User's Guide to the SAS Manitoba Multiple Tumour
Contents
1. References 1 Canadian Cancer Society s Steering Committee Canadian Cancer Statistics 2009 Toronto Canadian Cancer Society 2009 2 Fritz A Jack A Parkin DM et al eds nternational Classification of Diseases for Oncology Third Edition Geneva World Health Organization 2000 C 1l C SPAN SAS MB Multiple Tumour Data Extract Users Guide 2011 01 31 rev1 doc January 31 2011 92. 237163 health identification number tumourid tumour ID number best12 100391644 vitalst vital status alive deceased 1 a C 1l C SPAN SAS MB Multiple Tumour Data Extract Users Guide 2011 01 31 rev1 doc January 31 2011 3 PARTNERSH iL eal IE E an ve i CancerCare C SPAN Multiple Tumour Data Extract User s Guide ix orsonventince tr motos ou cancen V MANITOBA Description of the SAS Code 1 Create a working dataset called reg 1 Read in the input registry data may10dat sas7bdat in this example Define diagnosis year variable yydx using the date of diagnosis dxdt Select cases diagnosed from 1992 to 2006 inclusive Restrict to Manitoba residents at diagnosis where postal code at diagnosis dxpc begins with R Resulting SAS dataset reg 1 Invoke Alberta s IARC conversion macro shared by Alberta Health Services if working on CCR data skip this step 2 Define cancer site groupings based on ICDO 3 morph3 morphologies Read in dataset defined above reg 1 Define the morph2 variable by selecting the first 4 digits of the ICDO 3 morphology field morph3 Define the beh variable by selecting the behaviour code 5 digit of the ICDO 3 morphology field morph3 Ifthe first four digits of the ICDO 3 morphology code morph2 is less than 9590 then define cancer site groupings cancer grp according to topography code topog and group them using the topog2 format Ifthe first f
3. Testis C620 C629 Bladder C670 C679 Kidney C649 Brain Excludes other nervous system C710 C719 Thyroid C739 Hodgkin Lymphoma Histology types 9650 9667 Non Hodgkin Lymphoma Histology types 9590 9596 9670 9671 9673 9675 9678 9680 9684 9687 9689 9691 9695 9698 9 02 9705 9708 9709 9714 9719 9727 9729 9823 9827 all sites except C420 C421and C424 Multiple Myeloma Histology type 9731 9732 9734 Leukemia Histology type 9733 9742 9800 9801 9805 9820 9826 9831 9837 9840 9860 9861 9863 9866 9867 9870 9876 9891 9895 9897 9910 9920 9930 9931 9940 9945 9946 9948 9963 9964 9823 9827 for sites C420 C421 and C424 All Cancers All invasive sites and in situ bladder Site is defined by first making the appropriate histology exclusions or inclusions CDO 3 refers to the Third Edition of the International Classification of Diseases for Oncology 2000 Histology types 9050 9055 mesothelioma 9140 Karposi Sarcoma and 9590 9989 leukemia lymphoma and multiple myeloma are excluded from other specific organ sites C 1l C SPAN SAS MB Multiple Tumour Data Extract Users Guide 2011 01 31 rev1 doc January 31 2011 8 CANADIAN PARTNERSHIP P lt PARTENARIAT CANADIEN AGAINST CAMCER w p CONTRE LE CAN i CANCER SURVEILLANCE AND ioc fj ian erCare C SPAN Mutiple Tumour Data Exact User s Guide ssssememms Cancer Care
4. CANADIAN PARTNERSHIP AGAINST CANCER PARTENARIAT CANADIEN CONTRE LE CANCER _CANCER SURVEILLANCE AND EPIDEMIOLOGY NETWORKS RESEAUX DE SURVEILLANCE ET D EPIDEMIOLOGIE DU CANCER A CancerCare MANITOBA User s Guide to the SAS Manitoba Multiple Tumour Data Extraction Program Cancer Surveillance and Prevalence Analytic Network C SPAN Prepared by CancerCare Manitoba for the Canadian Partnership Against Cancer The Cancer Surveillance and Epidemiology Networks have been made possible through a financial contribution from Health Canada provided by the Canadian Partnership Against Cancer The views expressed herein do not necessarily represent the views of the Canadian Partnership Against Cancer nor that of Health Canada Material appearing in this report may be reproduced or copied without permission however the following citation must be used The Cancer Survival and Prevalence Analytic Network C SPAN is an initiative of the Canadian Partnership Against Cancer in collaboration with CancerCare Manitoba CANADIAN PARTNERSHIP 5 e PARTEHARIAT CANADIEN AGAINST CAMCER i p CONTRE LE CANC ER A a CANCER SURVIEILLANCT AAD EPIDEMIOLOGY NETWORKS T arx erCare C SPAN Multiple Tumour Data Extract User s Guide RISEAUX DE SURVEILLANCE ET DEPID MIOLOGIE DU CANCER i MANITOBA Table of Contents Accompanying Documentation ssseeccccssssssececcccsssssccecccsssssseccecccsssssecceccosssssseececosssssseceecosssssseee 3 FR
