📄 survan.doc
字号:
SURVAN Version 2.0 August 1990 Non-Parametric Survival Analysis Kaplan-Meier / Log Rank / Gehan-Breslow Proportional Hazards / Logistic Regression Barry W. Brown James Lovato M. Elizabeth Rozell David Tuttle This work was supported by grant CA-16672 from the National Cancer Institute. Copyright 1988 for The University of Texas, M.D. Anderson Cancer Center Department of Biomathematics, Box 237 1515 Holcombe Blvd. Houston, TX 77030 Contact: BWB at above address or bwb@mdaali.cancer.utexas.edu SURVAN can be freely copied and (noncommercially) distributed. Thanks to Our Supporters The creation of this code was supported in part by the Core Grant CA11672 of the National Cancer Institute of NIH to M. D. Anderson Hospital. Other support was provided by the State of Texas and the Larry and Pat McNeil Professorship in Cancer Research held by Dr. Brown. History SURVAN was written in FLECS (Fortran Language Extended Control Structures) which is a pre-processor similar in concept to RATFOR. The original version of FLECS was written by Terry Beyer of the University of Oregon and placed in the public domain. We added support for Fortran 77 and some new control structures. Those interested in acquiring FLECS should contact Barry Brown at the address below. The Fortran produced by FLECS is as ugly as that produced by RATFOR. The code was polished using pol from Toolpack1.2, a public domain package that can be acquired from NAG or via anonymous ftp to lambda.lanl.gov (128.165.4.4). Legalities of Use of SURVAN This code is copyright 1990 for: The University of Texas M. D. Anderson Cancer Center 1515 Holcombe Boulevard Houston, TX 77096 Correspondence concerning the code should be addressed to: Barry W. Brown Chief, Section of Computer Science Department of Biomathematics, Box 237 at the above address. We want to make this program as widely available as possible. Hence, executable code derived from the source provided, the source itself, and the accompanying documentation may be freely copied and distributed provided that this copyright section accompanies such copies. We encourage the distribution of this code on bulletin boards and as shareware. Modifications and improvements to the code can similarly be freely copied and distributed. Barry Brown would appreciate a copy of any such improvements for incorporation into later versions of SURVAN. Source code may be extracted and used for any non-commercial purpose. Non-commercial means that at most a modest distribution fee is charged, as shareware. We would appreciate acknowledgment of such use of our code. Commercial use of the source code requires written permission from the Department of Biomathematics at the above address. Our policy is generally to grant such permission provided appropriate credit is given to us in the documentation. NO WARRANTY WE PROVIDE ABSOLUTELY NO WARRANTY OF ANY KIND EITHER EXPRESSED OR IMPLIED, INCLUDING BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THIS PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION. IN NO EVENT SHALL THE UNIVERSITY OF TEXAS OR ANY OF ITS COMPONENT INSTITUTIONS INCLUDING M. D. ANDERSON HOSPITAL BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY LOST PROFITS, LOST MONIES, OR OTHER SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA OR ITS ANALYSIS BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY THIRD PARTIES) THE PROGRAM. (Above NO WARRANTY modified from the GNU NO WARRANTY statement.) INTRODUCTION SURVAN performs the following calculations:(1) Kaplan-Meier estimation of survival functions with confidence limits on estimates and quantiles.(2) Log rank test of the equality of survival functions of several groups (optionally over several strata).(3) Gehan Breslow test of the equality of survival functions ofseveral groups (optionally over several strata).(4) Proportional hazards model (Cox model) estimation of coefficients and underlying hazard and survival function.(5) Logistic regression estimation. In addition, SURVAN performs a few general purpose descriptivetasks that may be useful in setting up the above:(6) Frequency distributions.(7) Condescriptive statistics (mean, sd, min, max, etc.). References This document describes the use of the SURVAN computer program anddoes not explain any of the statistical theory behind SURVAN'scalculations. For such explanations, the reader is referred to thefollowing four references. The first two are written at an elementarylevel, the following two are advanced.Gross, A. J. and Clark, V.A. (1975) Survival Distributions:Reliability Applications in the Biomedical Sciences. New York: Wiley.Lee, E. T. (1980) Statistical Methods for Survival Data Analysis.Belmont, CA: Lifetime Learning Publications (A Division of Wadsworth,Inc.)Cox, D. R. and Oakes, D. (1984) Analysis of Survival Data. London:Chapman and Hall.Kalbfleish, J. D. and Prentice, R. L. (1980) The Statistical Analysisof Failure Time Data. New York: Wiley. Methods Most of the methods used by SURVAN are totally standard; thoseused where