strate: Calculate Incidence Rates from time-to-event data

Description

strate() calculates incidence rates and Corresponding 95\

Usage

strate(data, time, var, ..., fail = NULL, per = 1, digits = 5)

Arguments

data

Dataset

time

person-time variable

var

outcome variable: preferably 1 for event, 0 for censored

...

variables for stratified analysis

fail

a value or values to specify failure event

per

units to be used in reported rates

digits

Rounding of numbers

Details

Rates of event occurrences, known as incidence rates are outcome measures in longitudinal studies. In most longitudinal studies, follow-up times vary due to logistic reasons, different periods of recruitment, delay enrollment into the study, lost-to-follow-up, immigration or emigration and death.

Follow-up time in longitudinal studies

Period of observation (called as follow-up time) starts when individuals join the study and ends when they either have an outcome of interest, are lost-to- follow-up or the follow-up period ends, whichever happens first. This period is called person-year-at-risk. This is denoted by PY in strate function's output and number of event by D.

Rate

is calculated using the following formula: $λ = D / P Y$

Confidence interval of rate

is derived using the following formula:

$95 % C I (r a t e) = r a t e x E r r o r F a c t o r$ $E r r o r F a c t o r (r a t e) = e x p (1.96 / \sqrt{D})$

plot, if TRUE, produces a graph of the rates against the numerical code used for categories of by.

References

Betty R. Kirkwood, Jonathan A.C. Sterne (2006, ISBN:978<U+2013>0<U+2013>86542<U+2013>871<U+2013>3)

Examples

Run this code

# NOT RUN {
# }
# NOT RUN {
## Using the diet data (Clayton and Hills 1993) described in STATA manual
import diet data: require haven package to read dta format.
magrittr package for piping operation
diet <- haven::read_dta("https://www.stata-press.com/data/r16/diet.dta")

diet <- generate(diet, time, (dox - doe) / 365.25)
diet <- replace(diet, time, as.numeric(time))
diet <- generate(diet, age, as.numeric(doe - dob) / 365.25)
diet <- egen(diet, age, c(41, 51, 61, 71), new_var = ageband)
diet <- egen(diet, month, c(3, 6, 8), new_var = monthgrp)

## calculate overall rates and 95% Confidence intervals
strate(diet, time, fail, fail = c(1, 3, 13))

## per 100 unit
strate(diet, time, fail, fail = c(1, 3, 13), per = 100)

## calculate Stratified rates and 95% Confidence Intervals
strate(diet, time, fail, job, fail = c(1, 3, 13))
strate(diet, time, fail, job, ageband, monthgrp, fail = c(1, 3, 13))

## per 100 unit
strate(diet, time, fail, job, ageband, monthgrp, fail = c(1, 3, 13), per = 100)
# }
# NOT RUN {
# }