3.4 Baseline file

The Baseline File is optional. It should be specified if a baseline table is requested for any groups. The file enables users to customize the baseline table in the report. This file should contain 1 row per group. Table 3.6 contains detailed specifications for this file.

Table 3.6: BASELINEFILE specification
Parameter Field Name Description
GROUP/ANALYSISGRP/MILGRP to output GROUP Name of each cohort, analysis or mother-infant linkage group name to include in report.
Note 1: GROUP must correspond to values of GROUP/ANALYSISGRP/MILGRP included in file distributed input files. For Type 6 Manufacturer-level Product Utilization and Switching Patterns analyses, only ANALYSISGRP values can be specified.
Note 2: GROUP cannot be left blank.
Note 3: if MILGRP is specified, both exposure and reference cohorts are included in table.
Note 4: unadjusted baseline tables will be de-duplicated by PSESTIMATEGRP (i.e., if the same PSESTIMATEGRP is specified for multiple ANALYSISGRP values in the distributed PS Estimation File, only 1 unadjusted baseline table will be output to avoid duplicate tables in the report).
Note 5: unweighted IPTW baseline tables will be de-duplicated by PSESTIMATEGRP (i.e., if the same PSESTIMATEGRP is specified for multiple ANALYSISGRP values in the distributed PS Estimation File, only 1 unweighted IPTW baseline table will be output to avoid duplicate tables in the report).
Note 6: if baseline tables are de-duplicated (as in Note 4 and Note 5), the BASELINEFILE parameter values for the first unadjusted or unweighted IPTW table specfied in ORDER will be applied to the de-duplicated baseline table.
Format: SAS character $40; no special characters (e.g., commas, periods, hyphens, spaces, etc.) allowed, and underscores must be used to mark spaces.
Example: insulin
Run identifier RUNID RunID assigned to the run that corresponds to the group name listed in GROUP.
Note 1: RUNID must be equivalent to value specified in the distributed component of the request.
Format: SAS character $5
Example: r01
Group display order ORDER Order to display groups in report. If baseline tables for any groups are requested to appear on the same tab, they should be specified with the same ORDER value and BASELINEGROUPNUM must be specified.
Format: Numeric
Example: 1
Order of Baseline Tables to Appear on Same Tab BASELINEGROUPNUM Specifies order of baseline tables to appear on the same tab in report. BASELINEGROUPNUM order is applied to all GROUPs with the same value of ORDER. For example, the specification of GROUP, ORDER, and BASELINEGROUPNUM below would result in the GROUP A baseline table appearing on the Table 1a tab, the GROUP B and GROUP C baseline tables appearing on the Table 1b tab, and the Group D baseline table appearing on Table 1c tab. Only 2 groups can appear on the same tab. 
GROUP  ORDER  BASELINEGROUPNUM
A       1      .
B       2      1
C       2      2
D       3      .
Valid values are:
  • 1: group appears first on the tab
  • 2: group appears second on the tab
Note 1: for any group with a baseline table appearing on its own tab (i.e. one baseline table per tab), leave this field blank.
Note 2: BASELINEGROUPNUM values must be specified consecutively.
Note 3: if requesting two tables on the same tab, all BASELINEFILE parameter values applied to the first group are applied to the second group.
Note 4: BASELINEGROUPNUM functionality is restricted to groups within the same run ID.
Note 5: BASELINEGROUPNUM cannot be used with MILGRPs; when REPORTTYPE = T6L1, T2L2, or T4L2; or when INCLUDENONPREGNANT = Y.
Format: Numeric
Example: 1
Characteristics to display under “Health Characteristics” heading HEALTHCHAR Used to limit characteristics included in baseline tables and to specify which characteristics to include in a “Health Characteristics” section of the tables. Include a list of characteristics to include in this section, separated by a space.
Valid values are:
  • COVAR#: # corresponds to COVARNUM used for each covariate included in the distributed covariate input file.
  • COMORBIDSCORE: Charlson/Elixhauser Combined Comorbidity Score
  • NUMAV: Ambulatory encounters
  • NUMED: Emergency department encounters
  • NUMOA: Other ambulatory encounters
  • NUMIP: Inpatient hospital encounters
  • NUMIS: Non-acute institutional encounters
  • NUMCLASS: Unique drug classes dispensed
  • NUMGENERIC: Generics dispensed
  • NUMRX: Filled prescriptions
Note 1: if blank, a “Health Characteristics” section will not be included in baseline tables.
