Kirsch Cumulative Outcomes Ratio (KCOR), with Gamma Adjustment
Gamma-frailty adjustment
The initial view shows cumulative death counts. The Cohort 2 starting-dose slider assigns lower doses to Cohort 1 and the selected dose and higher doses to Cohort 2. The age slider selects a continuous range of 10-year groups; when it exactly matches ages 80+ or all ages, the detailed groups and the corresponding 80+ or all checkbox are checked together. The checkboxes still allow arbitrary selections. When gamma adjustment or fitting moves a line, the original is retained as a dashed line of the same width and the final value is solid. The lower RR chart shows the final value reflecting adjustment and fit and defaults to a logarithmic axis centered on 1. Clear “Log scale” to switch to a linear axis starting at 0.
The quiet-window start and end must remain at least four weeks apart and default to weeks 4–8. Fitting is separate from gamma adjustment: its start is fixed at week 1, and the blue line is scaled by the Cohort 2 / Cohort 1 ratio at the selected end week, making KCOR equal to 1 there. A fit end of week 0 does not fit; moving the end to week 4 or later applies it automatically.
Weekly risk populations and deaths are summed over the selected areas, ages, and doses within each cohort before theta and k are fitted. Osaka is available for ordinary cumulative death counts, but unavailable during gamma adjustment because its death-only source cannot provide a risk set.
This is a method-validation implementation. theta_zero and theta_upper_bound identify fits at the search boundary.