*CHANGE THE FOLDER NAME (CURRENTLY "MYFOLDER\PMRProgram") TO A PLACE YOU WANT TO STORE YOUR ANALYSES; options formdlim="*"; /*Replaces hard page breaks in the Output Window with rows of *.*/ libname public "C:\Users\MYFOLDER\NOMS\PMRProgram"; %include "C:\Users\MYFOLDER\NOMS\PMRProgram\formats_2020.sas"; /*Reads in SAS formats.*/ %include "C:\Users\MYFOLDER\PMRProgram\formats_ICD-10.sas"; /*Reads in SAS formats for ICD-10 codes.*/ proc contents data=public.public2019; /*To see the variable names, types, lengths, and formats.*/ run; /*************************************************************************************************************************/ /* */ /* The code below creates a macro to calculate unadjusted and indirectly standardized proportionate mortality ratios */ /* (PMR) and 95% confidence intervals (CI) for associations between an individual industry or occupation code or a */ /* group of industry or occupation codes and mortality from an individual cause of death of interest or a group of */ /* causes of death of interest. It also includes several example calls to the macro. Here are definitions and examples */ /* for each macro variable (all of which are provided by the user): */ /* */ /* dataset: the analysis dataset; must be provided by the user */ /* */ /* ifst: IF statement provided by the user to restrict the analysis dataset for sub-group analyses (e.g., females */ /* only, Hispanics only, ages 18-65 year only, etc.); make sure to include the semicolon at the end of the IF */ /* statement; leave blank if no restriction is desired */ /* */ /* ucodvar: variable that contains the underlying cause of death; must be provided by the user */ /* */ /* mcodvars: variables that contain the multiple causes of death; include a space between the variable names (e.g., */ /* Condition_1RA Condition_2RA); leave blank if only underlying cause of death is desired */ /* */ /* numICD: number of International Classification of Diseases, 10th Revision (ICD-10) (National Center for Health */ /* Statistics 2020), codes to be used to define the outcome (i.e., an individual cause of death of interest or */ /* a group of causes of death of interest); minimum value is one; must be provided by the user */ /* */ /* ICD: ICD-10 code(s) to be used to define the outcome (i.e., an individual cause of death of interest or a group */ /* of causes of death of interest); ICD-10 code(s) must be provided with quotes (e.g., "F03"); include a space */ /* between ICD-10 codes if more than one ICD-10 code is desired (e.g., "G112" "G35"); must provide at least one */ /* ICD-10 code; must be provided by the user */ /* */ /* numindocc: number of U.S. Census Bureau industry or occupation codes (U.S. Census Bureau 2020) to be used to */ /* define the exposure (i.e., an individual industry or occupation code or a group of industry or occupation */ /* codes); minimum value is one; must be provided by the user */ /* */ /* indocc: U.S. Census Bureau industry or occupation code(s) to be used to define the exposure (i.e., an individual */ /* industry or occupation code or a group of industry or occupation codes); include a space between industry or */ /* occupation codes if more than one industry or occupation code is desired (e.g., 8770 8780); only industry OR */ /* only occupation codes can be used for a single call of the macro (i.e., industry AND occupation codes cannot */ /* be provided in the same call to the macro); must provide at least one industry or occupation code; must be */ /* provided by the user */ /* */ /* indoccvar: variable that contains the U.S. Census Bureau industry or occupation codes; must be provided by the */ /* user */ /* */ /* covariates: categorical covariate(s) (i.e., potential confounder[s]) by which to indirectly standardize PMRs and */ /* 95% CIs (e.g., age, sex, race, ethniity, education, etc.); include a space between covariates if more than */ /* one covariate is desired (e.g., AGE_RECODE_9 Sex); leave blank if an unadjusted PMR and 95% CI is desired */ /* */ /* References */ /* */ /* Breslow NE, Day NE. Statistical Methods in Cancer Research. Volume II—The Design and Analysis of Cohort Studies. */ /* IARC Scientific Publications No. 82. Lyon, France: International Agency for Research on Cancer; 1987. */ /* https://publications.iarc.fr/Book-And-Report-Series/Iarc-Scientific-Publications/Statistical-Methods-In-Cancer- */ /* Research-Volume-II-The-Design-And-Analysis-Of-Cohort-Studies-1986. */ /* */ /* National Center for Health Statistics. Classification of Diseases, Functioning, and Disability