Quantifying Childhood Trauma: Causal Machine Learning Approaches to Mental Health Outcomes
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
We utilized the Behavioral Risk Factor Surveillance System (BRFSS) dataset for the following observational machine learning (ML) study. In contrast to recent literature using ML to predict specific outcomes, we used causal machine learning (CML) algorithms (LRSRegressors, DRLearner, Rlearner, Xregressor, and Tregressor) to quantify the cause-and-effect relationships between adverse childhood experiences (ACEs) and two different mental health targets (number of bad mental health days and whether the respondent had a depressive disorder). Specifically, we report each ACE’s average treatment effect (ATE) on each target. We also used Uplift Random Forest trees to calculate uplift scores. We found that growing up in a household where one parent had a depressive disorder and being forced into unwanted sex had the most significant effects on our chosen mental health targets. Growing up in a household where one parent had a depressive disorder increased the likelihood of being diagnosed with a depressive disorder by 30% and added 5.07 days of bad mental health in the last month. Repeatedly being forced to have sex added 5.77 days of poor mental health and increased the likelihood of being diagnosed with a depressive disorder by 33%. Uplift modeling indicated that individuals over 35 years of age earning under $100,000 USD annually were most susceptible to the effects of a parent diagnosed with depression by a factor of 1.1 compared to the average population. Male college graduates over 44 years of age earning under $100,000 USD were found to be most susceptible to the effects of childhood sexual abuse by a factor of 1.38. Future research should refine these models and employ more complicated algorithms to gain accurate, interpretable measurements and to understand the relationship between ACEs and mental health.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.014 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it