NHLBI Workshop

Data Needs for Cardiovascular Events, Management, and Outcomes

Disease Surveillance Questions - In designing surveillance systems suitable for global, national, or regional application:

What information is essential or indispensable?
What are the most valid, cost-efficient, and practical means for obtaining this information?
How can relevant data be obtained for potentially generalizable populations?
How can the system of surveillance become sustained?
How can this information continue to be made useful and of interest for different audiences?

Population-Based Approach to Surveillance

Broad-based perspective enhances generalizability and interpretation of findings
Ability to calculate incidence rates of disease and other pertinent health outcomes
Reflects more "real world" patients with disease, and their likely management practices, as compared to individuals studied in RCT's of more select patient samples with potentially narrow inclusion criteria

Non-Population Based Approach to Surveillance

Hospitals or clinics included for study may not be representative of centers from defined area
Patients hospitalized at select medical centers may have different characteristics from those seen in usual care settings
Management practices may not reflect those utilized in a community setting
Incidence rates of disease cannot be calculated with a "catch-can" ascertainment approach

"Cold" Pursuit Surveillance

Advantages
- Complete case ascertainment
- Cost efficiencies and minimal logistical complexities
Disadvantages
- Cannot obtain supplemental data not available from medical or other records for research purposes
- Need systematic lists and sampling frames for case selection

"Warm" Pursuit Surveillance

Advantages
- Supplemental data not included in medical records can be obtained through direct patient or surrogate interviews
- Can identify potential etiologic or prognostic factors in a more systematic and standardized manner
Disadvantages
- High potential for incomplete case ascertainment
- Greater logistical difficulties involved in identifying patients and ascertaining information
- Increased personnel costs

Endpoints to be examined in CHD Surveillance Systems

Obstacles to accessing and reviewing hospital and ambulatory care records

Ability to construct an appropriate sampling frame for sample selection
Reliability of ICD codes for case ascertainment
Selection of 1º or 2º discharge dx's of CHD for purposes of case ascertainment
Jumping into the "long and winding que" for accessing medical records (and importance of developing personal relationships)
Missing records
Incomplete records
Quality of information and lack of standardized questioning and recording
Quiet space for data abstractors

Data Abstraction Issues

Case definitions to be utilized
- Standardized criteria
- No upper age cap
- Decision on transferred cases
Hospital data elements
Post discharge data elements
- Mortality
- Morbidity
- QOL
- Medication adherence

HIPAA, Consent, and Confidentiality Issues

Type of study to be conducted (mailed questionnaire, medical record review, phone survey)
Initial IRB approvals and dealing with medical care centers either without an IRB or who meet infrequently
Patient identifiers and matching criteria for follow-up information
Assuring patient confidentiality

Event Adjudication

Determination of which events that need continual review (e.g., UA, HF) and those which may not need further review and adjudication (e.g., receipt of CABG)
Need for standardized definitions
Need for experienced clinicians
Adequate "case" information to validate or rule out from further consideration
Maintaining group interest in review process and emphasizing importance of reviewers task