The data set contains observed levels of ICF categories from the (comprehensive) ICF Core Set for chronic wide spread pain (CWP) and a physical health component summary measure for n = 420 patients.
The data frame has 420 rows and 68 columns. The first 67 columns contain observed levels of ICF categories from the (comprehensive) ICF Core Set for chronic wide spread pain (CWP). In the last column, the physical health component summary measure is given. Each row corresponds to one patient with CWP. ICF categories have discrete ordinal values between 0 and 4 (columns 1 - 50 and 67), or between -4 and 4 (columns 51 - 66). See the given references for details.
The original data set contained some missing values, which have been imputed
using R package
The data were collected within the study Validation of ICF Core Sets for chronic conditions, which was a collaboration effort between the ICF Research Branch of the collaborating centers for the Family of International Classifications in German, the Classification, Terminology and standards Team from the World Health Organization and the International Society for Physical and Rehabilitation Medicine.
Special thanks go to the following participating study centers: Ankara University, Turkey; Azienda Ospedaliera di Sciacca, Italy; Donauspital, Vienna, Austria; Drei-Burgen-Klinik, Bad Muenster, Germany; Edertal Klinik, Bad Wildungen, Germany; Fachklinik Bad Bentheim, Germany; Hospital das Clinicas, School of Medicine, University of Sao Paulo, Brazil; Hospital San Juan Bautista, Catamarca, Argentina; Istituto Scientifico di Montescano, Italy; Istituto Scientifico di Veruno, Italy; Kaiser-Franz-Josef-Spital, Vienna, Austria; Klinik am Regenbogen, Nittenau, Germany; Klinik Bavaria Kreischa, Germany; Klinik Hoher Meissner, Bad Sooden-Allendorf, Germany; Klinikum Berchtesgadener Land, Schoenau, Germany; Kuwait Physical Medicine and Rehabilitation Society, Safat, Kuwait; National Institute for Medical Rehabilitation, Budapest, Hungary; Neuro-Orthopaedisches Krankenhaus und Zentrum fuer Rehabilitative Medizin Soltau, Germany; Praxis fuer Physikalische Medizin und Rehabilitation, Goettingen, Germany; Rehabilitationsklinik Seehof der Bundesversicherungsanstalt fuer Angestellte, Teltow, Germany; Rehaklinik Rheinfelden, Switzerland; Spanish Society of Rheumatology, Madrid, Spain; University Hospital Zurich, Switzerland; University of Santo Tomas, Quelonchy, Philippines.
Most special thanks go to all the patients participating in the study.
If you use the data, please cite the following two references.
Cieza, A., G. Stucki, M. Weigl, L. Kullmann, T. Stoll, L. Kamen, N. Kostanjsek, and N. Walsh (2004). ICF Core Sets for chronic widespread pain. Journal of Rehabilitation Medicine, Suppl. 44, 63-68.
Gertheiss, J., S. Hogger, C. Oberhauser and G. Tutz (2011). Selection of ordinally scaled independent variables with applications to international classification of functioning core sets. Journal of the Royal Statistical Society C (Applied Statistics), 60, 377-395.
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
# load the data data(ICFCoreSetCWP) # available variables names(ICFCoreSetCWP) # adequate coding of x matrix (using levels 1,2,...) p <- ncol(ICFCoreSetCWP) - 1 n <- nrow(ICFCoreSetCWP) add <- c(rep(1,50),rep(5,16),1) add <- matrix(add,n,p,byrow=TRUE) x <- ICFCoreSetCWP[,1:p] + add # make sure that also a coefficient is fitted for levels # that are not observed in the data addrow <- c(rep(5,50),rep(9,16),5) x <- rbind(x,addrow) y <- c(ICFCoreSetCWP$phcs,NA) # some lambda values lambda <- c(600,500,400,300,200,100) # smoothing and selection modelICF <- ordSelect(x = x, y = y, lambda = lambda) # results plot(modelICF) # plot a selected ICF category (e.g. e1101 'drugs') # with adequate class labels plot(modelICF, whichx = 51, xaxt = "n") axis(side = 1, at = 1:9, labels = -4:4)