Note 2: can use “dash” notation to include range of COVAR# values in numeric order (i.e., COVAR1-COVAR9).
Format: SAS character - length can vary
Example: COVAR1-COVAR5 COVAR7
Characteristics to display under “Medical Product Use” heading MEDPRODUSE Used to limit characteristics included in baseline tables and to specify which characteristics to include in a “Medical Product Use” section of the tables. Include a list of characteristics to include in this section, separated by a space.
Valid values are:
  • COVAR#: # corresponds to COVARNUM used for each covariate included in the distributed covariate input file.
  • COMORBIDSCORE: Charlson/Elixhauser Combined Comorbidity Score
  • NUMAV: Ambulatory encounters
  • NUMED: Emergency department encounters
  • NUMOA: Other ambulatory encounters
  • NUMIP: Inpatient hospital encounters
  • NUMIS: Non-acute institutional encounters
  • NUMCLASS: Unique drug classes dispensed
  • NUMGENERIC: Generics dispensed
  • NUMRX: Filled prescriptions
Note 1: if blank, a “Medical Product Use” section will not be included in baseline tables.
Note 2: can use “dash” notation to include range of COVAR# values in numeric order (i.e., COVAR1-COVAR9).
Format: SAS character - length can vary
Example: COVAR1-COVAR5 COVAR7
Characteristics to display under “Laboratory Characteristics” heading LABCHARACTERISTICS Used to limit characteristics included in baseline tables and to specify which characteristics to include in a “Laboratory Characteristics” section of the tables. Include a list of characteristics to include in this section, separated by a space.
Valid values are:
  • COVAR#: # corresponds to COVARNUM used for each covariate included in the distributed covariate input file.
Note 1: if blank, a “Laboratory Characteristics” section will not be included in baseline tables.
Note 2: can use “dash” notation to include range of COVAR# values in numeric order (i.e., COVAR1-COVAR9).
Note 3: laboratory-based characteristics must have CODECAT = LB. Each characteristic must include either only numeric or only character labs.
Format: SAS character - length can vary
Example: COVAR1-COVAR5 COVAR7
Characteristics to display under “Health Service Utilization Intensity Metrics” heading UTILIZATIONINTENSITY Used to limit characteristics included in baseline tables and to specify which characteristics to include in a “Health Service Utilization Intensity Metrics” section of the tables. Include a list of characteristics to include in this section, separated by a space.
Valid values are:
  • COVAR#: # corresponds to COVARNUM used for each covariate included in the distributed covariate input file.
  • COMORBIDSCORE: Charlson/Elixhauser Combined Comorbidity Score
  • NUMAV: Ambulatory encounters
  • NUMED: Emergency department encounters
  • NUMOA: Other ambulatory encounters
  • NUMIP: Inpatient hospital encounters
  • NUMIS: Non-acute institutional encounters
  • NUMCLASS: Unique drug classes dispensed
  • NUMGENERIC: Generics dispensed
  • NUMRX: Filled prescriptions
Note 1: if blank, a “Health Service Utilization Intensity Metrics” section will not be included in baseline tables.
Note 2: can use “dash” notation to include range of COVAR# values in numeric order (i.e., COVAR1-COVAR9).
Format: SAS character - length can vary
Example: NUMAV NUMED NUMOA NUMIS NUMCLASS NUMGENERIC
Characteristics to display under “Pregnancy Characteristics” heading PREGNANCYCHAR Used to specify which pregnancy-related characteristics to include in a “Pregnancy Characteristics” section of the baseline table. Include a list of characteristics to include in this section, separated by a space.
Valid values are:
  • PREPOSTIND_PRE: Preterm
  • PREPOSTIND_TERM: Term
  • PREPOSTIND_POST: Postterm
  • PREPOSTIND_NONE: Unknown term
  • GA_BIRTH: Estimated gestational age at delivery
Note 1: only relevant for Type 4 analyses.
Note 2: if blank, a “Pregnancy Characteristics” section will not be included in baseline tables.
Format: SAS character - length can vary
Example: PREPOSTIND_PRE PREPOSTIND_TERM PREPOSTIND_POST PREPOSTIND_NONE GA_BIRTH
Characteristics to display under “Exposure Characteristics” heading EXPOSURECHAR Used to specify which pregnancy-related exposure characteristics to include in an “Exposure Characteristics” section of the baseline table. Include a list of characteristics to include in this section, separated by a space.
