Certainty of Results
Sensitivity analysis provides information on the level of certainty of the results, and which underlying parameters contribute to uncertainty. Based on the underlying distributions of all parameters, we can report the 95% credible interval of the total economic burden of $139 billion to be $112 to $174 billion.The credible interval is analogous to a confidence interval, and indicates that assuming correct estimates for parameter distributions, 95% of all possible economic burden estimates based on uncertainty in all parameters would fall in this range. The distribution of burden estimates is shown in the graph below. Most of the uncertainty in the total estimate is due to high uncertainty in productivity losses (95% C.I. $24.7-$80.6 billion). This uncertainty is a result of 1) high uncertainty in the prevalence of visual impairment and blindness, 2) high uncertainty in the productivity reduction attributable to low vision, and 3) the high cost estimate of productivity losses. Additional uncertainty is also attributable to the cost of diagnosed disorders (95% C.I. $33.9-$55 billion), long-term care costs (95% C.I. $12-$24.9 billion), the cost of undiagnosed low vision (95% C.I. $0.8-$9.1 billion) and the cost of informal care (95% C.I. $0.8-$3.8 billion). All other cost components exhibited less than $1 billion in uncertainty.