Examine very high percentile height and weight simulations

[eedrummer]

Requested Feature

Height and weight simulation under age 20 in Synthea is dependent on the CDC growth charts. To perform the simulation, steps often depend on converting an actual value into a percentile and vice-versa. CDC supplies data and a method for performing these conversions.

The method of conversion appears to break down at very high percentiles. It may be the same for low percentiles, too.

The concern is that the conversion from a high percentile into an actual value may result in a biologically implausible value.

As an example, using the data for females at 240 months, converting the 99.73th percentile into a BMI provides a result of ~387. While obviously not authoritative, a Wikipedia listing claims the highest BMI recorded as 251.1.

Another example is a test in the source code for converting percentile to height. The test looks for the 97th percentile height for males at 20 months. The return value is ~90.7cm. This differs from the actual 97th percentile listed in the data provided by CDC (~50.8cm). It also seems like a very unlikely value as the 97th percentile at 36 months is ~52.6cm.