alternatives exist are noted here. The handling of ties inthe proportional hazards model is as described in:Breslow, N. E. (1975) Covariance Analysis of Censored Survival Data.Biometrics, 30, 89-100. The underlying hazard and survival functions in the proportionalhazards model as estimated by the methods of:Kalbfleish, J. D. and Prentice, R. L. (1973) Marginal likelihoodsbased on Cox's regression and life model. Biometrika, 60, 267-278. Confindence intervals for Kaplan-Meier survival curve estimatesare by the method of:Simon, R. and Lee, Y. (1982) Nonparametric confidence limits forsurvival probabilities and the median. Cancer Treatment Reports, 66,37-42. Other People's Code Used in SURVAN Wherever possible, we use good code written by experts in place ofreinventing a poorer version ourselves. The following are used inSURVAN. All are available by electronic mail from netlib. Forinformation on netlib, send electronic mail whose content is 'sendindex' to netlib@research.att.com. The general function minimization routines written by David Gayare used in proportional hazards and logistic regression. The routineused, DMNH, is actually an update of the one described in thereference.Gay, D. M. (1983) Algorithm 611. Subroutines for UnconstrainedMinimization Using a Model/Trust-Region Approach. ACM Transactions onMathematical Software, 9, 503-524. The public portion of the PORT library was used for dynamic memoryallocation.Fox, P. A., Hall, A. D., and Schryer, N. L. (1978) The PORTMathematical Subroutine Library. ACM Transactions on MathematicalSoftware, 4, 104-126. The zero-finder described below is used in several places inSURVAN.Bus, J. C. P. and Dekker, T. J. (1975) Two Efficient Algorithms withGuaranteed Convergence for Finding a Zero of a Function. ACMTransactions on Mathematical Software, 1, 330-345. An Annotated Computer Run SURVAN Version 1.0 December 1990 Non-Parametric Survival Analysis Barry W. Brown James Lovato M. Elizabeth Rozell David Tuttle This work was supported by grant CA-16672 from the National Cancer Institute. Copyright 1988 for The University of Texas, M.D. Anderson Cancer Center Department of Biomathematics, Box 237 1515 Holcombe Blvd. Houston, TX 77030 Contact: BWB at above address or bwb@mdaali.cancer.utexas.edu SURVAN can be freely copied and (noncommercially) distributed.Press Return / Enter to continue:?Please enter a name of up to 8 characters for this run of SURVAN.sexampleEnter: (0) for no echo recording (1) To echo the entire session to a file (2) To echo the input only to a file (3) For functions (1) and (2)? 3Echo of session will be written to file: "sexample.ALLECHO "Echo of input will be written to file: "sexample.INECHO " ===== Begin Commentary ===== The user has an option to produce a transcript of the SURVANsession. Option 1 records everything seen on the screen includinguser prompts and user input. Option 2 records only the user input. SURVAN names files using the user supplied name followed by aperiod followed by a character string indicating the type of filebeing written. In MS-DOS, the length of the character stringfollowing the period is limited to three. ===== End Commentary =====Enter: (1) For results to be displayed at the terminal (2) For results to be written to a file (3) For both (1) and (2)? 3Results will be written to file: "sexample.RESULTS "Do you want page breaks in your results file? (Y/N)? y ===== Begin Commentary ===== Generally but not always, the results should be written to afile whether or not they are displayed to the terminal. The file canbe printed for reviewing the results. Also, SURVAN has no plotcapabilities so the user might extract Kaplan Meier estimates with atext editor and plot them using a different program. Selecting page breaks causes the results fileto be broken intopages at logical places. Page breaks are useful when the results fileis to printed. ===== End Commentary =====Do you want to restore a state from a previous run? (y/n)? n ===== Begin Commentary ===== SURVAN can read a state file which has been produced by a SAVEcommand on a previous SURVAN run. The state file contains thevariable names, definitions of survival variable, group, strata, andcompulsory variables as they were when the status was saved. Theywill be restored if the y option is chosen here. ===== End Commentary =====***** ENTER DATA *****Do you want to use an existing binary file (y/n)? nEnter name of binary file to receive a copy of the data.Specify file name or * for terminal output:? sexample.binCreated file sexample.bin -- -- P(roceed) or R(etry file specification)?? p ===== Begin Commentary ===== When SURVAN reads a data file, it produces a binary (machine butnot human readable) file whose name is specified by the user. Althoughthere is an option to type * for terminal output this option shouldnever be chosen for a binary file. Reading a binary file is fasterthan reading a raw data file, so if one is available from a previousSURVAN run, the user may want to use it.
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -