Survival Analysis Calculator

This calculator builds a Kaplan-Meier survival curve across four periods, starting with everyone at risk (Total Subjects) and multiplying in each period's survival fraction, (at-risk − events) / at-risk (line 44), to get a running Cumulative Survival (line 45). Subjects who leave through an event or through censoring both reduce who's at risk for the next period (line 54), but censoring recorded in the fourth period only ever reduces at-risk counts for a fifth period this calculator never computes, so it specifically never touches Survival Probability at T=4 — the value the atRisk running total feeds into (line 44) is captured before censoring is subtracted for that period. Censoring at T=4 still feeds Total Censored (a plain sum across all four periods, line 58) and the chart's own T=4 "At Risk" figure (line 51, computed from that same period's events and censored count).

If every subject has already had an event or been censored before a period starts, that period's survival estimate is not recalculated at all: an explicit guard (line 35) carries the last cumulative value forward and reports 0 subjects at risk, rather than dividing by an at-risk count of zero. Median Survival Period looks for the first period where cumulative survival drops to 0.5 or below (lines 61-67) and reports -1 if the curve never gets there — which is exactly what happens at the defaults, where survival only falls to 0.6203 by T=4. What this does not account for: any confidence interval around the survival estimates (no Greenwood's formula here), and it treats every subject's event or censoring time as falling cleanly within one of the four fixed periods rather than working from exact event times the way a full Kaplan-Meier table would.