mess/update_buttondown_29JULY2023.md

In arcaldwell49/SimplyAgree: Flexible and Robust Agreement and Reliability Analyses

SimplyAgree Update: v0.2.0

Aaron R. Caldwell

Welcome to version 0.2.0 (July 2023)

I am happy to announce that an early release of the SimplyAgree. As always, detailed documentation of the package can be found the website. This update mainly about two new functions: tolerance_limit and agreement_limit.

Rationale

I have never really liked that agreement limits are calculated from 3 different functions so I have created 1 function, agreement_limit, to calculate the 3 main types of agreement limits included in the package. Additionally, users will have the option of using Bland-Altman style confidence limits or Zou’s MOVER confidence limits.

The tolerance limits are inspired the recent work of Francq[1]. The tolerance limit is essentially a confidence limit on a prediction interval of the difference scores. While the concept might sound confusing the average user can just consider the prediction interval to be an equivalent of the agreement limit and the tolerance limit to be the equivalent of the confidence limits for agreement limits. This is advantageous for 2 reasons:

1. Tolerance and prediction intervals often have better coverage
2. Tolerance limits can be directly calculated from the estimated marginal means from a generalized least squares (GLS) model and are therefore more flexible

Agreement Limits

The agreement limits function is very simliar to the other “agree” functions in the package.

library(SimplyAgree)
data('reps')

# Simple
agreement_limit(x = "x", y ="y", data = reps)

MOVER Limits of Agreement (LoA)
95% LoA @ 5% Alpha-Level
Independent Data Points

   Bias           Bias CI Lower LoA Upper LoA            LoA CI
 0.4383 [-0.1669, 1.0436]    -1.947     2.824 [-3.0117, 3.8884]

SD of Differences = 1.217

# Replicates
agreement_limit(x = "x", y ="y", data = reps, id = "id", data_type = "rep")

MOVER Limits of Agreement (LoA)
95% LoA @ 5% Alpha-Level
Data with Replicates

   Bias           Bias CI Lower LoA Upper LoA            LoA CI
 0.7152 [-1.5287, 2.9591]    -2.232     3.662 [-7.5482, 8.9786]

SD of Differences = 1.5036

# Nested
agreement_limit(x = "x", y ="y", data = reps, id = "id", data_type = "nest")

MOVER Limits of Agreement (LoA)
95% LoA @ 5% Alpha-Level
Nested Data

   Bias           Bias CI Lower LoA Upper LoA            LoA CI
 0.7046 [-1.5572, 2.9664]    -2.153     3.562 [-7.4979, 8.9071]

SD of Differences = 1.4581

Tolerance Limit

The tolerance limit function operates in a similar fashion. The only difference is that it cannot handle data with replicates explicitly.

data('reps')

# Simple
tolerance_limit(x = "x", y ="y", data = reps)

Agreement between Measures (Difference: x-y)
95% Prediction Interval with 95% Tolerance Limits

   Bias           Bias CI Prediction Interval Tolerance Limits
 0.4383 [-0.1669, 1.0436]   [-2.1998, 3.0764] [-2.993, 3.8697]

# Nested
tolerance_limit(x = "x", y ="y", data = reps, id = "id")

Agreement between Measures (Difference: x-y)
95% Prediction Interval with 95% Tolerance Limits

   Bias           Bias CI Prediction Interval   Tolerance Limits
 0.7046 [-1.5571, 2.9663]   [-4.4686, 5.8778] [-8.6099, 10.0191]

Log transformation

Additionally, users can set the argument log_tf to TRUE, and this will log-transform the raw data. This can be advantageous in situations where the error is proportional to the mean (heteroscedascity). The interpretation is slightly different since the data is on the log-scale. When back transformed the differences between the two measures is on the ratio scale (e.g., x/y versus x-y).

tolerance_limit(x = "x", y ="y", 
                data = reps, id = "id",
                log_tf = TRUE)

Agreement between Measures (Ratio: x/y)
95% Prediction Interval with 95% Tolerance Limits

  Bias          Bias CI Prediction Interval Tolerance Limits
 1.117 [0.7251, 1.7202]    [0.4162, 2.9969] [0.1888, 6.6065]

Concluding Remarks

It has taken about half of this year to get tolerance limits into the package. I believe this function will greatly improve the ease of calculating agreement/tolerance limits. Just like the other agree functions there is a check and plot functions/methods so you can check your assumptions and plot your data.

If you have any questions, concerns, or have suggestions please feel free to reach out. I have a “Contact Me” form on my website, and feel free to send a message at any time. I would appreciate any feedback.

I hope you all enjoy the new SimplyAgree!

Cheers everyone,

AC

[1] Francq, B. G., Berger, M., & Boachie, C. (2020). To tolerate or to agree: A tutorial on tolerance intervals in method comparison studies with BivRegBLS R Package. Statistics in Medicine, 39(28), 4334-4349.

arcaldwell49/SimplyAgree documentation built on March 26, 2024, 2:26 p.m.