The Pearl Index quantifies how often a contraceptive fails in practice, allowing different methods to be ranked.
It is computed as accidental pregnancies normalised to 100 woman-years of exposure:
\[ PI = \frac{\text{accidental pregnancies} \times 1200}{\text{total months of use}} \]
where multiplying by $1200$ scales the result to $100$ women observed for $12$ months each. So if $5$ pregnancies occur across $100$ women followed for one year, the index is $5$. The smaller the number, the more reliable the method (e.g., IUCDs and implants have very low indices, while barrier methods are higher).
Choices framing it merely "per woman year," as a ratio of pregnancies to losses, or per $1000$ women misstate the accepted denominator.
\[\boxed{\text{Accidental pregnancies per 100 woman-years of exposure}}\]