QRP
 
Loading...
Searching...
No Matches
ms_cidacov_codeextraction.sas File Reference

This macro extracts codes for covariate evaluation and risk scores computation. More...

Detailed Description

This macro extracts codes for covariate evaluation and risk scores computation.

  • Create extraction lookup files for extraction needed.
  • Create enrollment dataset to use for shaving (infants will be assigned birth_date as Enr_Start).
  • Add indicator to determine an MI codepop or I codepop.
    • Add codepop and identify if I or M patid.
    • Recode infant patid to mother patids when it is an infant claim.
  • First step (DX/PX extraction):
    • Break out CaresettingPrincipal into Caresetting and Principal.
    • Extracting meds using EncType and PDX to meds.
    • Extract values for data extraction censoring.
    • Separate Risk Scores PX/DX from covariates.
    • Exract RX codes and separate risk scores codes from covariates.
  • Remaining steps (RX/ENC extraction, stockpiling, and enrollment shaving only needed for covariates codes):
    • Extract values for data extraction censoring and efficiency.
    • PX/DX data shaving and create final file if 1 enrollment criteria.
    • Labs data extraction and shaving:
      • For numeric labs, restrict to modifier=EQ.
      • Verify if at least one lab code was extracted.
        • Determine which covariates including lab codes are eligible to have additional variables computed.
          • For additional variables to be computed, a lab covariate must not have other code categories and have codedays set to 1.
        • Select random same day labs and assign missing/invalid ms_result_c to INVALID.
          • Make sure numeric results are not counted as missing/invalid categorical result.
    • Encounter data extraction and shaving:
      • Break out CaresettingPrincipal into Caresetting and Principal.
      • Extract by encounter type.
      • Shave claims.
        • If >1 enrollment criteria then output 1 file per enrollment criteria.
        • Else output final file.
    • Dispensing data extraction, shaving, and stockpiling:
      • Set aside covariatecode parameters to merge in later.
      • Define rxsupenddt.
      • Extract dispensing claims in the combo tool.
      • Shave claims and stockpile.
        • Because of potential for different enrollment criteria by group, loop through each unique enrollment group.
          • Apply stockpiling:
            • If there are no stockpiling file or 1 set of stockpiling parameters, can stockpile for each enrollment group.
            • If >1 set of stockpiling parameters, stockpile will be done separately for each group.
          • Reshave after stockpiling
          • Create final dataset if 1 enrollment group and <=1 stockpiling parameters.
    • Loop through each group if stockpiling differs by group or if there is >1 enrollment group.
      • Determine correct enrollment span for group.
      • Restrict data extract to only patients in the current group and assign group value.
      • Create final encounter dataset.
      • Create final meds dataset.
      • Create final labs dataset.
      • Dispensing data:
        • Stockpile and re-shave non-combo dispensing codes for each group if >1 set of stockpiling parameters.
          • Restricting to pts in current group.
            • Only process dataset if there are subsetted observations.
        • Stack stockgroups together.
        • Create final dispensing dataset.
          • Combo RX codes.
            • Stack combo and non-combo RX codes.
          • Merge in PreDrugParams.
Program inputs
  • dplocal.&RUNID._&DATA. (Dataset containing one record per individual per index date for every cohort specified in the CIDA module run.)
  • infolder.&MICOHORTFILE. (Dataset for Type 4 analysis with pregnancy cohorts among pregnancies matched to an infant specification.)
  • &RISKSCORESCODES. (Dataset with risk scores code list.)
  • &RUNID._&DATA. (Dataset containing first index date.)
  • work.baselinegrps (Dataset containing list of cohorts.)
  • work.cohortfile_for_enr (Dataset where same basecohort are grouped and assess order to process them correctly.)
  • work.covarcodes (Dataset with covariate code list.)
  • work.riskscorefile (Dataset containing unique risk scores.)
  • work.stockpile_covar (Dataset containing variables to stockpile (where stockgroups are only in covariatecodes file).)
  • worktemp.&ENR_DATA_FOR_SHAVING._&ENR. (Dataset containing each unique enrollment group to shave.)
  • worktemp.&ENR_DATA_FOR_SHAVING._&ENRNUM. (Dataset containing stacked enrollment group to reshave.)
  • worktemp.dem_infant_enr_&ENR. (Dataset containing infant only enrollment.)
  • worktemp.drugs (Dataset containing RX to extract.)
  • worktemp.drugscomb (Dataset containing combination items that behave like RX.)
  • worktemp.enr_&ENR. (Dataset containing enrollment.)
  • worktemp.labextract (Dataset containing labs to extract.)
  • worktemp.meds (Dataset containing combination of _diags and _procs.)
Program outputs
  • work.covarlabs_extravars (Dataset containing covariates including lab codes that are eligible to have additional variables computed.)
  • work.predrugs (Dataset containing dispensing covariate codes.)
  • work.preenc (Dataset containing encounter covariate codes.)
  • work.prelabs (Dataset containing lab covariate codes.)
  • worktemp._riskscores_rx (Dataset containing drug claims for scores computation.)
  • worktemp.premeds (Dataset containing med covariate codes.)
  • worktemp.riskscoresmeds (Dataset with separated diagnosis and procedure claims for scores computation.)

Usage 

%ms_cidacov_codeextraction();
Parameters
None.

SAS Macros Dependencies

Author
Sentinel Coordinating Center (info@.nosp@m.sent.nosp@m.inels.nosp@m.yste.nosp@m.m.org)