Classify row-level detection events — classify_detection

Appends CONCERT-owned detection fields. result_flag is driven only by numeric result, positivity, and the optional reporting limit comparison. Lab qualifiers, parser qualifiers such as "<", and narrative raw result text never change result_flag or detection_event_class; any non-empty qualifier value only raises follow-up/review context.

Usage

classify_detection_events(
  df,
  result_col = "result_value",
  reporting_limit_col = "reporting_limit_value",
  uncertainty_col = "uncertainty_value",
  qualifier_col = "qualifier",
  parser_qualifier_col = "result_qualifier",
  raw_result_col = "raw_result_text",
  measurement_type_col = "measurement_type",
  coverage_col = "uncertainty_coverage",
  result_unit_col = "result_unit",
  detect_operator = c(">", ">=")
)

Arguments

df: Data frame of parsed/harmonized observations.
result_col: Column containing numeric result values.
reporting_limit_col: Optional column containing numeric reporting limits.
uncertainty_col: Optional column containing numeric uncertainty values.
qualifier_col: Optional source qualifier column or columns.
parser_qualifier_col: Optional parser qualifier column, e.g. from parse_numeric_results().
raw_result_col: Optional raw result text column, preserved for context.
measurement_type_col: Optional measurement type column.
coverage_col: Optional uncertainty coverage column.
result_unit_col: Optional harmonized/original result unit column used to conservatively infer radiological rows from units such as pCi or pCi/L.
detect_operator: Reporting-limit comparison for result_flag. Default ">" means result == reporting_limit is non-detect.

Value

df with generated detection fields appended.