Attributable fractions for premature mortality avoidable through public health action: updated estimates by Australian public health physicians

Objective: To determine whether previously published estimates of attributable fractions of primary avoidable mortality (PAM) and secondary avoidable mortality (SAM) were valid to Australian public health physicians. Design: Self-administered survey. Setting: Australia-wide survey. Participants: 331 Fellows of the Australasian Faculty of Public Health Medicine (71.8% response rate). Results: Tobias and Jackson's cited estimates for PAM and SAM for lung cancer were the least contested by respondents (62% 'no change' for PAM estimate; 67% 'no change' for SAM estimate). By contrast, 39% indicated that the PAM estimate of 30% for diabetes required change. Those most-contested PAM estimates also typically elicited significantly different responses by gender (breast cancer, diabetes and falls), age (breast cancer, stroke and diabetes) and training (stroke, diabetes, mental health). The proportion of respondents who agreed with the cited estimates either for PAM or SAM for any of the 10 selected conditions never differed significantly according to current primary role ('practitioner' v 'researcher' v 'other'). Conclusions: PAM and SAM represent the key action areas for public health. We conclude that the estimates of Tobias and Jackson are generally robust among Australian public health physicians, invoking confidence that their advocacy will engage practitioners, policy makers and citizens seeking better governance of health system performance.