Valid values are:
  • GA_FIRST: Gestational age of first exposure, weeks
  • DISP_PRE: Number of dispensings in pre-pregnancy period
  • ADJUSTEDDISP_T1: Number of dispensings in first trimester
  • ADJUSTEDDISP_T2: Number of dispensings in second trimester
  • ADJUSTEDDISP_T3: Number of dispensings in third trimester
  • EXP_T1: Exposed during first trimester
  • EXP_T2: Exposed during second trimester
  • EXP_T3: Exposed during third trimester
  • EXP_PREGPRE: Exposed during user-defined pre-pregnancy period
Note 1: only relevant for Type 4 analyses.
Note 2: if blank, an “Exposure Characteristics” section will not be included in baseline tables.
Format: SAS character - length can vary
Example: GA_FIRST DISP_PRE ADJUSTEDDISP_T1 ADJUSTEDDISP_T2 ADJUSTEDDISP_T3 EXP_T1 EXP_T2 EXP_T3 EXP_PREGPRE
Infant characteristics indicator OUTPUTINFANTCHAR Indicator for inclusion of an infant section of the baseline table.
Valid values are:
  • Y: Include an infant section in the baseline table
  • N: Do not include an infant section in the baseline table
Note 1: only relevant for Type 4 analyses.
Format: SAS character $1
Example: Y
Indicator for inclusion of baseline table for non-pregnant cohort in L1 type 4 analyses INCLUDENONPREGNANT Indicator for inclusion of a non-pregnant cohort to the baseline table.
Valid values are:
  • Y: Include a non-pregnant cohort in the baseline table
  • N: Do not include a non-pregnant cohort in the baseline table
Note 1: only relevant for Type 4 analyses.
Format: SAS character $1
Example: Y
Covariate sort order COVARSORT Character indicator specifying the sort order for covariates in output.
Valid values are:
  • A: Sort by alphabetical order of covariate label
  • C: Sort by numeric assignment for covariate per contents of file represented by parameter COVARNUM
  • O: Sort by order that the covariates are specified in the parameters HEALTHCHAR, MEDPRODUSE, UTILIZATIONINTENSITY
Format: SAS character $1
Example: A
Compute absolute and standardized difference indicator COMPUTEBALANCE Compute absolute and standardized difference between cohorts.
Valid values are:
  • Y: Compute absolute and standardized difference
  • N: Do not compute absolute and standardized difference
Note 1: if COMPUTEBALANCE is requested for a Type 4 analysis, INCLUDENONPREGNANT must = Y or GROUP must include a MILGRP value.
Note 2: for L2 tables, absolute and standardized difference will always be output.
Note 3: COMPUTEBALANCE can be used with BASELINGROUPNUM
Format: SAS character $1
Example: Y
Standardized difference threshold SDTHRESHOLD Threshold for Table 1 standardized difference. Any values above the amount represented by this parameter will be output with font color blue.
Note 1: SDTHRESHOLD value cannot differ for ANALYSISGRPs with the same PSESTIMATEGRP.
Format: SAS Numeric
Example: 0.1
Covariates not in PS model COVNOTINPS List of covariates not included in the propensity score logistic regression model. Covariates listed will be indicated by a footnote in the Table 1 output for inferential analyses. COVNOTINPS can take the value of any METVAR value within the Adjusted Baseline dataset.
Note 1: COVNOTINPS value cannot differ for ANALYSISGRPs with the same PSESTIMATEGRP.
Note 2: Only relevant if REPORTTYPE = T2L2 or T4L2.
Format: SAS character - length can vary
Example: comorbidscore age age18-39 age60 race_0 hispanic_y year_2018 sex_f covar1 covar25 numip numrx
Profile Table output PROFILECOVARSTOINCLUDE List of covariates to include in a covariate profile table. Produces a table profiling the number of patients and episodes that qualify for every possible combination of covariates requested within one run of the QRP package.
Valid values are:
  • ALL: Include all covariates
  • COVAR#: # corresponds to COVARNUM used for each covariate included in the distributed covariate input file
Note 1: only relevant if profile datasets are requested.
Note 2: if blank, a profile table will not be output.
Note 3: can use “dash” notation to include range of COVAR# values in numeric order (i.e., COVAR1-COVAR9).
Format: SAS character $100
Example: COVAR1 COVAR2 COVAR8-COVAR20