The Atherosclerosis Risk in Communities (ARIC) Study

a. Identify major clinical events for participants lost to follow-up The longitudinally linked annual CMS Medicare claims reflect ongoing care provided to all Medicare-enrolled beneficiaries, independent of place of residence. As such, these data are ideally suited to ascertainment of disease incidence. Longitudinal studies, such as MESA, actively use the linked CMS Medicare data to ascertain events among study participants lost to follow-up. The combined ARIC-CMS Medicare hospitalization file, created with help from the ARIC CORC, serves a similar purpose and when updated annually, can be the basis for ongoing event ascertainment in general and especially among those lost to follow-up. Potential bias resulting from the limitation of disease ascertainment to diagnostic codes obtained from claims, would be minimal for coronary heart disease, but may be substantial for conditions such as heart failure, requiring use of calibration factors derived from existing cohort data. My role with respect to this contract objective is to supervise the annual creation and distribution to investigators of a final analytical data file that would supplement hospitalization records obtained from abstraction of cohort participants’ medical records with hospitalization records from CMS Medicare claims (the MedPAR files). b. For selected cardiovascular outcomes to be mutually agreed upon with the Project Office, compare estimates identified obtained using CMS Medicare claims data with estimates obtained using participants-reported hospitalizations. In the current healthcare environment an increasing proportion of hospitalized care is delivered during observation stays. The ARIC Study has not yet established clear protocols for the identification of observation stays from ongoing medical record abstraction. The CMS Medicare claims, however, can aid in the identification of conditions which are frequently managed as observation stays, including heart failure, transient ischemic attacks, and atrial fibrillation, through specific codes used to identify claims billed as observation stays. The contract objective to compare estimates of disease incidence from claims data and compare it to self-report of hospitalizations is currently fulfilled with respect to heart failure through research conducted by Ricky Camplain, a UNC Epidemiology student, under supervision provided jointly by Dr. Lisa Wruck and myself. My role for this component of the contract objectives is supervise annual identification of observation stays from claims data for conditions agreed upon with the Project Office and to provide guidance to field centers in efforts aimed at identification of observation stays from hospitals’ medical records. Additional work will involve preparation of documentation comparing estimates of self-reported hospitalizations, for conditions selected by mutual agreement with the NHLBI Project Office, with those obtained from CMS Medicare data. c. Enable analyses to address questions related to resource utilization among ARIC participants. Promotion of CMS Medicare data use by ARIC investigators will have to run on parallel tracks of the engagement in operational activities that will make the data more accessible and through identification of salient research questions which can be addressed with those data. In my role as the ARIC CMS data coordinator I have engaged in the following operational activities of the Coordinating Center which will enhance integration of claims-based analyses in ARIC. 1. Provide statistical oversight for the use of claims data in ARIC a. Solicit statistical code from investigators using CMS Medicare data and validate its use b. Supervise the creation of derived variables which will enable ARIC investigators to use claims data with lower analytical investment c. Engage statistical support for approved ARIC manuscripts utilizing CMS Medicare data I would further suggest targeted prioritization of CMS Medicare claims-based analyses which would fully leverage the potential of both the ARIC cohort data and the linked claims data. Potential topics can include (but should not be limited to) the following: 1. Healthcare resource utilization in dementia and MCI 2. Economic analyses of cost of care in heart failure 3. Patterns of healthcare utilization among caregivers 4. Equitable access to necessary care in heart failure. Although contract specifications do not identify tasks related to the use of the CMS Medicare data for the ARIC communities, the 2001-2013 data for the ARIC study communities have just been processed by the CC and have been distributed to investigators. Those data present a rich opportunity to examine the burden of chronic cardiovascular conditions and processes of care for those conditions in diverse geographic settings, with defined population denominators, while taking advantage of information on care provided in the inpatient as well as the outpatient setting. In my role as the ARIC CMS data coordinator, I will continue to promote the use of the CMS Medicare for the ARIC communities by personally engaging in research based on those data. I will further identify opportunities for research based on ARIC community CMS Medicare data that would support the mission of ARIC to examine factors contributing to the variation in medical care for cardiovascular disease.