SARC-F Score

Scope

Compute the 5-item SARC-F sarcopenia screening score (Strength, Assistance in walking, Rise from chair, Climb stairs, Falls). Items are expected on a 0–2 scale; total ranges 0–10. Flags sarc_f_high_risk when total ≥ 4. Includes mapping validation, NA policies, optional extreme screening/capping, and non-numeric coercion safeguards.

When to use this

You have SARC-F questionnaire responses and need the total score plus a high-risk flag.
You want controlled handling of missing or miscoded responses (0–2 expected) with optional capping.
You need outputs suitable for downstream modeling or screening summaries.

What you need (inputs & options)

Argument	Purpose / Options	Notes
data	Data frame/tibble with SARC-F responses	Columns mapped via `col_map`
col_map	Named list mapping required items	strength, walking, chair, stairs, falls
na_action	Missing-data policy	“keep” (default), “omit”, “error”, “ignore”, “warn”
verbose	Emit progress messages	Default FALSE

Required columns (col_map): strength, walking, chair, stairs, falls (expected 0/1/2). Mapped columns must exist.

Units: None. Inputs should already be coded 0–2; no transformation beyond coercion/capping/NA handling.

Handling and expectations

Validation & coercion: mapped columns must exist; non-numeric inputs are coerced to numeric with warnings if NAs are introduced; non-finite values become NA.
Missingness: keep/ignore/warn propagate NA to the total; omit drops rows with any required NA; error aborts if missing.
Out-of-range check: a warning is issued if values fall outside 0–2.
Scoring: sarc_f_score sums the five items; sarc_f_high_risk is TRUE when score ≥ 4, FALSE otherwise, NA if score is NA.
Empty result: if na_action = "omit" drops all rows, returns a zero-row tibble with expected columns.

Outputs

Returns a tibble with sarc_f_score (0–10, or NA) and sarc_f_high_risk (TRUE/FALSE/NA).
NA arises from missing inputs, NA introduced during coercion, NA set by extreme handling, or dropped rows under omit.

Worked example 1: Basic usage (keep NAs)

library(HealthMarkers)
library(tibble)

df <- tibble::tibble(
  Strength = c(0, 1, 2, NA),
  Walking  = c(0, 0, 2, 1),
  Chair    = c(0, 1, 2, 1),
  Stairs   = c(0, 1, 2, 1),
  Falls    = c(0, 0, 1, 0)
)

sarc_f_score(
  data = df,
  col_map = list(
    strength = "Strength",
    walking  = "Walking",
    chair    = "Chair",
    stairs   = "Stairs",
    falls    = "Falls"
  ),
  na_action = "keep",
  verbose = TRUE
)
#> # A tibble: 4 × 2
#>   sarc_f_score sarc_f_high_risk
#>          <dbl> <lgl>           
#> 1            0 FALSE           
#> 2            3 FALSE           
#> 3            9 TRUE            
#> 4           NA NA

Interpretation: Rows with complete items return scores; the row with an NA item yields sarc_f_score = NA and sarc_f_high_risk = NA.

Worked example 2: Detect miscoded values

df2 <- tibble::tibble(
  str = c(0, 3, -1),
  walk = c(1, 2, 1),
  chair = c(0, 2, 1),
  stairs = c(0, 2, 2),
  falls = c(0, 0, 1)
)

# Pre-recode out-of-range values before calling
df2 <- df2 %>% dplyr::mutate(dplyr::across(everything(), ~ ifelse(. < 0 | . > 2, NA_real_, .)))

sarc_f_score(
  data = df2,
  col_map = list(strength = "str", walking = "walk", chair = "chair", stairs = "stairs", falls = "falls"),
  na_action = "keep",
  verbose = TRUE
)
#> # A tibble: 3 × 2
#>   sarc_f_score sarc_f_high_risk
#>          <dbl> <lgl>           
#> 1            1 FALSE           
#> 2           NA NA              
#> 3           NA NA

Interpretation: Out-of-range values are set to NA before summing; scores with NA items return NA.

Troubleshooting & common pitfalls

Missing columns: ensure every required col_map key is mapped to an existing column.
Non-numeric inputs: coerced with warnings; resulting NAs propagate unless omit is used.
Out-of-range values: a warning is emitted; recode out-of-range items to NA before calling.
All NA scores: usually due to missing items or coercion to NA.

Verbose diagnostics

Enable verbose output to inspect column mapping, row counts, and result summaries during QC:

old_opt <- options(healthmarkers.verbose = "inform")
sarc_f_score(
  data.frame(Strength = 1, Walking = 1, Chair = 1, Stairs = 0, Falls = 0),
  col_map = list(strength="Strength", walking="Walking", chair="Chair", stairs="Stairs", falls="Falls"),
  verbose = TRUE
)
#> sarc_f_score(): reading input 'data' — 1 rows × 5 variables
#> sarc_f_score(): col_map (5 columns — 5 specified)
#>   strength          ->  'Strength'
#>   walking           ->  'Walking'
#>   chair             ->  'Chair'
#>   stairs            ->  'Stairs'
#>   falls             ->  'Falls'
#> sarc_f_score(): computing markers:
#>   sarc_f_score      [0-10 sum]
#>   sarc_f_high_risk  [score >= 4]
#> sarc_f_score(): results: sarc_f_score 1/1, sarc_f_high_risk 1/1
#> # A tibble: 1 × 2
#>   sarc_f_score sarc_f_high_risk
#>          <dbl> <lgl>           
#> 1            3 FALSE
options(old_opt)

Tips for best results

Recode out-of-range items to NA before calling when data may be noisy.
Choose na_action = "omit" when you need complete-case scores for modeling; use keep/warn for exploratory summaries where partial data is expected.
Treat sarc_f_high_risk as a screening flag (score ≥ 4); follow up with clinical assessment as appropriate.

Validation notes

Score is a simple sum of the five items; risk flag is score >= 4 when the score is finite.
No unit conversion or weighting is applied; inputs should already reflect the standard SARC-F coding.