Randomization for INTECOMM trial

A dependency of the INTECOMM trial EDC.

The INTECOMM trial is a cluster randomized trial where the unit of randomization is the patient group.

At screening, data for individual potential participants are stored in the intecomm_screening.PatientLog model. Eligible individual potential participants (model PatientLog) are added to a patient group (model intecomm_group.PatientGroup).

The data flow is PatientLog -> SubjectScreening -> if eligible -> SubjectConsent

Ideally, for a patient group to be considered for randomization, the group must contain between 9-14 screened and consented members where a count of chronic conditions of those in the group meets an approximate ratio of 2 : 1; that is, 2(DM/HTN) : 1(HIV). The site coordinators may override these values.

Once a PatientGroup is ready to randomize, the site staff open the PatientGroup form and click "randomize".

In the background, the Randomizer class calls its method randomize_group. randomize_group picks the next available record from the randomization_list (''intecomm_rando.RandomizationList) and inserts a uniquegroup_identifiervalue. A records is available ifgroup_identifierhas not been set. Ordering is ascending bysid. The PatientGroup is given its newly allocatedgroup_identifier. The subjects in this group may now be followed longitudinally starting with visit 1000. Thegroup_identifier, for subjects in a PatientGroup, is updated on the PatientLog record as well. * TheRegisteredGroupmodel holds thesidtogroup_identifierrelationship * TheRandomizationListmodel holds thesidtoassignmenttogroup_identifierrelationship *PatientLog`` links group_identifier and subject_identifier

See also tables: • Intecomm_rando_registeredgroup • Intecomm_rando_randomizationlist • intecomm_screening_patientlog • intecomm_group_patientlog