Skip to contents

Compute Expected Blood Pressure (\(\mu\))

Source: R/bp_family.R
bp_expected.Rd

Returns the expected blood pressure (mean/50th percentile) based on CYP sex, age, and height z-score. Currently, only NHBPEP Fourth Report regression models are available for use for children and adolescents (0-17 years). For ref = "NICE/BHF" or people >17 years old, expected BP is not defined and NA is returned.

Usage

bp_expected(..., .quiet = FALSE)

Arguments

...

Arguments passed on to .bp_validate_demo_input

bp_type

Character; "systolic" or "diastolic".

sex

Vector of sex codes in your data. Will be mapped to "male" or "female" using male_code and female_code. Values not matching these codes become NA.

male_code,female_code

Scalar values that identify the male and female codes in sex (e.g., 1 and 2, or "M" and "F").

height_z

Numeric vector of height z-scores. Must already be z-transformed (e.g., US CDC, UK-WHO).

height_limit

Positive numeric; z-scores with |z| > height_limit are treated as invalid and set to NA. Default 5.

age_years

Numeric vector of ages in years.

ref

Reference: "Fourth Report" (children/adolescents, 0-17) or "NICE/BHF" (adults, >17).

bp_value

Optional numeric vector of observed BP (mmHg); only used here to check that its length matches the other inputs.

.quiet

Logical; suppress validation messages (default FALSE).

Value

A numeric vector of expected BP values in mmHg (\(\mu\)). Returns NA for:

  • NICE/BHF reference (adults)

  • Ages outside reference range

  • Missing or invalid sex

  • Extreme height z-scores

Height Reference Standards for the Fourth Report

Fourth Report Regression Model:

The Fourth Report uses regression models with age (centred at 10 years) and height z-score polynomial terms (up to 4th degree) to calculate expected BP:

$$\mu = \alpha + \sum_{j=1}^{4} \beta_j (Age-10)^j + \sum_{k=1}^{4} \gamma_k (Z_{ht})^k$$

BMI References:

The Fourth Report was developed and validated using US CDC 2000 growth references for height percentiles. Your height z-scores should ideally be based on US CDC reference. However, UK clinical practice and NPDA audit analysis use UK-WHO growth charts as standard (historically the NPDA used the British 1990), so using the UK-WHO is also considered acceptable in UK contexts (based on the assumption that the two populations are similar enough).

To our knowledge, the Fourth Report BP percentiles have not been validated using UK-WHO height references. The clinical impact of applying Fourth Report regression coefficients with UK-WHO (rather than US CDC) height z-scores has not been formally evaluated. Users should be aware of this methodological discrepancy when interpreting results or comparing to US-based implementations.

Calculating Height Z-Scores:

This package requires height z-scores as input and does not calculate them internally. Two approaches I recommended for obtaining height z-scores are:

  1. childsds R package (https://mvogel78.r-universe.dev/childsds), which provides SDS/z-score calculations based on multiple growth standards (including UK–WHO, US CDC references).

  2. RCPCHGrowth Python library (https://growth.rcpch.ac.uk/products/python-library/), which is the official RCPCH implementation of UK-WHO growth charts, with API accessible from R.

References

See also

Other BP functions: .bp_message(), .bp_resolve_limit(), .bp_validate_demo_input(), bp_assess(), bp_category(), bp_family, bp_relative()

Examples

# Expected systolic BP for a 10-year-old boy at median height, and other some other children:
bp_expected(
  bp_type = "systolic",
  sex = c(1, 2, 1),
  male_code = 1, female_code = 2,
  height_z = c(0, 0, 1.5),
  age_years = c(10, 12, 8),
  ref = "Fourth Report"
)
#> [1] 102.1977 105.8496 102.5655

# Adults: returns NA with an informative message
bp_expected(
  bp_type = "diastolic",
  sex = "M", male_code = "M", female_code = "F",
  height_z = 0, age_years = 25, ref = "NICE/BHF"
)
#> NICE/BHF only provides category guidelines, expected BP not calculated.
#> [1] NA