---
title: "Pediatric AHA Scoring"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Pediatric AHA Scoring}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r setup, include = FALSE}
knitr::opts_chunk$set(collapse = TRUE, comment = "#>")
has_references <- requireNamespace("cdcanthro", quietly = TRUE) &&
  requireNamespace("pedbp", quietly = TRUE)
```

`score_le8()` scores children aged 4 months to 19 years with the pediatric
Life's Essential 8 (LE8) definitions from the American Heart Association's
2022 advisory: the children column of its Table 1, and the suggestions for
younger children in its Supplemental Appendix 2. Adults and children can be
scored in the same call; each row's `age` chooses the definitions, and
`le8_population` records which were used.

This article covers how LE8 scoring changes for children, which inputs to supply, and how to derive growth-chart and blood-pressure percentiles.

```{r packages, message = FALSE}
library(essential8)
library(dplyr)
```

## Pediatric Scoring Changes

Each component uses an age-appropriate rule. Some are defined only from a
certain age; before then, the component is `NA` and does not count toward
`min_components`.

| Component | Pediatric rule | Ages scored |
|---|---|---|
| Diet | MEPA: 6 items (6-23 months), 10 items (2-5 years), all 16 (6-19 years); or population percentiles | 6 months+ (percentiles: 2 years+) |
| Physical activity | Daily active play (1-5 years); weekly moderate or vigorous minutes, not doubled (6-19 years) | 1 year+ |
| Nicotine exposure | Never tried, tried over 30 days ago, current inhaled NDS, current combustible; secondhand smoke at any age | All |
| Sleep | Distance from the age-appropriate range (12-16 h at 4-11 months to 8-10 h at 13-19 years) | All |
| BMI | WHO weight-for-length percentile and rapid gain (under 2); CDC BMI percentile and % of 95th percentile (2-19 years) | All |
| Blood lipids | Pediatric non-HDL bands (<100, 100-119, 120-144, 145-189, 190+ mg/dL) | All |
| Blood glucose | Same as adults | All |
| Blood pressure | AAP 2017 percentile categories (1-12 years); adult bands (13-19 years) | 1 year+ |

AHA suggests that the composite for a child be the average of whatever
components are available and appropriate for measurement. Laboratory values
are often unavailable in children, and the NHANES analysis that accompanied the
advisory averaged as few as three components, so set `min_components` appropriately for your cohort.

## Deriving growth-chart and blood-pressure percentiles

For some pediatric samples, you will need blood-pressure percentiles and/or growth chart data. 

`add_pediatric_references()` derives the percentile inputs from raw
measurements, using the packages [cdcanthro](https://CRAN.R-project.org/package=cdcanthro)
(CDC and WHO growth charts) and [pedbp](https://CRAN.R-project.org/package=pedbp)
(AAP 2017 blood-pressure percentiles). It needs sex, height, and weight, and
uses `age_months` or `age * 12`.

```{r references, eval = has_references}
measurements <- tibble(
  id = c("toddler", "school_age", "teen"),
  age = c(1.5, 8.25, 15.5),
  sex = c("male", "female", "male"),
  height_cm = c(82, 128, 172),
  weight_kg = c(11.5, 27, 64)
)

measurements |>
  add_pediatric_references() |>
  select(id, bmi_percentile, bmi_percent_of_p95, infant_weight_percentile,
         systolic_bp_p90, systolic_bp_p95)
```

Blood-pressure references are returned only for ages 1 to 12, and only when
height is known.

Rapid infant weight gain compares visits, so it cannot be derived from one
row. With longitudinal data, compute it from `infant_weight_z` within each
child:

```{r z-change, eval = FALSE}
visits |>
  add_pediatric_references() |>
  arrange(child_id, age) |>
  mutate(
    infant_weight_z_change = infant_weight_z - lag(infant_weight_z),
    .by = child_id
  )
```

AHA compares against the z-score at prior routine visits, for example since
age 1 month; choose the comparison visit to match your study design.

## School-age children and adolescents (6-19 yrs)

From age 6, diet uses all 16 MEPA items and the adult scoring bands. Weekly moderate and vigorous minutes count once each (adults count vigorous minutes twice). The pediatric nicotine categories describe any nicotine product: `"never"` = never tried, `"former"` = tried but not within 30 days, and `"current"` = combustible use within 30 days. Blood pressure switches to the adult bands at 13, while the other components stay pediatric through age 19.

## A preschool child (2-5 yrs)

At ages 2-5, ten of the MEPA items are scored, at 10 points per item met, with pediatric thresholds:

* vegetables: `green_leafy_vegetables + other_vegetables`, 2+ per day
* fruit: `berries + other_fruit`, 14+ per week (2 per day)
* `whole_grains` 3+ and `olive_oil` 1+ per day
* `fish` 1+ and `nuts` 4+ per week
* `meat`, `butter_cream`, and `sweets_and_pastries` 3 or fewer per week; for
  children under 6, sweets should include sugar-sweetened beverages
* `fast_food` 2 or fewer times per month

`chicken`, `cheese`, `beans`, `alcohol`, and `sex` are not used before age 6
and may be omitted from data containing only young children.

This example also uses active play in minutes per day and blood-pressure
percentiles. The 90th and 95th percentiles (`systolic_bp_p90` and so on) are
the child's reference values for age, sex, and height, in mmHg.

