Prevention & lipidsCohort study
Comparison of Lipid Results From 2007 to 2009 via Commercial Clinical Laboratory Testing Versus a Population-Based Cohort: Proof-of-Concept for Population Health Monitoring of Lipids Using Commercial Lab Data
- Comparison of commercial laboratory testing versus a population-based cohort
- proof-of-concept
- sensitivity analysis
- logistic regression adjusting for age and sex
Authors' conclusion
Within the same geographic region, the differences between lipids observed in commercial laboratory testing and in a population-based cohort were negligible (high-density lipoprotein cholesterol) to moderate (low-density lipoprotein cholesterol, triglycerides). This proof-of-concept study suggests that commercial laboratory testing may be useful for monitoring population lipid levels and assessing regional and temporal variation.
Study context
- Population
- Adult patients in Dallas County, TX, September 2007-September 2009commercial laboratory population 59,419 · DHS-2 3,147
- Participants
- 59,419 vs 3,147
- Exposure
- commercial laboratory population vs DHS-2
- commercial laboratory populationn = 59,419
- DHS-2n = 3,147
Main finding
As reported in the abstract
Distribution plots of high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and triglycerides commercial laboratory population median age [interquartile range] 48 [33-58] years · DHS-2 median age [interquartile range] 50 [42-58] years
women
Observational data: associations, not effects.
Further findings
| Outcome | Estimate | |
|---|---|---|
| Median lipid levels — commercial laboratory population vs DHS-2 · lower | difference in median high-density lipoprotein cholesterol 1 mg/dL (95% CI 0.6–2) | |
| Median lipid levels — commercial laboratory population vs DHS-2 · lower | difference in median low-density lipoprotein cholesterol 5 mg/dL (95% CI 3–6) | |
| Median lipid levels — commercial laboratory population vs DHS-2 · higher | difference in median triglycerides 12 mg/dL (95% CI 9–14) | |
| −1001020Difference, mg/dL (linear scale) |
Sensitivity analysis excluding those with diabetes or cardiovascular disease
- Similar results were observed
Abnormal levels of all 3 lipids by ZIP code
- Higher probabilities for 5 ZIP codes; odds ratios [95% CI] ranged from
- 1.07 [1.001-1.15] to 1.84 [1.48-2.28]
- commercial laboratory population; adjusted for age and sex
Original publication
Summarised from the published abstract; every number is checked against it