The research team, led by Tong Xiao and Yiru Wang, who contributed equally to the work, with Yanhong Zhang as corresponding author, searched ten databases spanning the English-language and Chinese biomedical literature, including PubMed, Embase, Web of Science, CINAHL, the Cochrane Library, PsycINFO, and four major Chinese databases. Their search covered everything published from database inception through August 15, 2025. From that sweep they identified 27 studies containing 58 distinct prediction models for adolescent non-suicidal self-injury. Seventeen of the source studies were conducted in hospital settings and ten in community settings, a split that matters because the predictors available and the base rates of self-injury can differ substantially between a psychiatric clinic and a school population.

The centerpiece of the analysis is the pooled area under the receiver operating characteristic curve, or AUC, the most widely used summary of how well a model separates those who will experience an outcome from those who will not. An AUC of 0.5 reflects performance no better than a coin flip, while 1.0 represents perfect discrimination. Across the 17 models for which training-set data could be pooled, the meta-analysis found a combined AUC of 0.850, with a 95 percent confidence interval of 0.810 to 0.890. For the far smaller set of five models evaluated on independent validation data, the pooled AUC was 0.840, with a wider interval of 0.760 to 0.930. In conventional terms, that places these tools in the moderate to good range of predictive performance, comparable to risk models used in other areas of medicine.

But the authors are quick to point out that a strong AUC on a training set is not the same as a trustworthy clinical instrument. Only a handful of the 58 models had ever been externally validated, meaning tested on data not used to build them, and the confidence interval for the validation estimate is wide enough to encompass everything from mediocre to excellent performance. Discrimination is only half the story in prediction research. A model must also be well calibrated, meaning that when it says a patient has a 30 percent risk, roughly 30 percent of similar patients should actually experience the outcome. Calibration was rarely assessed or reported in the studies the team reviewed, a gap that has plagued prediction modeling across medicine and that the authors single out as a priority for future work.

Methodological quality emerged as the study’s most sobering finding. Using the PROBAST + AI tool, a risk-of-bias instrument designed specifically for prediction models, including those built with artificial intelligence, the reviewers judged 24 of the 27 studies to be of low overall quality and 23 to be at high risk of bias. Two studies were rated at high risk of concerns regarding applicability, meaning their data or predictors might not transfer well to the populations clinicians actually want to serve. Common sources of bias in prediction research include small sample sizes relative to the number of predictors, inappropriate handling of missing data, optimistic performance estimates from evaluating models on the same data used to develop them, and selective reporting of the best-performing model among many that were tried.

What drives the risk, according to the models themselves? Across the included studies, a consistent cluster of predictors surfaced again and again. Childhood trauma was among the most powerful signals, echoing decades of research linking early adversity to self-directed harm. Depressive mood and anxiety, the emotional workhorses of adolescent risk screening, appeared frequently, as did a history of prior non-suicidal self-injury, which is unsurprising given that past behavior is one of the strongest predictors of future behavior in nearly every domain of mental health. Sleep disorders, stressful life events, and female gender rounded out the list of common predictors. The recurrence of these variables across diverse settings suggests a reasonably stable risk architecture, even if the models that combine them vary widely in construction.

The implications for clinical practice are double-edged. On one hand, the finding that pooled AUCs hover around 0.85 suggests that a data-informed approach to identifying at-risk adolescents is feasible. Schools, pediatric clinics, and emergency departments could, in principle, deploy brief screening tools that combine a small number of well-validated predictors to flag young people who warrant deeper assessment. On the other hand, the authors conclude that the existing models perform poorly in methodological quality and clinical applicability, and they explicitly call for future research to strengthen validation and calibration of existing models and to improve methodological rigor in line with PROBAST criteria. Until that happens, they caution, these tools should not be treated as ready-made clinical decision instruments.

The review also highlights a structural weakness of the field: fragmentation. With 58 models across 27 studies, many built on small, single-center samples and using different predictor sets, statistical techniques, and outcome definitions, there is little opportunity to compare approaches head to head or to pool their strengths. The authors suggest that the path forward lies less in building yet more novel models and more in rigorously validating, calibrating, and refining the ones that already exist, ideally in large, diverse, prospectively collected cohorts. Reporting standards such as TRIPOD, the transparency checklist for prediction studies, and the PROBAST framework for assessing bias offer a shared language for judging whether a model deserves clinical trust.

