction (SAF) during cohort follow up, and smoking-attributable deaths 2009-2018, for adults aged ≥45 years

	2008 smoking status (proportion)			Past-smoker: sex-combined mortality RR			Current-smoker: sex-combined RR				Smoking-attributable fraction (SAF)b			National smoking-attributable deaths 2009-18b
	Current	Past	Never	RR	LCI	UCI	RR	LCI	UCI	National total deaths 2009-2018a	Estimate (∝=0.90)	Lower bound (∝=0.80)	Upper bound (∝=1.00)	Estimate (∝=0.90)	Lower bound (∝=0.80)	Upper bound (∝=1.00)
Males
45–64 years	0.462	0.338	0.201	3.62	1.53	8.55	6.92	2.98	16.04	2953	70.5	62.7	78.4	2083	1851	2314
65–74 years	0.262	0.520	0.218	2.53	1.18	5.45	5.51	2.36	13.15	3105	59.8	53.1	66.4	1856	1650	2062
≥65 years	0.120	0.721	0.159	0.95	0.50	1.78	1.97	0.83	4.66	5052	7.9	7.0	8.8	400	355	444
						Male deaths at age ≥45 years				11 110	54.4	48.3	60.4	6042	5371	6714
						Male deaths at all ages				15 890	38.0	33.8	42.3	6042	5371	6714
Females
45–64 years	0.437	0.254	0.309	3.62	1.53	8.55	6.92	2.98	16.04	1993	68.8	61.2	76.5	1372	1219	1524
65–74 years	0.208	0.264	0.528	2.53	1.18	5.45	5.51	2.36	13.15	2484	51.6	45.8	57.3	1281	1139	1423
≥65 years	0.166	0.401	0.433	0.95	0.50	1.78	1.97	0.83	4.66	5738	11.1	9.9	12.4	638	567	709
						Female deaths at age ≥45 years				10 215	46.0	40.8	51.1	4694	4173	5216
						Female deaths at all ages				13 181	35.6	31.7	39.6	4694	4173	5216
Persons
						Total deaths at age ≥45 years				21 325	50.3	44.8	55.9	10 737	9544	11 930
						Total deaths at all ages				29 071	36.9	32.8	41.0	10 737	9544	11 930

Current smoking prevalence includes daily and non-daily smokers; the vast majority of current-smokers are daily smokers.

RR, relative risk; LCI, lower confidence interval; UCI, upper confidence interval; SAF, smoking-attributable fraction.

^aNational estimates of annual total deaths by sex were extracted from the Australian Bureau of Statistics’ ABS.Stat [http://stat.data.abs.gov.au/]. Age-group-specific deaths data are only available for five of the eight states/territories in Australia (New South Wales, Queensland, South Australia, Western Australia and the Northern Territory) due to data quality limitations. To estimate national age-sex group deaths, we applied a scale factor (total number of deaths by sex/five state-territory deaths by sex) to the number of deaths in each age-sex group in the five state-territory data. The mortality statistics used may underestimate deaths in this population, due to potential misclassification of Aboriginal and/or Torres Strait Islander deaths as non-Aboriginal, and lags in death registration. It has been estimated that over 2001-15, 13.5% of all male and 13.9% of all female Aboriginal and Torres Strait Islander deaths were misclassified as non-Indigenous deaths.²⁰ Misclassification was particularly common among those aged ≥65 years (19.3% and 17.7%, respectively), but still high among those aged 45–64 years (11.9% and 11.0%, respectively). If this rate of misclassification was consistent over the 2009-18 period, we would have underestimated deaths by 15.0% (missing 3204 deaths ≥45 years) (Table S12).

^bAccording to assumptions that 90%, 80% and 100% of excess deaths among smokers are smoking-attributable. We calculated the smoking-attributable fraction (SAF) using the prevalence-based method, for the age groups 45-64, 65–74, and ≥75 years: SAF(%)=100x[P_p(RR_p-1)+P_c(RR_c-1)]/[P_p(RR_p-1)+P_c(RR_c-1)+1]. Here, P_p and P_c are the prevalence of past and current smoking, respectively, and RR_p and RR_c are the RRs for mortality among past- and current-smokers, respectively, compared with never-smokers. The smoking-attributable fraction is calculated using sex-combined RR² results for participants aged 45–64, 65–74 and ≥75 years, adjusted for age as the underlying time variable, sex, education and remoteness.