Cross-Intervention-Evidence-Comparison-v0.3.8
v0.3.9 provenance note: Program-specific PCIT-CDI-CU results in this v0.3.8 comparison were extracted from a nonpublisher article-text mirror. For Evidence-Gated Genealogy v1, use only the primary-source-hardened 2024 claim: the three parent-child interventions were associated with post-intervention improvement in sadness recognition and skin-conductance reactivity. Treat condition-specific superiority as provisional until primary full text is inspected.
type: evidence-comparison version: 0.3.8 status: complete publish: true
Cross-Intervention Evidence Comparison — v0.3.8
| Branch | Allocation | Independent unit problem | Main target | Direct endpoint | Follow-up | Strongest support | Principal limitation |
|---|---|---|---|---|---|---|---|
| 2018 school prevention | school-level random allocation | only 3 schools; 1 treated cluster | broad emotional/social/self-control skills | CU scores, CD symptoms, impulsivity, paternal involvement, friend support | 3 and 9 months | distal trait/symptom change | treatment confounded with one intervention school; child-level ANOVA does not model clustering |
| 2024 PCIT-CDI-CU | individual/family randomization | small post-dropout cells | parental warmth / positive dyadic interaction | sadness recognition + SC reactivity | ~3 months | intervention-specific SC change; promising sadness-recognition pattern | sadness accuracy lacked significant Time × Condition interaction; no CU/CD outcome |
| 2024 CARES | individual/family randomization | small post-dropout cells | emotional literacy / cue decoding | sadness recognition + SC reactivity | ~3 months | no clear intervention-specific effect in this study | small n; no distal outcome |
| 2024 Emotional Engagement | individual/family randomization | small post-dropout cells | emotional engagement / eye gaze | sadness recognition + SC reactivity | ~3 months | no clear intervention-specific effect in this study | small n; no distal outcome |
| HRV biofeedback | individual intervention vs wait-list | modest sample / subgroup slicing | physiological regulation | SC and HRV-derived stress responses | none beyond immediate post in audited dissertation | task/subgroup-specific physiology change | no durable symptom/trait outcome demonstrated; LF/HF interpretive limits |
| cTBS / AMT | real/sham and attention-training comparisons | community samples | attention / cortical processing | gaze, attention, emotion recognition; anxiety in one trial | short-term | proximal mechanism manipulation | no APD/ASPD symptom reduction; no added cTBS anxiety benefit |
| VR ecosystem | RCT evidence exists adjacent to Fanti; direct Fanti attribution unresolved | attribution boundary | emotion/social learning | CU/conduct/reactive aggression/emotion-recognition in Impact VR papers | up to 3 months | adjacent translational efficacy evidence | not direct Fanti-authored evidence for composite 2026 VR claim |
Core lesson for the genealogy
Evidence strength is multidimensional. "Later" and "more technological" do not mean "more clinically demonstrated."
The atlas should render at least four separate axes:
- randomization/causal control;
- endpoint distance from real-world behavior;
- durability/follow-up;
- generalizability/replication.