5. OG ECOL VAL TAOIES S sani niv eR HERE HITS TETTE HUS TE 3 Description of the SAS Code sarson ere ESE RENEE Ea EEEN EEES 4 I Create a working dataset called reg 1 esses nnne nnns 4 2 Define cancer site groupings based on ICDO 3 morph3 morphologies 4 3 Make additional restrictions and clean data eese esee nennen eee 5 4 Define exit date death indicator and death date status essen 2 5 Calculate exact age at diagnosis and survival time using an exact interval macro shared by Latry Ellison at Statishices Canada osa siandamee icu siaeiu a bti betamdasctentcdM vigi ubt etam dM ute U iub 6 6 Define additional variables required for Paul Dickman s macro ssec 6 7 Output data to a permanent datasets sias erre arp eRa tha adi U sta IRR MR d ran exevaasaeaaeraves Table 1 C SPAN Cancer Site Definitions following the site groupings agreed upon by Cancer Surveillance and Epidemiology Networks December 2010 8 PREP EP S M E S cae ccsonciecesdesceccacsstecescucdecciesabecasvesaescavessesaevessesainesiessowessesausuens scenes seseasustecevesncasescatnoseveuseces 9 C 1l C SPAN SAS MB Multiple Tumour Data Extract Users Guide 2011 01 31 rev1 doc January 31 2011 2 SAM ADLAN PARTNERSHIP AGAINST CANCER zw PARTENARIAT CANADIEN CONTRE LE CANCER Y ms prEMIOLOGY NETW ORXES Sf ar Care C SPAN Multiple Tumou
6. RES T arx erCare C SPAN Multiple Tumour Data Extract User s Guide RISEAUX DE SURVEILLANCE ET DEPID MIOLOGIE DU CANCER i MANITOBA Define variable survival time surv day as dur1 surv day dur1 Age at diagnosis is referred to as age and gender is referred to as sex in Paul Dickman s macro Resulting SAS dataset reg 6 7 Output data to a permanent dataset Read in dataset created above reg 6 Resulting SAS dataset reg multumour C 1l C SPAN SAS MB Multiple Tumour Data Extract Users Guide 2011 01 31 rev1 doc January 31 2011 7 SANADAN PARTNERSHIP 3 EMT CANADIEN A GAINST CANCER i po ONTRE E CANCER aan ancerCare C SPAN Multiple Tumour Data Extract User s Guide wiser orsurvettance oto ou cancer V MANITOBA Table 1 C SPAN Cancer Site Definitions iollowing t the site groupings agreed upon by Cancer Surveillance and Epidemiology Networks December 2010 Index tumours a selected using the above extraction process are categorized by cancer site in the following way Cancer Site ICDO 3 Site Histology Type Oral C000 C148 Esophagus C150 C159 Stomach C160 C169 Colorectal Excludes Anus C180 C189 C199 C209 C260 Liver C220 Pancreas C250 C259 Larynx C320 C329 Lung C340 C349 Melanoma C440 C449 Histology Types 8720 8790 Breast C500 C509 Cervix C530 C539 Body of Uterus C540 C549 Excludes uterus not otherwise specified Ovary C569 Prostate C619
7. ataset reg 4 5 Calculate exact age at diagnosis and survival time using an exact interval macro shared by Larry Ellison at Statistics Canada Read in dataset created above reg 4 Parse date variables into components variables for use in exact interval algorithm which returns an imputed mean age or survival time in the case of missing month or day information Call the exact interval macro to compute age at diagnosis Define the variable agedx by rounding the resulting of age at diagnosis dur divided by 365 25 to 3 decimal places lf the integer value of agedx is not equal to the Cancer Registry age at diagnosis dxage then set agedx to the Cancer Registry age Exclude cases with diagnosis date prior to birth date Call the exact interval macro to compute survival time Exclude cases with death date prior to diagnosis date Define the new variable dur1 by rounding the resulting survival time dur to the nearest integer Exclude cases where survival time is zero dur 0 and diagnosis method is autopsy dxmethod autopsy Resulting SAS dataset reg 5 6 Define additional variables required for Paul Dickman s macro Read in dataset created above reg 5 C 1l C SPAN SAS MB Multiple Tumour Data Extract Users Guide 2011 01 31 rev1 doc January 31 2011 6 L CANADIAN PARTNERSHIP 5 e PARTEHARIAT CANADIEN AGAINST CAMCER i p CONTRE LE CANC ER A CANCER SURVIEILLANCT AAD EPIDEMIOLOGY ALTWO
