Convert between IOTF, WHO and CDC prevalence rates for child thinness, overweight and obesity
Child thinness, overweight and obesity are defined as the child's body mass
index (BMI) lying beyond a pre-specified reference cutoff. Three references
are compared here: IOTF (International Obesity Task Force), WHO (World Health
Organization) and CDC (US Centers for Disease Control and Prevention), each
of which have their own cutoffs. ob_convertr takes age-sex-specific
prevalence rates of thinness, overweight and obesity based on one cutoff, and
converts them to rates based on a different cutoff, using a novel estimation
algorithm.
ob_convertr(
prev = 50,
age,
sex,
from,
to,
prev_true = NA,
report = c("vector", "wider", "longer"),
plot = c("no", "density", "compare"),
data = parent.frame()
)prev |
vector of age-sex-specific percentage prevalence rates. |
age |
vector of ages between 2 and 18 years corresponding to each rate. |
sex |
vector of the sexes corresponding to each rate, coded as either 'boys/girls' or 'male/female' or '1/2' (upper or lower case, and only the first character considered). |
from |
the BMI cutoff (see Details) on which the prevalence is based. |
to |
the BMI cutoff (see Details) on which to base the converted prevalence. |
prev_true |
optional vector of known percentage prevalence rates
corresponding to |
report |
character controlling the format of the returned data: 'vector'
for the estimated prevalence rates, 'wider' for the working tibble in wide
format, i.e. the |
plot |
character controlling what if anything is plotted: 'no' for no
plot, 'density' to display the BMI density distributions and cutoffs
corresponding to |
data |
data frame containing |
The IOTF cutoffs correspond to the value of BMI (kg/m^2) at age 18: IOTF35 (morbid obesity), IOTF30 (obesity), IOTF25 (overweight), IOTF18.5 (grade 1 thinness), IOTF17 (grade 2 thinness) and IOTF16 (grade 3 thinness).
The WHO cutoffs correspond to BMI z_scores. Age 5-19 years, WHO+2 (obesity), WHO+1 (overweight) and WHO-2 (thinness). Age 0-5 years, WHO+3 (obesity), WHO+2 (overweight) and WHO-2 (thinness).
The CDC cutoffs correspond to BMI centiles: CDC95 (obesity), CDC85 (overweight) and CDC5 (thinness).
Note 1: the overweight category needs to be analysed as overweight plus obesity. To predict overweight excluding obesity, first calculate predicted overweight plus obesity then subtract predicted obesity.
Note 2: the category labels are harmonised and not necessarily as originally defined.
The conversion algorithm exploits the fact that all three references are based on the LMS method, which allows prevalence to be converted to a common BMI centile and z-score scale.
The algorithm is commutative, which means that converting a prevalence rate from cutoff A to cutoff B and then from B to A returns the original value.
Either the converted prevalence rates or a plot visualizing the
findings, depending on the report and plot settings.
Tim Cole tim.cole@ucl.ac.uk
## convert 10% IOTF overweight prevalence (cutoff IOTF25) in 8-year-old boys
## to the overweight prevalence based on WHO, i.e. cutoff WHO+1
ob_convertr(prev = 10, age = 8, sex = 'boys', from = 'IOTF25', to = 'WHO+1')
## compare the BMI density functions and cutoffs for IOTF25 and WHO+1
ob_convertr(prev = 10, age = 8, sex = 'boys', from = 'IOTF25', to = 'WHO+1', plot = 'density')
#' ## convert IOTF overweight prevalence to WHO overweight prevalence
## and compare with true value - boys and girls age 7-17
## note the need to first add obesity prevalence to overweight prevalence
data(deren)
deren <- within(deren, {
IOTF25 = IOTF25 + IOTF30
`WHO+1` = `WHO+1` + `WHO+2`})
ob_convertr(prev = IOTF25, age = Age, sex = Sex, from = 'IOTF25', to = 'WHO+1',
prev_true = `WHO+1`, data = deren, plot = 'compare')Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.