```{r preschool}
preschooler <- tibble(
  id = "preschooler",
  age = 4,
  # Servings per day
  green_leafy_vegetables = 1,
  other_vegetables = 1,
  whole_grains = 2,
  olive_oil = 1,
  # Servings per week
  berries = 7,
  other_fruit = 14,
  sweets_and_pastries = 2,
  meat = 2,
  fish = 1,
  nuts = 2,
  butter_cream = 1,
  # Times per week
  fast_food = 0.25,
  active_play_minutes = 160,
  smoking_status = "never",
  current_inhaled_nds = FALSE,
  secondhand_smoke_home = FALSE,
  sleep_hours = 13.5,
  bmi_percentile = 60,
  bmi_percent_of_p95 = 82,
  systolic_bp = 108,
  diastolic_bp = 60,
  systolic_bp_p90 = 103,
  systolic_bp_p95 = 107,
  diastolic_bp_p90 = 63,
  diastolic_bp_p95 = 67,
  antihypertensive_treatment = FALSE
)

preschooler |>
  score_le8(min_components = 5) |>
  select(mepa_total, ends_with("_score"), n_components, le8_category)
```

Eight of the ten items are met (80 points); `mepa_total` counts items met on
the screener for the child's age. Sleep is half an hour above the 10-13 hour
range (90 points). The systolic pressure of 108 mm Hg is at or above this
child's 95th percentile (107 mm Hg) but below 95th + 12 mm Hg, which is
stage 1 (50 points). Lipids and glucose were not measured, so the composite
averages six components; the warning says so.

## Infants and toddlers (6-23 months)

From 6 to 23 months, diet uses six items: vegetables, fruit, and whole grains
(at least once a day before 12 months, twice a day from 12 months), no sweets
or sugar-sweetened beverages, no fast food, and the milk source recommended for
age. `milk_source` is the one column used only at these ages: `"human"`,
`"cow"`, `"formula"`, or `"other"`. Human milk meets the item before 12 months,
and human or cow's milk from 12 months.

Weight uses the WHO weight-for-length (or BMI-for-age) percentile together
with the change in its z-score since the prior routine visit. An increase of
0.67 or more counts as rapid weight gain.

```{r infant}
infant <- tibble(
  id = "infant",
  age = 0.75,
  age_months = 9,
  green_leafy_vegetables = 0.5,
  other_vegetables = 0.5,
  berries = 3,
  other_fruit = 4,
  whole_grains = 1,
  sweets_and_pastries = 0,
  fast_food = 0,
  milk_source = "human",
  smoking_status = "never",
  current_inhaled_nds = FALSE,
  secondhand_smoke_home = TRUE,
  sleep_hours = 13,
  infant_weight_percentile = 98.5,
  infant_weight_z_change = 0.4
)

infant |>
  score_le8(min_components = 4) |>
  select(mepa_total, ends_with("_score"), n_components, le8_category)
```

`age_months` is optional; when supplied it sets the pediatric age bands more
precisely than `age * 12`. Physical activity and blood pressure are not scored
before age 1, so they are `NA` here regardless of the inputs.

## Where the sources conflict

The advisory's main table and its supplement are not always consistent. `essential8` follows the more specific rule and documents each choice in
`?score_le8`:

* Ages 2-5 use the supplement's 10 points per item met rather than Table 1's
  banded 0-10 scale (for example, 7 items score 70 rather than 80).
* The supplement describes the 16-item MEPA questionnaire as, "used in older
  children and adults", so ages 6-19 use the adult bands.
* At 6-23 months, weight at or above the 99.7th percentile with rapid gain
  scores 0, the supplement's explicit worst state, rather than the 25 its
  listed decrements would give.
* Nineteen-year-olds use the 13-18 year sleep range (8-10 hours), the last
  range AHA lists for the pediatric column.
* The packaged NHANES diet references describe adults, so they are refused
  for children. Use percentiles in `diet_value` or a pediatric reference.