Gradebook data may flag substance use risk in middle schoolers

Teacher-recorded classroom engagement scores, a routinely collected but rarely analyzed form of school data, distinguished middle school students who went on to report alcohol or cannabis use from those who did not more accurately than grade point average (GPA), according to a study published in the Journal of School Health.¹ A model combining gradebook engagement scores with student sex achieved a cross-validated area under the curve (AUC) of 0.79, compared with 0.63 for a GPA-only model.¹

Engagement and participation scores were "more accurate signals for identifying students at risk of substance use," said Benjamin P. L. Meza, MD, assistant professor of medicine and pediatrics at UCLA Health and one of the study's lead authors.² He added that, if replicated, the approach could help reach children who rarely see a physician and may miss routine screening.²

Gradebook data and adolescent substance use: LIFT trial secondary analysis

Investigators conducted a secondary analysis of the Linking Information and Families Together (LIFT) trial (NCT02129153), a teacher-masked randomized trial in which parents received weekly messages about their child's missed assignments, grades, and behavior.¹,³ Students were recruited in sixth grade from 4 public middle schools in Los Angeles, California, in 2014 and 2015 and followed through eighth grade.¹

The primary outcome was self-reported lifetime alcohol or cannabis use by eighth grade.¹ Predictors included transcript-recorded GPA, parent-reported GPA, and 3 gradebook domains: summative (quiz and test scores), formative (homework), and engagement (classroom behavior and participation).¹

Of 234 students in the complete-case analysis, 51.7% were female, 82.1% were Latino, and nearly half of families reported annual household income below $15,000.¹ Forty-seven students reported lifetime substance use by eighth grade.¹

Classroom engagement scores outperformed GPA in predicting substance use

In adjusted models, higher engagement scores were associated with lower odds of substance use (OR, 0.84 per point; P < .001), and formative homework scores showed a smaller association (OR, 0.95; P = .02).¹ Transcript-recorded GPA was not significantly associated with the outcome in unadjusted analyses.¹

Because 1 school did not record engagement data, engagement models were limited to 170 students.¹ In that subgroup, the engagement model's cross-validated AUC was 0.79 (95% CI, 0.68–0.87), and adding parent-reported GPA did not improve discrimination (AUC, 0.74).¹ At the threshold maximizing the Youden index, sensitivity was 96.0% and specificity was 55.2%.¹ Multiple imputation sensitivity analyses produced similar estimates.¹

Why school data may matter for adolescent substance use screening

The American Academy of Pediatrics recommends substance use screening, brief intervention, and referral to treatment as part of routine adolescent health care.⁴ Yet about one-third of adolescents do not attend an annual well visit, the authors note.¹ Rebecca N. Dudovitz, MD, general pediatrics division chief at UCLA Mattel Children's Hospital and a study coauthor, called that gap "a critical blind spot for early substance abuse screening and intervention."²

The authors frame gradebook data as a passive early indicator, rhetorically analogous to a vital sign, that requires no added effort from teachers or students.¹ They envision it as one input for multidisciplinary school support teams rather than a stand-alone screening tool or a basis for discipline.¹

Limitations of gradebook-based risk prediction

The sample was small, drawn from a single urban region, and predominantly Latino and low-income, which limits generalizability.¹ The outcome was self-reported lifetime use rather than misuse or substance use disorder, and data were collected between 2014 and 2016.¹ Grading practices vary across classrooms and schools and may reflect implicit bias; the authors call for local validation, model transparency, and privacy safeguards consistent with the Family Educational Rights and Privacy Act before any implementation.¹

The engagement model's modest specificity also means a gradebook-based flag would generate false positives, supporting its use as a prompt for supportive follow-up rather than a diagnosis.¹

References

1. DiPentino R, Meza BPL, Jackson N, Dosanjh KK, Dudovitz RN, Wong MD. Could school function be an adolescent health vital sign? Associations between gradebook data and substance use among middle school students. J Sch Health. 2026;96(10):e70261. doi:10.1111/josh.70261

2. Cruz R. UCLA Health study finds gradebook data can help predict risk of drug use in adolescents. News release. UCLA Health. November 10, 2026. Accessed November 11, 2026. https://www.uclahealth.org/news/release/ucla-health-study-finds-gradebook-data-can-help-predict-risk

3. Bergman P, Dudovitz RN, Dosanjh KK, Wong MD. Engaging parents to prevent adolescent substance use: a randomized controlled trial. Am J Public Health. 2019;109(10):1455-1461. doi:10.2105/AJPH.2019.305240

4. Levy SJ, Williams JF; Committee on Substance Use and Prevention. Substance use screening, brief intervention, and referral to treatment. Pediatrics. 2016;138(1):e20161211. doi:10.1542/peds.2016-1211