 |
FHS Methodology:
FHS Data:
FHS Bibliographies:
Quality Control (QC)/ Quality Assurance (QA):
The Framingham Heart Study Echocardiography Laboratory adheres to many quality assurance techniques including those listed below.
- External validity: In developing our protocol we reviewed the manuals and spoke with investigators from other epidemiological investigations (i.e. CHS and ARIC).
- Written Procedure Manual: We had extensive written procedure manual that included the following elements: Scanning protocol; Computer protocol, and interpretation sequence; In-depth interpretation guidelines; Posted guidelines for what studies require MD over reading & what inconsistencies will trigger data cleaning; Data cleaning protocol; Reproducibility manual.
- Abnormal studies: All abnormal studies are over-read by a physician.
- Regular staff meetings: Throughout the examination cycle we had regular meetings to review reproducibility studies, discuss interpretation issues, and over-read abnormal studies together (to minimize interpreter drift).
- Data Entry: One individual performed the data entry and returned incomplete or ambiguous forms to interpreters. All studies were subsequently double keyed by an additional staff member.
- Data Cleaning: Data cleaning steps consisted of checking the accuracy of key identifying data, and examining the interpretations for logic and consistency. Data cleaning occurred on an ongoing basis, to give timely feedback to readers.
- Reproducibility: We assessed of intra- and inter-observer reproducibility of echocardiographic measurements and interpretations 1-2 times per year, 20 studies each time; We required that personnel meet acceptable reproducibility standards before they were certified to read independently. We assessed secular drift of intra- and inter-observer reproducibility of echocardiographic measurements and interpretation, by reading a calibration set annually. We analyzed sonographer variability by having each sonographer scan 20 subjects twice.
-
Descriptive statistics: On routine studies we examined sonographer and interpreter means for quantitative phenotypes by 6-month increments to look for systematic differences between observers, and longitudinal variability within observer over time.
Back to Project 5
Participants
|