Homepage. */ /* http://www.cdc.gov/nchs/icd/. Updated June 20, 2020. Accessed May 6, 2021. */ /* */ /* National Institute for Occupational Safety and Health. LTAS Manual. */ /* https://www.cdc.gov/niosh/ltas/pdf/LTAS-manual-2014.pdf. Updated February 21, 2019. Accessed May 6, 2021. */ /* */ /* U.S. Census Bureau. Industry and Occupation Code Lists & Crosswalks. */ /* https://www.census.gov/topics/employment/industry-occupation/guidance/code-lists.html. Updated August 31, 2020. */ /* Accessed May 6, 2021. */ /* */ /*************************************************************************************************************************/ %macro PMR(dataset,ifst,ucodvar,mcodvars,numICD,ICD,numindocc,indocc,indoccvar,covariates); /*Creates a macro that calculates unadjusted or indirectly standardized proportionate mortality ratios (PMR) and 95% confidence intervals (CI) for associations between an individual industry or occupation code or a group of industry or occupation codes and mortality from an individual cause of death of interest or a group of causes of death of interest.*/ data work.restricted; set &dataset; /**********************************************************************************************************************/ /* Subsets the dataset (e.g., restrict to females only, Hispanics only, ages 18-65 years only [i.e., the working */ /* population], etc.). */ /**********************************************************************************************************************/ %if &ifst ^= %then %do; &ifst %end; /**********************************************************************************************************************/ NewID = _n_; /*Creates an ID variable.*/ /**********************************************************************************************************************/ /* Determines the outcome/cause(s) of death of interest (e.g., an individual or a group of ICD-10 code[s]) using */ /* underlying cause of death only or underlying and multiple causes of death. */ /**********************************************************************************************************************/ count_outcome = 0; /*Initializes the count variable for the outcome/cause(s) of death of interest.*/ array mcod {*} &ucodvar &mcodvars; /*Underlying (and multiple) cause(s) of death.*/ %if &numICD = 1 %then %do; array ICDcodes {*} $ a1 (&ICD); /*One ICD-10 code (i.e., an individual cause of death of interest).*/ %end; %else %do; array ICDcodes {*} $ a1-a&numICD (&ICD); /*More than one ICD-10 code (i.e., a group of causes of death of interest).*/ %end; do a = 1 to dim(mcod); do b = 1 to dim(ICDcodes); if mcod{a} = ICDcodes{b} then count_outcome = count_outcome + 1; /*Increments the count variable for the outcome/cause(s) of death of interest when the ICD-10 code(s) is found in the underlying (or multiple) cause(s) of death.*/ else count_outcome = count_outcome; end; end; if count_outcome > 0 then outcome = 1; /*Cause(s) of death of interest.*/ else outcome = 0; /*Not the cause(s) of death of interest.*/ /**********************************************************************************************************************/ /**********************************************************************************************************************/ /* Determines the exposure (e.g., an individual or a group of industry or occupation codes[s]). */ /**********************************************************************************************************************/ count_exposure = 0; /*Initializes the count variable for the exposure (i.e., industry or occupation code[s]).*/ %if &numindocc = 1 %then %do; array indocccodes {*} b1 (&indocc); /*One industry or occupation code.*/ %end; %else %do; array indocccodes {*} b1-b&numindocc (&indocc); /*More than one industry or occupation code (i.e., a group of industry or occupation codes).*/ %end; do c = 1 to dim(indocccodes); if &indoccvar = indocccodes{c} then count_exposure = count_exposure + 1; /*Increments the count variable for the exposure (i.e., industry or occupation code[s]) when the industry or occupation code(s) is found in the variable that contains the industry or occupation codes.*/ else count_exposure = count_exposure; end; if count_exposure > 0 then exposure = 1; /*Specified industry or occupation code(s).*/ else exposure = 0; /*Not the specified industry or occupation code(s).*/ /**********************************************************************************************************************/ %if &numICD = 1 and &numindocc = 1 %then %do; /*Drops unwanted variables.