8. h vital status coded as deceased and death date equal to missing vitalst 2 d and deathat Resulting SAS dataset reg 3 4 Define exit date death indicator and death date status Read in dataset created above reg 3 gt Ifthe Vital Status field indicates a person is deceased vitalst d and the death date occurs before or on the end of study date deathat lt 31Dec2006 d then set the exit date to the date of death exit deathdt and set the death status indicator to deceased d 7 and the death date occurs after the end of study date deathat gt 31Dec2006 8 then set the exit date to the end of study date exit 31Dec2006 a and set the death status indicator to alive C 1l C SPAN SAS MB Multiple Tumour Data Extract Users Guide 2011 01 31 rev1 doc January 31 2011 5 L CANADIAN PARTNERSHIP we PARTEHARIAT CANADIEN s AGAINST CAMCER p CONTRE LE CANCER A WEILLANCE T tPIDEMIOLOGY NETWORKS FN al i C Care i Du CANCER C SPAN Multiple Tumour Data Extract User s Guide sis oro 4 MANITOBA d 0 and set the death date status as complete dthdatst C in order to use exact interval macro gt Ifthe Vital Status field indicates a person is alive vitalst a set the exit date to the date of death exit 317Dec2006 d set the death status indicator to alive d 0 and set the death date status as complete dthdatst C in order to use exact interval macro Resulting SAS d
9. our digits of the ICDO 3 morphology code morph2 is greater than or equal to 9590 then define cancer site groupings cancer grp according to morphology morph2 and group them using the morph3f format Select only invasive cancers and in situ bladder cancers Identify melanomas basal and squamous special cases for Non Hodgkin Lymphomas Leukemia and other non specified cancers Exclude mesothelioma and special brain cancer cases from organ specific site groupings C 1l C SPAN SAS MB Multiple Tumour Data Extract Users Guide 2011 01 31 rev1 doc January 31 2011 4 L CANADIAN PARTNERSHIP 5 e PARTEHARIAT CANADIEN AGAINST CAMCER i p CONTRE LE CA CAMC HCER A a CANCER SURVDILLANCE AND EPIDEMIOLOGY NITWCHRES i arx erCare C SPAN Multiple Tumour Data Extract User s Guide _ vx memento ov eancen V MANITOBA Exclude adolescent bone cancer Exclude basal and squamous skin cancer Resulting SAS dataset reg 2 3 Make additional restrictions and clean data Read in dataset created above reg 2 Exclude cases with invalid genital organs le Ovarian cancer in a patient with sex male Select cases with age at diagnosis between 15 and 99 years of age 15 lt dxage lt 99 Exclude cases where the method of diagnosis was by death certificate only DCO dxmethod death cert Exclude cases with missing sex sex Exclude cases with missing date of birth birdatst Y Exclude cases wit
10. r Data Extract User s Guide ven otto ov cance 4 MANITOBA Overview This document describes SAS code seeC SPAN MB Extract Multiple Tumour 92 06 2011 01 31 sas used by the Cancer Surveillance and Prevalence Analytic Network C SPAN to extract and prepare data from the Manitoba Cancer Registry for use in the C SPAN period or cohort survival programs The SAS code selects all primary invasive malignant cancers excluding basal and squamous cell skin cancers and in situ bladder cancers diagnosed during the period 1992 2006 The code also groups tumours following the same site agreed upon by the Cancer Surveillance and Epidemiology Networks in December of 2010 see Table 1 does data cleaning and makes data exclusions and creates new variables appropriate to survival analysis Accompanying Documentation All formats used in C SPAN programs are defined in the accompanying format file C SPAN Formats 2011 01 31 sas The exact interval macro used to compute exact age at diagnosis and survival time is in the accompanying macro C SPAN Exact Interval Macro sas Required Variables The input dataset of cancer cases must include the following variables to run the program Variable Name Description Format Example birdatst birth date status 1 C deathat death date yymmdd10 2001 06 27 dthdatst death date status 1 C dxat diagnosis date yymmdd10 1993 05 19 dxpc postal code at diagnosis 6 R3G2L8 sphin patients scrambled personal 6
Download Pdf Manuals
Related Search