Melissa J. Bak

Melissa J. Bak

Auto-generated profile, based on publicly available information

Melissa J. Bak

Northwestern University Feinberg School of Medicine, 420 E. Superior St., Chicago, IL 60611, USA

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
3 June 2024 in Fertility and Sterility
Verified
Journal Article
Closed access
MBMelissa J. Bak
Melissa J. Bak
DFDoris Farje
Doris Farje
ETElizabeth L Tsui
Elizabeth L Tsui
TLTimothy B. Lautz
Timothy B. Lautz
Monica M LarondaMonica M Laronda
Monica M Laronda
ERErin E Rowell
Erin E Rowell

7 June 2025 in Journal of Pediatric Surgery
Verified
Journal Article
Open access
LCLauren E. Corona
Lauren E. Corona
MBMelissa Bak
Melissa Bak
DFDoris Farje
Doris Farje
IRIlina Rosoklija
Ilina Rosoklija
TKTara Kennedy
Tara Kennedy
ERErin E Rowell
Erin E Rowell
Clinical guidelines recommend discussing infertility risk and referral to fertility preservation (FP) specialists in all patients facing cancer treatment, including children. We sought to determine whether sociodemographic or clinical factors led to differences in referral and utilization of FP services at a children's hospital with a formal FP program. We hypothesized that sociodemographic differences would exist in FP utilization, but not in referral rates. We abstracted sociodemographic and clinical details on patients receiving gonadotoxic treatment at a large urban children's hospital. Primary outcome was referral for FP consultation or documentation of fertility-related discussion. Secondary outcome was FP completion when Pediatric Initiative Network Risk Stratification System (PINRSS) risk of infertility was significant or high. Outcomes were compared using descriptive statistics, and a Poisson regression with robust variance analysis used to determine adjusted risks of outcomes. 66 % (114/172) of the total cohort and all (80/80) significant- and high-risk patients had an FP consult or documentation of fertility discussion. On adjusted analysis, only minimal risk by PINRSS was associated with a lower probability of fertility-related discussion or consult and no variables independently predicted FP completion. Sociodemographic parity existed in both outcomes. The FP program supported our multi-disciplinary efforts for FP consults; when treatment-related increased risk of infertility was significant or high, a risk-related discussion/consult always occurred. Fertility discussions were less likely when infertility risks were minimal. FP consults and risk-related discussions were distributed equitably, and no sociodemographic differences were observed between those who did and did not pursue FP.

17 January 2025 in medRxiv
Verified
Preprint
Open access
PDF file
VGVijeeth Guggilla
Vijeeth Guggilla
MKMengjia Kang
Mengjia Kang
MBMelissa J Bak
Melissa J Bak
STSteven D Tran
Steven D Tran
APAnna Pawlowski
Anna Pawlowski
TIThe NU SCRIPT Study Investigators
The NU SCRIPT Study Investigators
Identifying immunosuppressed patients using structured data can be challenging. Large language models effectively extract structured concepts from unstructured clinical text. Here we show that GPT-4o outperforms traditional approaches in identifying immunosuppressive conditions and medication use by processing hospital admission notes. We also demonstrate the extensibility of our approach in an external dataset. Cost-effective models like GPT-4o mini and Llama 3.1 also perform well, but not as well as GPT-4o.

30 June 2024 in medRxiv
Verified
Preprint
Open access
PDF file
SFSamuel W Fenske
Samuel W Fenske
APAlec Peltekian
Alec Peltekian
MKMengjia Kang
Mengjia Kang
NMNikolay S Markov
Nikolay S Markov
MZMengou Zhu
Mengou Zhu
TIthe NU SCRIPT Study Investigators
the NU SCRIPT Study Investigators
Background Criteria to identify patients who are ready to be liberated from mechanical ventilation are imprecise, often resulting in prolonged mechanical ventilation or reintubation, both of which are associated with adverse outcomes. Daily protocol-driven assessment of the need for mechanical ventilation leads to earlier extubation but requires dedicated personnel. We sought to determine whether machine learning applied to the electronic health record could predict successful extubation. Methods We examined 37 clinical features from patients from a single-center prospective cohort study of patients in our quaternary care medical ICU who required mechanical ventilation and underwent a bronchoalveolar lavage for known or suspected pneumonia. We also tested our models on an external test set from a community hospital ICU in our health care system. We curated electronic health record data aggregated from midnight to 8AM and labeled extubation status. We deployed three data encoding/imputation strategies and built XGBoost, LightGBM, logistic regression, LSTM, and RNN models to predict successful next-day extubation. We evaluated each model’s performance using Area Under the Receiver Operating Characteristic (AUROC), Area Under the Precision Recall Curve (AUPRC), Sensitivity (Recall), Specificity, PPV (Precision), Accuracy, and F1-Score. Results Our internal cohort included 696 patients and 9,828 ICU days, and our external cohort had 333 patients and 2,835 ICU days. The best model (LSTM) predicted successful extubation on a given ICU day with an AUROC 0.87 (95% CI 0.834-0.902) and the internal test set and 0.87 (95% CI 0.848-0.885) on the external test set. A Logistic Regression model performed similarly (AUROC 0.86 internal test, 0.83 external test). Across multiple model types, measures previously demonstrated to be important in determining readiness for extubation were found to be most informative, including plateau pressure and Richmond Agitation Sedation Scale (RASS) score. Our model often predicted patients to be stable for extubation in the days preceding their actual extubation, with 63.8% of predicted extubations occurring within three days of true extubation. We also tested the best model on cases of failed extubations (requiring reintubation within two days) not seen by the model during training. Our best model would have identified 35.4% (17/48) of these cases in the internal test set and 48.1% (13/27) cases in the external test set as unlikely to be successfully extubated. Conclusions Machine learning models can accurately predict the likelihood of extubation on a given ICU day from data available in the electronic health record. Predictions from these models are driven by clinical features that have been associated with successful extubation in clinical trials.