OCMAP-PLUS for Windows A Microsoft Windows-Based Program

OCMAP-PLUS for Windows A Microsoft ® Windows-Based Program for Comprehensive Analysis of Occupational Cohort Data Ada O. Youk, Ph. D Gary M. Marsh, Ph. D Graduate School of Public Health Department of Biostatistics April 8, 2003 University of Pittsburgh 1

OCMAP Occupational Cohort Mortality Analysis Program n n - Descriptive analysis of cohort death rates and SMRs by levels of demographic and selection variables Descriptive analysis of exposure-response for selected agents (alone or in combination) and for selected causes of death Examine variability in SMRs as based on geographic-specific general populations Provide input data for other OCMAP program modules or independent statistical programs (e. g. , SAS, S-PLUS, Stata) Poisson regression modeling of cohort death Relative risk regression modeling (Cox models) University of Pittsburgh 2

Evolution of OCMAP (1979 -80) OCMAP-PC (1986) OCMAP-PLUS (1994 -) OCMAP-Plus for Windows (2000 -) University of Pittsburgh 3

Rate Files n n Mortality or incidence rates from any geographic region can be used Must be specific format to be compatible with OCMAP University of Pittsburgh 4

OCMAP User Community (approximate figures) Total OCMAP (users and purchasers ) Domestic 128 91 Academic (38), Government (12), Private (41) Foreign 37 Canada (6), France (2), Hong Kong (1), Italy (7), Netherlands (2), South Africa (1), Sweden (6), United Kingdom (12) University of Pittsburgh 5

OCMAP Primary Limitations of Earlier Versions § Did not permit analysis in relation to multiple/diverse work history and exposure measures (no more than 2 exposures/run) § Large memory requirements for person-day arrays § Restrictive upper bounds on some processing parameters (e. g. , no. of selection variables) § No death rate file link to geographic area § Need to reprocess person-days in multiple runs § Highly centralized task-based processing § Limited control over report generation § Absence of data exporting for other statistical packages University of Pittsburgh 6

OCMAP-PLUS Features Enables comprehensive, flexible and efficient analysis of incidence/mortality measures in relation to multiple and diverse work history / exposure indicators Separate analytic modules Module 0 Cohort data editing routines, rate file interface Module 1 Cohort mortality or incidence rates and standardized ratios (SIRs and SMRs) Module 2 Cohort proportional mortality and standardized ratios (PMRs) Module 3 Grouped data generator for Poisson or logistic regression Module 4 Risk set generation for relative risk regression (Cox models) University of Pittsburgh 7

OCMAP-PLUS Features § Produces output files for input to standard statistical/graphics software (e. g. , SAS, S-PLUS, Stata) § Produces hard copy reports suitable for publication or presentation § Processes large cohort files efficiently in microcomputer environment University of Pittsburgh 8

OCMAP Methodology Input Cohort Data File Layout Record Type 1 (ID & demographic) Gender/race Start/stop date for person-years Cause of death (ICD code) No. of jobs in work history Birth date Vital status Hire date Record Type 2 (Selection variables - optional) University of Pittsburgh 9

OCMAP Methodology Input Cohort Data File Layout Record Type 3 (Work history and exposure variables) For each job in the work history (hypothetical data) Jobj 1 2 3. . n Dates 1950 -54 1955 -59 1960 -69. . 1980 -84 Job. Code 125 1 376 1 234 1. . 765 1 University of Pittsburgh (Asb) (HCHO) (Dur) (Y, N) (fibers/cc) (ppm) E 1 j__E 2 j E 3 j. . . Ekj 0. 02 1. 05 1. 35. 85 0. 66 1. 50. . . 1. 34. 45 10

OCMAP Plus Other Special Features Proportional allocation of person-years for study members of unknown race (PAM) (Modules 1 -3) § Person-years of members with unknown race allocated to white and nonwhite categories in proportion to the personyears distribution of members with known race § Allocation performed within strata defined by geographic area, sex, age group and time period § Total numbers of allocated members are weighted averages of stratum-specific assignments University of Pittsburgh 11

OCMAP-PLUS Other Special Features Time dependent stratification by active and inactive work status § The variable “employment status” (ES) can potentially confound the association between workplace exposure and health outcome § Person-days allocated at any time point according to whether member was active or inactive (concurrently with other time-dependent variables) § Optional switch allows users to extend the “active” status a period of k days beyond termination date University of Pittsburgh 12

OCMAP-Plus for Windows Features Allows users to input necessary information via a menu-driven graphical window rather than a control file University of Pittsburgh 13

University of Pittsburgh 14

University of Pittsburgh 15

OCMAP-Plus for Windows Features Multiple configurations can be set up and re-used for future runs without editing control files University of Pittsburgh 16

OCMAP-Plus for Windows Features Latency origins can be specified University of Pittsburgh 17

OCMAP-Plus for Windows Features Report print control University of Pittsburgh 18

OCMAP-Plus for Windows Features § Control files and system files are transparent to the user § File management is handled by the program § Capability to convert ‘old’ OCMAP-Plus control files for use with Windows version § Incorporation of the 10 th revision of the ICDA § Capability to modify date ranges for the ICDA cause of death codes § PAM method for imputing race can be performed in one run rather than multiple runs § Selected table printer and selected case lister § Online help feature with search capability University of Pittsburgh 19

OCMAP / MPDS Published Articles Marsh GM and Preininger ME: OCMAP: A User-Oriented Occupational Cohort Mortality Analysis Program. American Statistician 34: 245 -246, 1980. Caplan RJ, Marsh GM, and Enterline PE: A Generalized Effective Exposure Modeling Program for Assessing Dose-Response in Epidemiologic Investigations. Computers and Biomedical Research 16: 587 -596, 1984. Marsh GM, Ehland J, Paik M, Preininger M, Caplan R: OCMAP/PC: A User Oriented Cohort Mortality Analysis Program for the IBM PC. The American Statistician 40: 308 -309, 1986. Marsh GM, Co-Chien H, Rao BR, and Ehland J: OCMAP: Module 6 - A New Computing Algorithm for Proportional Mortality Analysis. American Statistician 43: 127 -128, 1989. Marsh GM, Youk AO, Stone RA, Sefcik S, Alcorn C: OCMAP-PLUS, A Program for the Comprehensive Analysis of Occupational Cohort Data. Journal of Occupational and Environmental Health 40: 351 -362, 1998. University of Pittsburgh 20