OBJECTIVE: 'The Norwegian Women and Cancer Study' (NOWAC) was created as a national population-based cohort study by taking advantage of the existing population registers in Norway. Thus, the women in the NOWAC study should be representative for the entire female population in the corresponding age-groups. The aim of this paper was to study the external validity of the cohort as a necessary premise for the estimation of population attributable risk. METHODS: Between 1991 and 1997 women were sampled randomly from the national Central person registry in Norway. A total of 179,388 women were invited, of whom 102,443 can be included in the forthcoming follow-up analysis. The response-rates were around 60% in the age-groups 30-34 years till 55-59 years, while 44.7% responded among those aged 65-70 years. Information was collected by postal questionnaires. Follow-up information was based on linkage to national end-point registers with use of the unique national identification number given all Norwegians. RESULTS: Validation of the information on parity and education by linkage to national registers showed only minor differences for these factors between the responders and the total sample of women. A postal survey among non-responders indicated that the most important reasons for not returning the questionnaire were lack of time and concern about privacy. The results showed no differences in life-style factors between the original responders and the non-responders. No significant differences were found between the observed incidence rates for all cancer sites or cancer of the breast compared with national figures for the year 1999. CONCLUSION: The analysis revealed no major source of selection bias that could seriously invalidate the estimation of population attributable risk.
OBJECTIVE: 'The Norwegian Women and Cancer Study' (NOWAC) was created as a national population-based cohort study by taking advantage of the existing population registers in Norway. Thus, the women in the NOWAC study should be representative for the entire female population in the corresponding age-groups. The aim of this paper was to study the external validity of the cohort as a necessary premise for the estimation of population attributable risk. METHODS: Between 1991 and 1997 women were sampled randomly from the national Central person registry in Norway. A total of 179,388 women were invited, of whom 102,443 can be included in the forthcoming follow-up analysis. The response-rates were around 60% in the age-groups 30-34 years till 55-59 years, while 44.7% responded among those aged 65-70 years. Information was collected by postal questionnaires. Follow-up information was based on linkage to national end-point registers with use of the unique national identification number given all Norwegians. RESULTS: Validation of the information on parity and education by linkage to national registers showed only minor differences for these factors between the responders and the total sample of women. A postal survey among non-responders indicated that the most important reasons for not returning the questionnaire were lack of time and concern about privacy. The results showed no differences in life-style factors between the original responders and the non-responders. No significant differences were found between the observed incidence rates for all cancer sites or cancer of the breast compared with national figures for the year 1999. CONCLUSION: The analysis revealed no major source of selection bias that could seriously invalidate the estimation of population attributable risk.
Authors: Elizabeth E Hatch; Kristen A Hahn; Lauren A Wise; Ellen M Mikkelsen; Ramya Kumar; Matthew P Fox; Daniel R Brooks; Anders H Riis; Henrik Toft Sorensen; Kenneth J Rothman Journal: Epidemiology Date: 2016-01 Impact factor: 4.822