I’m a high-school student in Bangkok learning how data, technology, and policy might improve decisions in health systems. I want to share what I found, be honest about what I still do not know, and listen to how others see these problems.
Evidence briefs
Can we see where a system fails clearly enough to design a better response? Scan the signal first; open any brief for the full academic record.
Medical imaging · Computer vision · 2023–2024
A student-built object detector could localise fractures in the reported test set.
YOLOv7 · retrospective image dataset
No external or prospective clinical validation; two reported metrics require clarification.
Computational pathology · Deep learning · 2025–2026
Attention-based multiple-instance learning was feasible, but discrimination remained limited.
Attention MIL · whole-slide pathology
The analytical sample and class counts remain incompletely documented; no external clinical validation.
Emergency medicine · Health financing · 2025–2026
Surgery increased after financing expansion; measured mortality did not significantly change.
National claims · retrospective cohort
Unreviewed pre/post analysis with major residual confounding from injury severity.
Health policy · Political analysis · 2026
Party platforms reveal different allocations of rights, risk, and responsibility.
Document review · welfare-state analysis
A time-bound interpretive analysis—not a party ranking, endorsement, or implementation forecast.
How I try to work
Lead with the central result and its denominator so the substance is visible within seconds.
Limitations sit beside the result they qualify, not buried at the end of the page.
Published work, competition reports, completed studies, and unreviewed drafts are explicitly distinguished.
Every brief includes the question, methods, authorship, governance, data statement, and citation.
03 / CONNECT
If we met at a conference—or if you work on emergency systems, health data, medical AI, or public policy—please send a question, a critique, or a problem you are trying to solve. I would genuinely like to learn from the conversation.
thanakorn@thanakorn.site