1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
|
#' Convert Between Odds Ratios and Risk Ratios
#'
#' @param OR,RR Risk ratio of `p1/p0` or Odds ratio of `odds(p1)/odds(p0)`,
#' possibly log-ed. `OR` can also be a logistic regression model.
#' @param p0 Baseline risk
#' @param ... Arguments passed to and from other methods.
#' @inheritParams oddsratio_to_d
#' @inheritParams cohens_d
#'
#' @return Converted index, or if `OR` is a logistic regression model, a
#' parameter table with the converted indices.
#'
#' @family convert between effect sizes
#' @seealso [oddsratio()] and [riskratio()]
#'
#' @examples
#' p0 <- 0.4
#' p1 <- 0.7
#'
#' (OR <- probs_to_odds(p1) / probs_to_odds(p0))
#' (RR <- p1 / p0)
#'
#' riskratio_to_oddsratio(RR, p0 = p0)
#' oddsratio_to_riskratio(OR, p0 = p0)
#'
#' m <- glm(am ~ factor(cyl),
#' data = mtcars,
#' family = binomial()
#' )
#' oddsratio_to_riskratio(m, verbose = FALSE) # RR is relative to the intercept if p0 not provided
#' @references
#'
#' Grant, R. L. (2014). Converting an odds ratio to a range of plausible
#' relative risks for better communication of research findings. Bmj, 348,
#' f7450.
#'
#' @export
oddsratio_to_riskratio <- function(OR, p0, log = FALSE, verbose = TRUE, ...) {
UseMethod("oddsratio_to_riskratio")
}
#' @export
oddsratio_to_riskratio.numeric <- function(OR, p0, log = FALSE, verbose = TRUE, ...) {
if (log) OR <- exp(OR)
RR <- OR / (1 - p0 + (p0 * OR))
if (log) RR <- log(RR)
return(RR)
}
#' @export
oddsratio_to_riskratio.default <- function(OR, p0, log = FALSE, verbose = TRUE, ...) {
mi <- .get_model_info(OR, ...)
if (!mi$is_binomial || !mi$is_logit) {
insight::format_error("Model must be a binomial model with a logit-link (logistic regression).")
}
RR <- parameters::model_parameters(OR, exponentiate = !log, effects = "fixed", ...)
RR$SE <- NULL
RR$z <- NULL
RR$df_error <- NULL
RR$p <- NULL
used_intercept <- missing(p0)
if (used_intercept) {
p0 <- RR[["Coefficient"]][RR$Parameter == "(Intercept)"]
if (!log) p0 <- log(p0)
p0 <- plogis(p0)
if (verbose) {
insight::format_warning(
"'p0' not provided.",
sprintf(
"RR is relative to the intercept (p0 = %s) - make sure your intercept is meaningful.",
insight::format_value(p0)
)
)
}
}
trans_cols <- colnames(RR) %in% c("Coefficient", "CI_low", "CI_high")
RR[, trans_cols] <-
lapply(RR[, trans_cols, drop = FALSE],
oddsratio_to_riskratio,
p0 = p0, log = log
)
if (verbose && any(c("CI_low", "CI_high") %in% colnames(RR))) {
insight::format_alert("CIs are back-transformed from the logit scale.")
}
RR[RR$Parameter == "(Intercept)", "Coefficient"] <- p0
RR[RR$Parameter == "(Intercept)", c("CI_low", "CI_high")] <- NA
if (!used_intercept) {
RR[RR$Parameter == "(Intercept)", "Parameter"] <- "(p0)"
}
attr(RR, "coefficient_name") <- if (log) "Log-RR" else "Risk Ratio"
return(RR)
}
#' @rdname oddsratio_to_riskratio
#' @export
riskratio_to_oddsratio <- function(RR, p0, log = FALSE, verbose = TRUE, ...) {
if (log) RR <- exp(RR)
OR <- RR * (1 - p0) / (1 - RR * p0)
if (log) OR <- log(OR)
return(OR)
}
|
