The exact text we saved

This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.

TRUVACE RECORD VERSION
record: TRV-2026-2815
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-11-12T17:28:09.392777Z
status: published
lens: g_space
sector: education
headline: Effects of an AI-enhanced BOPPPS teaching model in nursing courses: a meta-analysis of randomized controlled trials
dek: Aim To evaluate the effectiveness of an artificial intelligence (AI)–enhanced BOPPPS teaching model in nursing courses, with a focus on academic performance, self-directed learning ability, and teaching satisfaction. Design Systematic review and meta-analysis of randomized controlled trials (RCTs) conducted in accordance with PRISMA and registered in INPLASY (INPLASY202590123). Methods Chinese- and English-language databases were searched for eligible RCTs comparing AI-enhanced BOPPPS with traditional teaching (…
gain_title: A meta-analysis of four randomized trials with 459 nursing students found AI-enhanced BOPPPS teaching raised academic performance (SMD 1.06) and tripled the odds of an 'excellent' self-directed learning rating (RR 3.28) versus traditional teaching.
problem_title: (none)
trace_subject: (none)
gain_reading: A meta-analysis of four randomized trials with 459 nursing students found AI-enhanced BOPPPS teaching raised academic performance (SMD 1.06) and tripled the odds of an 'excellent' self-directed learning rating (RR 3.28) versus traditional teaching.
gain_evidence: Academic performance showed a significant improvement in the AI-enhanced BOPPPS group (SMD = 1.06, 95% CI 0.63–1.48; p < 0.05) | Self-directed learning ability (3 studies; “excellent” category) was significantly improved in the AI-enhanced BOPPPS group (RR = 3.28, 95% CI 2.14–5.02; p < 0.001)
problem_reading: (none)
problem_evidence: (none)
quick_read: A systematic review and meta-analysis pooled four randomized controlled trials of 459 nursing students in China-linked courses. It compared BOPPPS, a six-phase lesson structure, enhanced with AI against traditional teaching or BOPPPS alone. Academic performance improved in the AI group (SMD 1.06, 95% CI 0.63–1.48).

Students in the AI-enhanced groups were more likely to be rated 'excellent' in self-directed learning (RR 3.28) and 'very satisfied' with teaching (RR 1.80, 95% CI 1.27–2.55). The authors say this suggests value lies in learning processes and student experience rather than test scores alone.

The evidence base is small. Heterogeneity in the academic performance result was substantial (I² = 72.8%), and the AI tools varied from ChatGPT-assisted teaching to a 5G smart classroom, so the pooled effect cannot be tied to one kind of AI.
limitation: The review included only four mostly single-center trials with short-term interventions and no standardized way of integrating AI, and blinding was inadequate in some trials.
tag: Evidence-backed gain
key_points: The review searched PubMed, Web of Science, Embase, the Cochrane Library, CNKI, WanFang and VIP through October 14, 2025, and was registered with INPLASY. | Four RCTs were included, with 219 students in intervention groups and 212 in control groups taught by conventional teacher-centered lecturing. | Three studies reported academic performance, with a pooled SMD of 1.06 and heterogeneity of 72.8%. | Three studies reported self-directed learning; the share rated 'excellent' was 3.28 times higher with AI, with no heterogeneity (I² = 0.0%). | Three studies reported satisfaction; 'very satisfied' ratings were 1.80 times as likely, with I² of 37.4%. | Two trials did not adequately report or implement blinding of participants and personnel, and most did not describe allocation concealment.
rundown: The included trials covered different AI approaches. One used an AI-enhanced BOPPPS model, another a BOPPPS-based 5G smart classroom, and two used ChatGPT-assisted BOPPPS. Courses included neurosurgery, thoracic nursing and pathophysiology, and outcomes ranged from theory scores to internship performance.

Risk-of-bias review rated the trials as low risk overall, with complete outcome data and no selective reporting. Randomization procedures were often unclear, and blinding of participants was a problem in two trials, which the authors say may have introduced observer bias.

The authors attribute the variable academic results to differences in AI implementation, course type, assessment format and student baselines. They call for standardized AI components, multicenter trials and long-term follow-up on clinical competence.
sources:
- peer_reviewed | Frontiers in Medicine | https://doi.org/10.3389/fmed.2026.1767911 | 2026-02-18
prev: 0000000000000000000000000000000000000000000000000000000000000000

SHA-256 hash

ffd8be070552b53668d3269498d9a0a5c37d5dd93859db819b3319b9dcb5caab

Previous version’s hash

0000000000000000000000000000000000000000000000000000000000000000