This macro creates control groups. More...

Detailed Description

This macro creates control groups.

Only create non-pregnant group if requested.
- Create basic control population.
- Initialize minPregDur parameter.
- If there are no pregnant patients, create empty _controlgroup dataset end exit macro.
- If there are pregnant patients, execute macro:
  - Calculate minimum pregnancy duration value from pregnant cohort.
  - Apply age group and enrollment duration requirements.
  - Separate patients who had pregnancy outcomes from those who did not.
    - For efficiency purposes, patients who did not have pregnancy outcomes should not be processed by the ms_shaveinside macro.
  - Remove treatment episodes segments from the control population enrollment episodes.
  - Get demographic info back.
  - Reapply the criterion since removing treatment episodes may have created too short segments.
  - The shaveinside macro does not only shave the enrollment episode where the delivery occurs, but shaves also previous enrollment episodes.
    - For example, enrollment episode 1 (01AUG2004 - 31MAY2006) and episode 2 (01AUG2006 - 31JAN2007), delivery 19NOV2006.
    - ShaveInside will truncate episode 1 at 19FEB2006 even if the delivery occured in episode 2.
      - Create a copy to have control info in the case of exclusion only.
  - Compute complex HOI:
    - Extract highest CONDFROM value among exclusions (if any).
      - Apply if GROUP has incl-excl codes, otherwise do not change _controlpopulation before matching.
    - In pregnancy, enrdays is from delivery date, not episode start date and it is already applied.
    - Restrict eligibility to study period.
    - Initially truncate start eligibility enrollment start by max of lookback periods.
    - Create MaxLookBack macro variable.
    - If there is inclusion or exclusion use DenomEnr instead of CStart CEnd.
    - Inclusion required (POV3 Incl) for eligibility.
      - For each condition, shave periods for all subconditions.
      - Determine how many levels of subconditions needed and if condition is an inclusion.
      - Determine if there is subexclusion in condition in order to calculate new maxlookback.
        
        If there is subexclusion, determine new maxlookback for that specific condition before shaving.
      - If the subcond condfrom is greater in absolute value than the maxlookback value, then readjust denominator period.
      - Restrict eligibility to study period.
      - Initially truncate start eligibility enrollment start by max of lookback periods.
      - Determine if subcond is an inclusion or exclusion.
      - Merge time periods for each subcondition and shave for each.
        
        Only keep time periods that meet codedays criteria.
      - Only keep time periods that meet codedays criteria.
    - Output dataset for inclusion=1 and for exclusions=0 and then (a) shaveoutside (b) shaveinside denominator.
      - First shaveoutside denom file but need to merge inclusion periods together.
        
        _Denom&GROUP._stack_inc includes all time periods where the potential index dates will meet at least one of the cond inclusions.
      - Merge all periods of ineligibilities.
      - _Denom&Group._stack_excl includes all time periods where the potential index dates will meet at least one of the cond exclusions.
      - Remove all these periods from the ELIG.
      - If there was exclusion, other variables lost in shaveinside so merge back in.
        
        Match treatment with control.
    - Make sure the number of enrollment days remaining after each potential elig days is sufficient.
      - Adjust eligibility end date.
  - Get treatment minimum and maximum delivery age.
  - Create control datasets according to each age.
  - Loop on age.
    - Match treatment with control.
    - Keep one control per treatment.
  - Do not run the matching macro if _TreatmentGroupForMatch is empty and create an empty table instead.
  - Process zip codes.
  - Make sure patients used more than once in the matching process are considered distinct deliveries.

Program inputs

infolder.&ZIPFILE. (Lookup table required if a request requires stratification of results by geographic location.)
work._inclexcl (Dataset with a list of control patients and demo data.)
work._it&INCLUSIONCODES. (Dataset with inclusion/exclusion definition.)
work._ptsmasterlist (Dataset containing the patient master list.)
work.pregnancyoutcomes (Dataset containing defined group pregnancy episodes.)
worktemp.enr_&ENROLLMENTNUM. (Dataset containing enrolment episodes.)

Program outputs

work._controlgroup (Datasets with a list of control patients and demo data.)
work._controlpopulation_&COND. (Dataset containing new maxlookback for that specific subexclusion condition.)
work._eligincl_&SUBCOND. (Dataset containing eligible time periods for each subcondition.)
work._name (Dataset with variables lost in shaveinside to merge back in.)
work._overlap_spans_&SUBCOND. (Dataset containig only the time periods that meet codedays criteria.)
work._ptslist (Dataset containig patients who did not have a pregnancy outcomes.)
work._treatmentgroupformatch (Dataset containing treatment minimum and maximum delivery age.)
work._uneligexcl_&SUBCOND. (Dataset containing non-eligible time periods for each subcondition.)

Usage

%ms_createnopreggroup();

Parameters: None.

SAS Macros Dependencies

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

	QRP