Learn R Programming

Epi (version 2.56)

steno2: Clinical trial: Steno2 baseline and follow-up.

Description

Steno-2 was a clinical trial conducted at Steno Diabetes Center 1993-2001. The intervention was intensified treatment versus conventional treatment of diabetes patients with micro-albuminuria. The datsets here concern the extended follow-up of the trial population till 2015. Three files are provided: steno2 with one record per person, st2clin with one record per clinical visit and st2alb with one record per transition between states of albuminuria.

These dataset are entirely simulated, but designed to give approximately the same results as the original.

Usage

data("steno2")
       data("st2clin")
       data("st2alb")

Arguments

Format

steno2 is a data frame with 160 observations on the following 14 variables:

id

person id, numeric

allo

Original trial allocation, a factor with levels Int Conv

sex

Sex, a factor with levels F M

baseCVD

0/1 indicator of preexisting CVD at baseline

deathCVD

0/1 indicator whether cause of death was CVD

doBth

Date of birth, a Date

doDM

Date of diabetes diagnosis, a Date

doBase

Date of entry to study, a Date

doCVD1

Date of 1st CVD event, a Date

doCVD2

Date of 2nd CVD event, a Date

doCVD3

Date of 3rd CVD event, a Date

doESRD

Date of end stage renal disease, a Date

doEnd

Date of exit from follow-up, a Date

doDth

Date of death, a Date

st2clin is data frame with 750 observations on clinical measurements at different clinical visits:

id

person id, numeric

doV

Date of clinical visit, a Date

a1c

Glycosylated hemoglobin, mmol/mol

chol

Total cholesterol, mg/mol

crea

Creatinine, mg/mol

st2alb is data frame with 307 observations of changes in complication (albuminuria) state

id

person id, numeric

doTr

Date of transition, a Date

state

State of albuminuria, factor with levels Norm, Mic, Mac. All persons begin in the state Micro-albuminuria.

Details

The data are not the original; all values of measurements and dates have been randomly perturbed, to prevent identifiability of individuals. Analysis of these data will give only (very) approximately the same results as in the published article, and only some of the aspects of data are included.

References

P. Gaede, J. Oellgaard, B. Carstensen, P. Rossing, H. Lund-Andersen, H. H. Parving & O. Pedersen: Years of life gained by multifactorial intervention in patients with type 2 diabetes mellitus and microalbuminuria: 21 years follow-up on the Steno-2 randomised trial. Diabetologia (2016), 59, pp 2298-2307

Examples

Run this code
data(steno2)
data(st2alb)
L2 <- Lexis( entry = list(per = doBase,
                          age = doBase - doBth),
              exit = list(per = doEnd),
       exit.status = factor(deathCVD + !is.na(doDth),
                            labels=c("Mic","D(oth)","D(CVD)")),
                id = id,
              data = cal.yr(steno2) )
summary(L2)
#
# Cut at intermediate transitions
cut2 <- data.frame(lex.id = st2alb$id,
                      cut = cal.yr(st2alb$do),
                new.state = st2alb$state)
L3 <- rcutLexis(L2, cut2)
summary(L3)
#
# no direct transitions Mic <-> Mac allowed, so put a cut in between:
dd <- subset(L3, (lex.Cst == "Mac" & lex.Xst =="Norm") |
                 (lex.Cst =="Norm" & lex.Xst == "Mac"))
# artificial visits to the middle state Mic: 
cut3 <- data.frame( lex.id = dd$lex.id,
                       cut = dd$per + dd$lex.dur/2,
                 new.state = "Mic")
L4 <- rcutLexis(L3, cut3)
summary(L4)
#
# Show all transitions
boxes(L4, boxpos = list(x = c(15,15,15,85,85),
                        y = c(50,15,85,25,75)),
          show.BE = TRUE, scale.R = 1000,
          cex=0.8, pos.arr=0.7, font=1, font.arr=1)

Run the code above in your browser using DataLab