Framework to construct and interpret latent class trajectory modelling
Objectives
Latent class trajectory modelling (LCTM) is a relatively new methodology in epidemiology to describe life-course exposures, which simplifies heterogeneous populations into homogeneous patterns or classes. However, for a given dataset, it is possible to derive scores of different models based on number of classes, model structure and trajectory property. Here, we rationalise a systematic framework to derive a ‘core’ favoured model.
Methods
We developed an eight-step framework: step 1: a scoping model; step 2: refining the number of classes; step 3: refining model structure (from fixed-effects through to a flexible random-effect specification); step 4: model adequacy assessment; step 5: graphical presentations; step 6: use of additional discrimination tools (‘degree of separation’; Elsensohn’s envelope of residual plots); step 7: clinical characterisation and plausibility; and step 8: sensitivity analysis. We illustrated these steps using data from the NIH-AARP cohort of repeated determinations of body mass index (BMI) at baseline (mean age: 62.5 years), and BMI derived by weight recall at ages 18, 35 and 50 years.
Results
From 288 993 participants, we derived a five-class model for each gender (men: 177 455; women: 111 538). From seven model structures, the favoured model was a proportional random quadratic structure (model F). Favourable properties were also noted for the unrestricted random quadratic structure (model G). However, class proportions varied considerably by model structure—concordance between models F and G were moderate (Cohen κ: men, 0.57; women, 0.65) but poor with other models. Model adequacy assessments, evaluations using discrimination tools, clinical plausibility and sensitivity analyses supported our model selection.
Conclusion
We propose a framework to construct and select a ‘core’ LCTM, which will facilitate generalisability of results in future studies.
Top-30
Journals
|
1
2
3
4
5
6
7
|
|
|
PLoS ONE
7 publications, 2.37%
|
|
|
Scientific Reports
6 publications, 2.03%
|
|
|
Nutrients
5 publications, 1.69%
|
|
|
Journal of Affective Disorders
5 publications, 1.69%
|
|
|
Frontiers in Medicine
4 publications, 1.36%
|
|
|
JAMA network open
4 publications, 1.36%
|
|
|
BMJ Open
4 publications, 1.36%
|
|
|
Journal of the American Heart Association
3 publications, 1.02%
|
|
|
Frontiers in Endocrinology
3 publications, 1.02%
|
|
|
Frontiers in Public Health
3 publications, 1.02%
|
|
|
Environmental Research
3 publications, 1.02%
|
|
|
Obesity Science and Practice
3 publications, 1.02%
|
|
|
Diabetic Medicine
3 publications, 1.02%
|
|
|
Rheumatology
3 publications, 1.02%
|
|
|
RMD Open
3 publications, 1.02%
|
|
|
Cancers
2 publications, 0.68%
|
|
|
Frontiers in Nutrition
2 publications, 0.68%
|
|
|
BMC Cancer
2 publications, 0.68%
|
|
|
BMC Medical Research Methodology
2 publications, 0.68%
|
|
|
Chemosphere
2 publications, 0.68%
|
|
|
Science of the Total Environment
2 publications, 0.68%
|
|
|
Arthritis and Rheumatology
2 publications, 0.68%
|
|
|
Arthritis Care and Research
2 publications, 0.68%
|
|
|
British Journal of Clinical Pharmacology
2 publications, 0.68%
|
|
|
Clinical Epidemiology
2 publications, 0.68%
|
|
|
Journal of Statistical Computation and Simulation
2 publications, 0.68%
|
|
|
Psychological Medicine
2 publications, 0.68%
|
|
|
Diabetes Care
2 publications, 0.68%
|
|
|
Diabetology and Metabolic Syndrome
2 publications, 0.68%
|
|
|
1
2
3
4
5
6
7
|
Publishers
|
10
20
30
40
50
60
70
80
|
|
|
Elsevier
75 publications, 25.42%
|
|
|
Springer Nature
55 publications, 18.64%
|
|
|
Wiley
41 publications, 13.9%
|
|
|
Frontiers Media S.A.
18 publications, 6.1%
|
|
|
Oxford University Press
14 publications, 4.75%
|
|
|
BMJ
12 publications, 4.07%
|
|
|
Taylor & Francis
10 publications, 3.39%
|
|
|
Ovid Technologies (Wolters Kluwer Health)
10 publications, 3.39%
|
|
|
MDPI
10 publications, 3.39%
|
|
|
Public Library of Science (PLoS)
9 publications, 3.05%
|
|
|
Cold Spring Harbor Laboratory
6 publications, 2.03%
|
|
|
SAGE
5 publications, 1.69%
|
|
|
American Medical Association (AMA)
5 publications, 1.69%
|
|
|
Cambridge University Press
4 publications, 1.36%
|
|
|
JMIR Publications
3 publications, 1.02%
|
|
|
American Diabetes Association
2 publications, 0.68%
|
|
|
Research Square Platform LLC
2 publications, 0.68%
|
|
|
American Academy of Pediatrics
1 publication, 0.34%
|
|
|
Australian Physiotherapy Association
1 publication, 0.34%
|
|
|
The Endocrine Society
1 publication, 0.34%
|
|
|
American Association for Cancer Research (AACR)
1 publication, 0.34%
|
|
|
S. Karger AG
1 publication, 0.34%
|
|
|
Baishideng Publishing Group
1 publication, 0.34%
|
|
|
Mark Allen Group
1 publication, 0.34%
|
|
|
Human Kinetics
1 publication, 0.34%
|
|
|
American Chemical Society (ACS)
1 publication, 0.34%
|
|
|
10
20
30
40
50
60
70
80
|
- We do not take into account publications without a DOI.
- Statistics recalculated weekly.