QualityMeasure (version 2.0.1)

calcSSR: Calculate reliability using split-sample method

Description

This function estimates reliability using the split-sample method.

Usage

calcSSR(
  df = NULL,
  model = NULL,
  entity = "entity",
  y = "y",
  data.type = "binary",
  ctrPerf = controlPerf(),
  ctrRel = controlRel()
)

Value

A list containing:

• entity: list of entities

• n: entity sample sizes

• icc: Spearman-Brown-adjusted intraclass correlation coefficients for each resample

• icc.lb: lower bound on confidence interval for Spearman-Brown-adjusted intraclass correlation coefficients for each resample

• icc.ub: upper bound on confidence interval for Spearman-Brown-adjusted intraclass correlation coefficients for each resample

• est.SSR: reliability estimate based on a single split

• est.PSSR: mean reliability estimate across resamples

If a risk-adjustment model is included then, the outputted list will contain:

• entity: list of entities

• n: entity sample sizes

• icc.oe: Spearman-Brown-adjusted intraclass correlation coefficients for OE ratios for each resample

• icc.oe.lb: lower bound on confidence interval for Spearman-Brown-adjusted intraclass correlation coefficients for OE ratios for each resample

• icc.oe.ub: upper bound on confidence interval for Spearman-Brown-adjusted intraclass correlation coefficients for OE ratios for each resample

• icc.pe: Spearman-Brown-adjusted intraclass correlation coefficients for PE ratios for each resample

• icc.pe.lb: lower bound on confidence interval for Spearman-Brown-adjusted intraclass correlation coefficients for PE ratios for each resample

• icc.pe.ub: upper bound on confidence interval for Spearman-Brown-adjusted intraclass correlation coefficients for PE ratios for each resample

• est.SSR.oe: reliability estimate for OE ratio based on a single split

• est.PSSR.oe: mean reliability estimate for OE ratio across resamples

• est.SSR.pe: reliability estimate for PE ratio based on a single split

• est.PSSR.pe: mean reliability estimate for PE ratio across resamples

Arguments

df

observation-level data; if null, will use the dataframe from the model object

model

model; if null, will use an unadjusted model

entity

data column containing the accountable entity identifier

y

data column containing the outcome variable

data.type

acceptable values are binary for 0/1 data and continuous for continuous data (default: binary)

ctrPerf

parameters to control performance measure calculation

ctrRel

parameters to control reliability estimation

Author

Kenneth Nieser ([e-mail])

References

Nieser KJ, Harris AH. Comparing methods for assessing the reliability of health care quality measures. Statistics in Medicine. 2024 Oct 15;43(23):4575-94.

Examples

# Simulate data
df <- simulateData(n.entity = 50, n.obs = 100, mu = .2, r = .7)

# Calculate reliability
out <- calcSSR(df = df, entity = 'entity', y = 'y', ctrRel = controlRel(n.resamples = 10))
out$est.PSSR

# Distribution of estimates obtained from the permutation sampling.
hist(out$icc)
summary(out$icc)