*/ drop count_outcome a1 a b count_exposure b1 c; %end; %else %if &numICD = 1 and &numindocc ^= 1 %then %do; drop count_outcome a1 a b count_exposure b1-b&numindocc c; %end; %else %if &numICD ^= 1 and &numindocc = 1 %then %do; drop count_outcome a1-a&numICD a b count_exposure b1 c; %end; %else %do; drop count_outcome a1-a&numICD a b count_exposure b1-b&numindocc c; %end; run; /*************************************************************************************************************************/ /* Sorts the dataset by the covariates (i.e., potential confounders; e.g., age, sex, race, ethnicity, education, etc.) */ /* that will be used to indirectly standardize the proportionate mortality among the exposed to the total population of */ /* all deaths (i.e., the reference population). */ /*************************************************************************************************************************/ %if &covariates ^= %then %do; proc sort data=work.restricted; by &covariates; run; %end; /*************************************************************************************************************************/ /*************************************************************************************************************************/ /* Calculates (i.e., counts) the number of deaths from all causes within strata of the covariates used for indirect */ /* standardization and outputs these numbers to a new dataset. */ /*************************************************************************************************************************/ proc means data=work.restricted noprint; var NewID; by &covariates; output out=work.dac n(NewID) = deaths_all_causes; run; /*************************************************************************************************************************/ /*************************************************************************************************************************/ /* Calculates (i.e., counts) the number of deaths from the cause(s) of interest within strata of the covariates used */ /* used for indirect standardization and outputs these numbers to a new dataset. */ /*************************************************************************************************************************/ proc means data=work.restricted noprint; var NewID; by &covariates; where outcome = 1; /*Cause(s) of death of interest (i.e., case).*/ output out=work.doi n(NewID) = deaths_of_interest; run; /*************************************************************************************************************************/ /*************************************************************************************************************************/ /* Calculates (i.e., counts) the number of deaths from all causes among the exposed within strata of the covariates */ /* used for indirect standardization and outputs these numbers to a new dataset. */ /*************************************************************************************************************************/ proc means data=work.restricted noprint; var NewID; by &covariates; where exposure = 1; /*Exposed (specified industry or occupation code[s]).*/ output out=work.dace n(NewID) = deaths_all_causes_exposed; run; /*************************************************************************************************************************/ /*************************************************************************************************************************/ /* Calculates (i.e., counts) the number of deaths from the cause(s) of interest among the exposed within strata of the */ /* covariates used for indirect standardization and outputs these numbers to a new dataset. */ /*************************************************************************************************************************/ proc means data=work.restricted noprint; var NewID; by &covariates; where exposure = 1 and outcome = 1; /*Exposed (specified industry or occupation code[s]) and cause(s) of death of interest (case).