Analyzing Blood Pressure Trends Using Partial Least Squares Regression
Author Information
Author(s): Tu Yu-Kang, George Davey Smith, Mark S. Gilthorpe
Primary Institution: University of Leeds
Hypothesis
Can partial least squares regression effectively separate the effects of age, period, and cohort on blood pressure trends?
Conclusion
The study found that both systolic and diastolic blood pressure increased with age, while those born later had lower blood pressure.
Supporting Evidence
- Both systolic and diastolic blood pressure increased with age.
- Students born later had lower blood pressure.
- PLS analysis provides a flexible modeling strategy for age-period-cohort analysis.
- Men and women showed similar trends in blood pressure over time.
- Participants with missing values in confounders were excluded from the analysis.
- Ethics approval was obtained for the study.
- Data were collected from health screenings conducted between 1948 and 1968.
- PLS regression allows for the analysis of collinear variables.
Takeaway
This study looked at how age, the year you were born, and the year you were examined affect blood pressure. It found that older people generally have higher blood pressure, but younger people tend to have lower blood pressure.
Methodology
The study used partial least squares regression to analyze blood pressure data from the Glasgow Alumni cohort, focusing on age, year of examination, and year of birth.
Potential Biases
Potential biases may arise from the historical nature of the cohort and the lack of consent collected at the time of data collection.
Limitations
The study excluded participants born before 1925 or older than 25 at examination, which may limit the generalizability of the findings.
Participant Demographics
The cohort consisted of 15,322 students, including 11,755 men and 3,567 women, who received medical screening between 1948 and 1968.
Statistical Information
P-Value
−0.17 mmHg/per year for men; −0.25 mmHg/per year for women
Confidence Interval
−0.19 to −0.15 for men; −0.28 to −0.22 for women
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website