For a behavior as hidden and as consequential as adolescent self-injury, the promise of early, data-driven identification is genuinely compelling. Non-suicidal self-injury is associated with serious adverse effects on individuals, families, and society, and it frequently goes undetected until it has become entrenched. A validated screening model could shift the clinical posture from reaction to prevention, directing scarce mental health resources toward the adolescents who need them most. This meta-analysis provides the clearest picture yet of what that future might look like: a field with real predictive signal, a recognizable set of risk factors, and performance statistics that look respectable on paper, but one that still must clear the harder hurdles of external validation, honest calibration, and methodological discipline before its algorithms earn a place in the clinic.

The study, funded by the Nanjing Health Science and Technology Development Special Fund Project and published open access, arrives at a moment when machine-learning tools are being proposed for nearly every corner of mental health care. Its message generalizes well beyond self-injury. Impressive AUCs are easy to produce and hard to trust; the difference lies in validation, calibration, and transparent reporting. For the researchers, clinicians, and families hoping that prediction models can help protect vulnerable teenagers, the new review offers both a map of what has been achieved and a candid account of how much work remains before those achievements can be safely put to use.

Subject of Research: Risk prediction models for non-suicidal self-injury among adolescents

Article Title: Risk prediction models of non-suicidal self-injury among adolescents: a systematic review and meta-analysis

Article References: Risk prediction models of non-suicidal self-injury among adolescents: a systematic review and meta-analysis. (n.d.). https://doi.org/10.1186/s12888-026-08719-1

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08719-1

Keywords: non-suicidal self-injury, adolescents, risk prediction models, systematic review, meta-analysis, machine learning, PROBAST, AUC, childhood trauma, depression, anxiety, mental health screening

Preeclampsia is a pregnancy complication characterized by new-onset hypertension accompanied by proteinuria, maternal end-organ dysfunction, or uteroplacental dysfunction. When it goes unrecognized or untreated, it remains one of the leading causes of maternal mortality. The consequences extend well beyond delivery: hypertensive complications in pregnancy are linked to long-term maternal cardiovascular disease and reduced quality of life, while severe disease increases the risk of fetal growth restriction and preterm birth through underlying placental insufficiency. Timely identification of women at risk is therefore a central goal of modern obstetric medicine, and a growing library of prediction models has been developed to flag high-risk pregnancies before symptoms appear.

Why would ART pregnancies be more vulnerable in the first place? The review outlines several converging explanations. The need for ART itself may reflect underlying reproductive characteristics associated with elevated preeclampsia risk, including advanced maternal age, nulliparity after prolonged infertility, and specific infertility diagnoses. More provocatively, recent evidence suggests that certain ART-related factors may causally contribute to risk. Programmed frozen embryo transfer cycles, for example, lack a corpus luteum and its vasoactive products, and multiple embryo transfers can produce multiple gestations, both of which are associated with hypertensive complications. Altered maternal-fetal immune interactions, particularly evident in oocyte donation pregnancies, have also been hypothesized to play a role in the development of the disease.

Despite this well-established link, the researchers found that preeclampsia prediction models almost never account for how a pregnancy was conceived. The team, led by Géraldine Lafeber, systematically searched MEDLINE for studies published between January 2017 and July 2023, updating a previous systematic review from 2019. Out of 15,305 records initially identified, 161 articles were assessed in full text, ultimately yielding fifteen eligible studies; eleven more were added from the earlier review, for a total of twenty-six articles describing models that included ART as a predictor. Remarkably, across the roughly 231 prediction models identified in the literature, not a single one had been developed specifically for the ART population.

The included studies spanned 2009 to 2023 and collectively involved 625,001 participants, of whom 20,142 developed preeclampsia. Most were conducted in Europe, with six studies from Asia and one from North America. The models fell into two broad statistical families. Frequentist models estimate parameters based on observations alone, using metrics such as confidence intervals and p-values, while Bayesian models incorporate prior knowledge, updating beliefs as new data become available to produce probability distributions for model parameters. Some studies even applied machine learning approaches, including random forest, XGBoost and support vector machines. Predictors typically combined clinical parameters with biomarkers such as the uterine artery pulsatility index, mean arterial pressure, placental growth factor and pregnancy-associated plasma protein A.