*/ output out=work.doie n(NewID) = deaths_of_interest_exposed; run; /*************************************************************************************************************************/ /*************************************************************************************************************************/ /* Merges the four datasets containing the four quantities calculated above, drops unwanted variables, and calculates */ /* the proportion of deaths from the cause(s) of interest within strata of the covariates used for indirect */ /* standardization. Also calculates the expected number of deaths from the cause(s) of interest among the exposed */ /* within strata of the covariates used for indirect standardization. */ /*************************************************************************************************************************/ data work.expected; merge work.dac (drop = _TYPE_ _FREQ_) work.doi (drop = _TYPE_ _FREQ_) work.dace (drop = _TYPE_ _FREQ_) work.doie (drop = _TYPE_ _FREQ_); by &covariates; proportion_deaths_of_interest = deaths_of_interest/deaths_all_causes; deaths_of_interest_exposed_exp = proportion_deaths_of_interest*deaths_all_causes_exposed; run; /*************************************************************************************************************************/ /*************************************************************************************************************************/ /* Calculates the overall observed number of deaths from the cause(s) of interest among the exposed by summing the */ /* stratum-specific observed number of deaths from the cause(s) of interest among the exposed and outputs it to a new */ /* dataset. */ /*************************************************************************************************************************/ proc means data=work.expected noprint; var deaths_of_interest_exposed; output out=work.odoie sum(deaths_of_interest_exposed) = observed_deaths_interest_exposed; run; /*************************************************************************************************************************/ /*************************************************************************************************************************/ /* Calculates the overall expected number of deaths from the cause(s) of interest among the exposed by summing the */ /* stratum-specific expected number of deaths from the cause(s) of interest among the exposed and outputs it to a new */ /* dataset. */ /*************************************************************************************************************************/ proc means data=work.expected noprint; var deaths_of_interest_exposed_exp; output out=work.edoie sum(deaths_of_interest_exposed_exp) = expected_deaths_interest_exposed; run; /*************************************************************************************************************************/ /*************************************************************************************************************************/ /* Merges the two datasets containing the overall observed and expected number of deaths from the cause(s) of interest */ /* among the exposed, drops unwanted variables, and calculates the unadjusted or indirectly standardized proportionate */ /* mortality ratio and its 95% confidence interval and multiplies them by 100. */ /*************************************************************************************************************************/ data work.PMR; merge work.odoie (drop = _TYPE_ _FREQ_) work.edoie (drop = _TYPE_ _FREQ_); Z = probit(0.975); /*Standard normal distribution critical value (i.e., approximately 1.96) needed to calculate a two-sided 95% confidence interval for the unadjusted or indirectly standardized proportionate mortality ratio.*/ PMR = (observed_deaths_interest_exposed/expected_deaths_interest_exposed)*100; /*Calculates a two-sided 95% confidence interval for the unadjusted or indirectly standardized proportionate mortality ratio using formulas based on Byar's approximation to the exact Poisson test (Breslow and Day 1987, p. 69; National Institute for Occupational Safety and Health 2019, p. 66).*/ LCL = ((observed_deaths_interest_exposed*((1 - (1/(9*observed_deaths_interest_exposed)) - (Z/(3*sqrt(observed_deaths_interest_exposed))))**3))/expected_deaths_interest_exposed)*100; UCL = (((observed_deaths_interest_exposed + 1)*((1 - (1/(9*(observed_deaths_interest_exposed + 1))) + (Z/(3*sqrt((observed_deaths_interest_exposed + 1)))))**3))/expected_deaths_interest_exposed)*100; drop Z; run; /*************************************************************************************************************************/ proc print data=work.PMR noobs; /*To see the unadjusted or indirectly standardized proportionate mortality ratio and its 95% confidence intervals.