mic_uncensor: Uncensor MICs

In MIC: Analysis of Antimicrobial Minimum Inhibitory Concentration Data

Uncensor MICs

Description

Uncensor MICs

Usage

mic_uncensor(
  mic,
  method = "scale",
  scale = 2,
  ab = NULL,
  mo = NULL,
  distros = NULL
)

Arguments

mic	vector of MICs to uncensor; will be coerced to MIC using AMR::as.mic
method	method to uncensor MICs (scale, simple, or bootstrap)
scale	scalar to multiply or divide MIC by (for method = scale)
ab	antibiotic name (for method = bootstrap)
mo	microorganism name (for method = bootstrap)
distros	dataframe of epidemiological distributions (only used, optionally, for method = bootstrap)

Details

Censored MIC data is generally unsuitable for modelling without some conversion of censored data. The default behaviour (method = scale) is to halve MICs under the limit of detection (<=) and double MICs above the limit of detection (>). When used with method = simple, this function effectively just removes the censoring symbols, e.g., <=2 becomes 2, and >64 becomes 64.

The bootstrap method is the more complex of the three available methods. It attempts to use a second (uncensored) MIC distribution to sample values in the censored range. These values are then used to populate and uncensor the MIC data provided as input (mic). The second (uncensored) MIC distribution is ideally provided from similar experimental conditions. Alternatively, epidemiological distributions can be used. These distributions should be provided as a dataframe to the distros argument. The format for this dataframe is inspired by the EUCAST epidemiological distributions, see: https://www.eucast.org/mic_and_zone_distributions_and_ecoffs. The dataframe should contain columns for antimicrobial (converted using AMR::as.ab), organism (converted using AMR::as.mo), and MIC concentrations. An example is provided in the 'ecoffs' dataset available with this pacakge. Currently, only Escherichia coli is available in this dataset. Each observation (row) consists of the frequency a particular MIC concentration is observed in the distribution. If such a dataframe is not provided to distros, the function will attempt to use 'ecoffs', but remains limited to E. coli.

Value

vector of MICs in AMR::mic format

References

https://www.eucast.org/mic_and_zone_distributions_and_ecoffs

Examples

mic_uncensor(c(">64.0", "<0.25", "8.0"), method = "scale", scale = 2)

MIC documentation built on April 12, 2025, 2:26 a.m.