Number needed to harm (NNH) — sign-reversed NNT

Number needed to harm (NNH) — sign-reversed NNT

Usage

number_needed_to_harm(a, b, c, d, confidence = 0.95)

Arguments

a, b, c, d

Cell counts of the 2x2 table: a = exposed with outcome, b = exposed without, c = unexposed with outcome, d = unexposed without.

confidence

Confidence level. Default 0.95.

Value

A morie_effect_size.

Examples

# Exposure raises the bad outcome from 10% to 20% -> NNH = 1/0.10 = 10.
res <- number_needed_to_harm(20, 80, 10, 90)
res$estimate                 # people exposed to cause one extra harm
#> [1] 10

# NNH is the reciprocal of the (absolute) risk difference, like NNT but
# for a harmful exposure.
1 / abs(risk_difference(20, 80, 10, 90)$estimate)
#> [1] 10