Performance varied enormously. Discrimination, measured by the area under the receiver operating characteristic curve, ranged from a nearly useless 0.58 to an almost perfect 0.99. A value of 0.5 indicates performance no better than random guessing, while 1.0 indicates perfect discrimination. Bayesian models consistently improved when biomarkers were added to maternal factors alone. Yet the headline numbers conceal deeper problems: calibration, which assesses whether predicted probabilities actually match observed outcome frequencies, was rarely reported. Only seven of the twenty-six models had ever been externally validated in an independent population, a step widely considered essential before any prediction model is used in clinical practice.

The methodological quality assessment, conducted with the PROBAST tool, revealed that risk of bias was unclear in most studies, twenty-one of the twenty-six, largely due to analytical issues. Some studies excluded patients with missing outcome data without justification, and many reported too few events per variable, raising the risk of model overfitting. Outcome definitions of preeclampsia varied widely across studies, drawing on different guidelines published over decades, which complicates comparability and generalizability. Reporting quality, assessed against the TRIPOD checklist, ranged from 41 to 89 percent of applicable items met. Only seven studies described their variable selection and model building process in detail, and just three published their full models with individual coefficients, making independent reproduction effectively impossible.

One of the most intriguing findings concerns how ART itself was handled. In most frequentist models, ART was reduced to a simple yes-or-no variable, and in some Bayesian models the method of conception was buried within unspecified maternal factors, requiring the reviewers to contact study authors for clarification. Almost no model distinguished between IVF, ICSI, frozen or thawed embryo transfer and oocyte donation, despite substantial variation in preeclampsia risk across these treatment modalities. The risk of hypertensive complications in oocyte donation is more than double that seen with conventional IVF and ICSI, and embryo or double donation may carry an even higher risk. Treating ART as a single homogeneous category, the authors argue, may seriously compromise risk assessment for the very patients who need it most.

There was also a telling inverse relationship between apparent performance and transparency. The models reporting the highest discrimination had the lowest reporting scores, leaving it unclear whether outcome and predictor assessment were blinded, how predictions were calculated, how missing data were handled, or how the model should be used. None of the studies presented their models as online calculators for clinicians or patients. The reviewers note that despite the publication of the TRIPOD statement in 2015, which was designed to improve exactly these shortcomings, reporting quality does not appear to have improved over time.

The review does have limitations the authors acknowledge, including reliance on a single database, MEDLINE, and a gap between the July 2023 search end date and manuscript completion. Nonetheless, its conclusion is unambiguous: no existing prediction model can be endorsed for clinical use in ART pregnancies. The team argues that future research should prioritize developing a high-quality model that incorporates ART, ideally disaggregated by treatment type, and that such a model should be developed and validated within ART populations specifically. Better risk assessment could support periconceptional counselling, improve early recognition of disease, deepen understanding of preeclampsia pathophysiology, and even help minimize donor exposure to the risks associated with oocyte retrieval. The researchers themselves are already working toward a dedicated prediction model for hypertensive complications in oocyte and double donation pregnancies, following PROBAST and TRIPOD guidelines. For the growing population of families conceived through reproductive technology, that work cannot come soon enough.

Subject of Research: Prediction models for preeclampsia risk in assisted reproductive technology pregnancies

Article Title: Assisted Reproductive Technology as Predictor for Preeclampsia: A Systematic Review of Current Clinical Prediction Models

Article References: Lafeber, G. C. M., van der Hoorn, M.-L. P., le Cessie, S., & Lashley, E. E. L. O. (2026). Assisted Reproductive Technology as Predictor for Preeclampsia: A Systematic Review of Current Clinical Prediction Models. Reproductive Sciences. https://doi.org/10.1007/s43032-026-02168-7

Image Credits: AI Generated

DOI: 10.1007/s43032-026-02168-7

Keywords: assisted reproductive technology, preeclampsia, prediction models, IVF, oocyte donation, systematic review, PROBAST, TRIPOD, hypertensive disorders, maternal health, reproductive medicine, external validation