Core concept - Utilize temporal relationships between events
CLG’s patented capability to define patient groups with sophisticated rules for temporal relationships (distance and direction in time) between clinical events allows users to apply a variety of inclusionary and exclusionary criteria to cohort definitions. CLG’s Event Canvas can associate events of any type that are available in the clinical data repository. Event Canvas cohort definitions also have logical operators and parenthetical sub-groupings. Patients who have clinical events that fit the definition of the group are included as cohort subjects. Each subject also has a singular index date, specific to that person's care, which is used in analysis the way a study enrollment date is used—it is the start point for observing outcome events.
Because a CLG cohort is essentially a collection of patient ID’s and index dates from groups sharing common characteristics, cohorts can also be uploaded to CLG from other sources. These include EMR's, registers, external reports, or patient work lists.
Some analysis requires non-uniqueness of patient identity. Encounters by department, for example, pose a throughput question requiring a collection of events rather than a cohort. CLG’s event collection capability addresses this question, defining a collection in the same way as a cohort except that multiple instances of a qualifying event(s) are defined for each patient. For example, a user could create an event collection of all CHF admissions in 2007 or all glucose readings for stroke patients in the ICU in the last week. CLG Analytic Methods such as List, Pivot, and Time to Outcome can then be applied to this collection.
Real-world application - User defined studies
A clinical researcher has been asked to probe the risk of blindness in diabetic patients by discovering all diabetic patients that didn’t have an eye exam in the last year. The researcher wants to focus on those diabetic patients with high cholesterol, suggesting higher severity of illness. At her discretion, she defines diabetics by any one of three events: a Clinic Visit with diagnosis code of DM, a Hemoglobin A1c result greater than 7, or a Problem List Issue of DM. The earliest of any of those will become the date of diagnosis of DM and will be the Index Date for each subject.
Additionally, the researcher decides that high cholesterol is indicated by a cohort member having an LDL cholesterol > 130 within 30 days of the diabetes diagnosis. Finally, those diabetics with high LDL who have NOT had an eye exam in a given year are included in the study. Normally, such an analysis would require a database programmer and be delayed by the wait-time for requests in the IT queue. CLG enables clinical managers, analysts, and researchers to implement these rules in a cohort definition and perform this analysis themselves without delay.
Business value - Respond to changes in metrics
Ever-changing rules for patient group definition can be easily created to keep pace with changes to generally accepted and nationally mandated metrics and empower internal strategies for quality improvement . Because database programmers are not needed, results can be quickly delivered and the cost for the requisite skills dramatically reduced.