*/ run; %mend; /*Unadjusted PMR and 95% CI for the association between the automotive repair and maintenance industry (i.e., U.S. Census Bureau 2012 industry code 8770) and unspecified dementia (i.e., ICD-10 code F03) using the underlying cause of death only. dataset: public.public2019 ifst: ucodvar: UNDERLYING_CAUSE_OF_DEATH mcodvars: numICD: 1 ICD: "F03" numindocc: 1 indocc: 8770 indoccvar: Census2012IndustryCode covariates: */ %PMR(public.public2019,,UNDERLYING_CAUSE_OF_DEATH,,1,"F03",1,8770,Census2012IndustryCode,) /*Indirectly standardized (by age, sex, race, ethnicity, and education) PMR and 95% CI for the association between the automotive repair and maintenance industry (i.e., U.S. Census Bureau 2012 industry code 8770) and unspecified dementia (i.e., ICD-10 code F03) using the underlying cause of death only. dataset: public.public2019 ifst: if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .; ucodvar: UNDERLYING_CAUSE_OF_DEATH mcodvars: numICD: 1 ICD: "F03" numindocc: 1 indocc: 8770 indoccvar: Census2012IndustryCode covariates: Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4 */ %PMR(public.public2019,if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .;,UNDERLYING_CAUSE_OF_DEATH,,1,"F03",1,8770,Census2012IndustryCode,Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4) /*Restricted to females and indirectly standardized (by age, race, ethnicity, and education) PMR and 95% CI for the association between the automotive repair and maintenance industry (i.e., U.S. Census Bureau 2012 industry code 8770) and unspecified dementia (i.e., ICD-10 code F03) using the underlying cause of death only. dataset: public.public2019 ifst: if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= . and Sex = 1; ucodvar: UNDERLYING_CAUSE_OF_DEATH mcodvars: numICD: 1 ICD: "F03" numindocc: 1 indocc: 8770 indoccvar: Census2012IndustryCode covariates: Age_Recode_19 Race_Recode_5 Hispanic_Origin_2 Education_2003_4 */ %PMR(public.public2019,if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= . and Sex = 1;,UNDERLYING_CAUSE_OF_DEATH,,1,"F03",1,8770,Census2012IndustryCode,Age_Recode_19 Race_Recode_5 Hispanic_Origin_2 Education_2003_4) /*Indirectly standardized (by age, sex, race, ethnicity, and education) PMR and 95% CI for the association between the automotive repair and maintenance industry (i.e., U.S. Census Bureau 2012 industry code 8770) and unspecified dementia (i.e., ICD-10 code F03) using underlying and multiple causes of death. dataset: public.public2019 ifst: if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .; ucodvar: UNDERLYING_CAUSE_OF_DEATH mcodvars: Condition_1RA Condition_2RA Condition_3RA Condition_4RA Condition_5RA Condition_6RA Condition_7RA Condition_8RA Condition_9RA Condition_10RA Condition_11RA Condition_12RA Condition_13RA Condition_14RA Condition_15RA Condition_16RA Condition_17RA Condition_18RA Condition_19RA Condition_20RA numICD: 1 ICD: "F03" numindocc: 1 indocc: 8770 indoccvar: Census2012IndustryCode covariates: Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4 */ %PMR(public.public2019,if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .;,UNDERLYING_CAUSE_OF_DEATH,Condition_1RA Condition_2RA Condition_3RA Condition_4RA Condition_5RA Condition_6RA Condition_7RA Condition_8RA Condition_9RA Condition_10RA Condition_11RA Condition_12RA Condition_13RA Condition_14RA Condition_15RA Condition_16RA Condition_17RA Condition_18RA Condition_19RA Condition_20RA,1,"F03",1,8770,Census2012IndustryCode,Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4) /*Indirectly standardized (by age, sex, race, ethnicity, and education) PMR and 95% CI for the association between the automotive repair and maintenance industry (i.e., U.S. Census Bureau 2012 industry code 8770) and demyelinating diseases of the central nervous system (i.e., ICD-10 codes G35, G36, G360, G361, G368, G369, G37, G370, G371, G372, G373, G374, G375, G378, G379) using the underlying cause of death only. dataset: public.public2019 ifst: if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .; ucodvar: UNDERLYING_CAUSE_OF_DEATH mcodvars: numICD: 15 ICD: "G35" "G36" "G360" "G361" "G368" "G369" "G37" "G370" "G371" "G372" "G373" "G374" "G375" "G378" "G379" numindocc: 1 indocc: 8770 indoccvar: Census2012IndustryCode covariates: Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4 */ %PMR(public.public2019,if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .;,UNDERLYING_CAUSE_OF_DEATH,,15,"G35" "G36" "G360" "G361" "G368" "G369" "G37" "G370" "G371" "G372" "G373" "G374" "G375" "G378" "G379",1,8770,Census2012IndustryCode,Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4) /*Indirectly standardized (by age, sex, race, ethnicity, and education) PMR and 95% CI for the association between the other services, except public administration, industries (i.e., U.S. Census Bureau 2012 industry codes 8770, 8780, 8790, 8870, 8880, 8890, 8970, 8980, 8990, 9070, 9080, 9090, 9160, 9170, 9180, 9190, 9290) and unspecified dementia (i.e., ICD-10 code F03) using underlying cause of death only. dataset: public.public2019 ifst: if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .; ucodvar: UNDERLYING_CAUSE_OF_DEATH mcodvars: numICD: 1 ICD: "F03" numindocc: 17 indocc: 8770 8780 8790 8870 8880 8890 8970 8980 8990 9070 9080 9090 9160 9170 9180 9190 9290 indoccvar: Census2012IndustryCode covariates: Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4 */ %PMR(public.public2019,if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .;,UNDERLYING_CAUSE_OF_DEATH,,1,"F03",17,8770 8780 8790 8870 8880 8890 8970 8980 8990 9070 9080 9090 9160 9170 9180 9190 9290,Census2012IndustryCode,Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4) /*Indirectly standardized (by age, sex, race, ethnicity, and education) PMR and 95% CI for the association between the crane and tower operators occupation (i.e., U.S. Census Bureau 2010 occupation code 9510) and unspecified dementia (i.e., ICD-10 code F03) using underlying cause of death only. dataset: public.public2019 ifst: if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .; ucodvar: UNDERLYING_CAUSE_OF_DEATH mcodvars: numICD: 1 ICD: "F03" numindocc: 1 indocc: 9510 indoccvar: Census2010OccupationCode covariates: Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4 */ %PMR(public.public2019,if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .;,UNDERLYING_CAUSE_OF_DEATH,,1,"F03",1,9510,Census2010OccupationCode,Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4) /*Indirectly standardized (by age, sex, race, ethnicity, and education) PMR and 95% CI for the association between the material moving occupations (i.e., U.S. Census Bureau 2010 occupation codes 9510, 9520, 9560, 9600, 9610, 9620, 9630, 9640, 9650, 9720, 9730, 9740, 9750) and unspecified dementia (i.e., ICD-10 code F03) using underlying cause of death only. dataset: public.public2019 ifst: if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .; ucodvar: UNDERLYING_CAUSE_OF_DEATH mcodvars: numICD: 1 ICD: "F03" numindocc: 13 indocc: 9510 9520 9560 9600 9610 9620 9630 9640 9650 9720 9730 9740 9750 indoccvar: Census2010OccupationCode covariates: Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4 */ %PMR(public.public2019,if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= .;,UNDERLYING_CAUSE_OF_DEATH,,1,"F03",13,9510 9520 9560 9600 9610 9620 9630 9640 9650 9720 9730 9740 9750,Census2010OccupationCode,Age_Recode_19 Sex Race_Recode_5 Hispanic_Origin_2 Education_2003_4) /*Restricted to females and indirectly standardized (by age, race, ethnicity, and education) PMR and 95% CI for the association between the material moving occupations (i.e., U.S. Census Bureau 2010 occupation codes 9510, 9520, 9560, 9600, 9610, 9620, 9630, 9640, 9650, 9720, 9730, 9740, 9750) and demyelinating diseases of the central nervous system (i.e., ICD-10 codes G35, G36, G360, G361, G368, G369, G37, G370, G371, G372, G373, G374, G375, G378, G379) underlying and multiple causes of death. dataset: public.public2019 ifst: if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= . and Sex = 1; ucodvar: UNDERLYING_CAUSE_OF_DEATH mcodvars: Condition_1RA Condition_2RA Condition_3RA Condition_4RA Condition_5RA Condition_6RA Condition_7RA Condition_8RA Condition_9RA Condition_10RA Condition_11RA Condition_12RA Condition_13RA Condition_14RA Condition_15RA Condition_16RA Condition_17RA Condition_18RA Condition_19RA Condition_20RA numICD: 15 ICD: "G35" "G36" "G360" "G361" "G368" "G369" "G37" "G370" "G371" "G372" "G373" "G374" "G375" "G378" "G379" numindocc: 13 indocc: 9510 9520 9560 9600 9610 9620 9630 9640 9650 9720 9730 9740 9750 indoccvar: Census2010OccupationCode covariates: Age_Recode_19 Race_Recode_5 Hispanic_Origin_2 Education_2003_4 */ %PMR(public.public2019,if Age_Recode_19 ^= . and Hispanic_Origin_2 ^= . and Education_2003_4 ^= . and Sex = 1;,UNDERLYING_CAUSE_OF_DEATH,Condition_1RA Condition_2RA Condition_3RA Condition_4RA Condition_5RA Condition_6RA Condition_7RA Condition_8RA Condition_9RA Condition_10RA Condition_11RA Condition_12RA Condition_13RA Condition_14RA Condition_15RA Condition_16RA Condition_17RA Condition_18RA Condition_19RA Condition_20RA,15,"G35" "G36" "G360" "G361" "G368" "G369" "G37" "G370" "G371" "G372" "G373" "G374" "G375" "G378" "G379",13,9510 9520 9560 9600 9610 9620 9630 9640 9650 9720 9730 9740 9750,Census2010OccupationCode,Age_Recode_19 Race_Recode_5 Hispanic_Origin_2 Education_2003_4)