MPMila Pastrak

Mila Pastrak

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Mila Pastrak has a total of 26 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Evaluation of ChatGPT Performance on Emergency Medicine Board Examination Questions: Observational Study

    by Mila Pastrak, Sten Kajitani, Anthony James Goodings, Austin Drewek, Andrew LaFree, Adrian Murphy

    Background: The ever-evolving field of medicine has highlighted the potential for ChatGPT as an assistive platform. However, its use in medical board examination preparation and completion remains unclear. Objective: This study aimed to evaluate the performance of a custom-modified version of ChatGPT-4, tailored with emergency medicine board examination preparatory materials (Anki flashcard deck), compared to its default version and previous iteration (3.5). The goal was to assess the accuracy of ChatGPT-4 answering board-style questions and its suitability as a tool to aid students and trainees in standardized examination preparation. Methods: A comparative analysis was conducted using a random selection of 598 questions from the Rosh In-Training Examination Question Bank. The subjects of the study included three versions of ChatGPT: the Default, a Custom, and ChatGPT-3.5. The accuracy, response length, medical discipline subgroups, and underlying causes of error were analyzed. Results: The Custom version did not demonstrate a significant improvement in accuracy over the Default version (P=.61), although both significantly outperformed ChatGPT-3.5 (P<.001). The Default version produced significantly longer responses than the Custom version, with the mean (SD) values being 1371 (444) and 929 (408), respectively (P<.001). Subgroup analysis revealed no significant difference in the performance across different medical subdisciplines between the versions (P>.05 in all cases). Both the versions of ChatGPT-4 had similar underlying error types (P>.05 in all cases) and had a 99% predicted probability of passing while ChatGPT-3.5 had an 85% probability. Conclusions: The findings suggest that while newer versions of ChatGPT exhibit improved performance in emergency medicine board examination preparation, specific enhancement with a comprehensive Anki flashcard deck on the topic does not significantly impact accuracy. The study highlights the potential of ChatGPT-4 as a tool for medical education, capable of providing accurate support across a wide range of topics in emergency medicine in its default form.


  • Open Access

    Physician and Endoscopy Nurse Perspectives on Tap Water for Gastrointestinal Endoscopy: A Cross-Sectional Survey on Support, Perceived Barriers, and Intent to Advocate

    by Anthony James Goodings, Allison Dana Chhor, Hannah Anderson, Mila Pastrak, Sten Kajitani, Matthew Schultzel, Aoife O’sullivan, Ann-Marie Eustaceryan

    Background: The standard water used for endoscopic irrigation is sterile water. Minimal evidence exists regarding sterile water use where there is access to clean water. The WHO has declared the climate crisis as the greatest global health crisis today; we must re-examine our practices and adapt them to promote environmental stewardship while maintaining safety. Method and Objective: We surveyed physicians and endoscopy nurses to determine their attitudes toward tap water use for irrigation in gastrointestinal endoscopic procedures. Results: There were 88 complete responses collected from June to November 2024. The majority of respondents and endoscopy-performing consultants expressed comfort with tap water use (59% and 84%, respectively), perceived viability (62% and 68%, respectively), and an interest to implement (73% and 94%, respectively); however, discussions on the topic remained infrequent (77% and 81%, respectively). 82% of overall respondents and 93% of consultants were aware of potential cost-savings, with 69% and 87% more willing to consider tap water based on this. Respondents (60%) and consultants (73%) agree there is a lack of guidelines regarding tap water use and feel that policy barriers will hinder change (59% and 73% respectively). Overall, 59% of respondents and 73% of consultants are likely to advocate for change. Conclusion: The majority of respondents support tap water as a viable, cost-effective alternative with environmental benefits. A strong intention to advocate for change highlights the presence of potential leaders in this space. By promoting and supporting these leaders through education and institutional change, a more sustainable future for endoscopy exists. What is already known on this topic: There is an urgent need to address the immense impact of medicine on the environment. Gastrointestinal endoscopy involves the use of large volumes of sterile water to irrigate a non-sterile space in the body. What this study adds: Allows us to understand the views of physicians on the potential use of tap water in GI endoscopy, as well as understand perceived feasibility and barriers. How this study might affect research, practice, or policy: With an understanding of physician support and perceived barriers, specific actions to address these can be taken by regulators. By increasing awareness around the topic, experts can deliberate on the idea and choose to advocate for change, providing leadership in an intimidating discipline.

  • Open Access

    Generative AI’s Performance on Emergency Medicine Boards Questions: Observational Study (Preprint)

    by Sten Kajitani, Mila Pastrak, Anthony Goodings, Austin Drewek, Andrew Lafree, Adrian Murphy

    BACKGROUND The ever evolving field of medicine has highlighted the potential for ChatGPT as an assistive platform. However, its use in medical board exam preparation and completion remains divided. OBJECTIVE This study aimed to evaluate the performance of a custom-modified version of ChatGPT-4, tailored with emergency medicine board exam preparatory materials (Anki deck), compared to its default version and previous iteration (3.5). The goal was to assess the accuracy of ChatGPT-4 answering board- style questions and its suitability as a tool for medical education. METHODS A comparative analysis was conducted using a random selection of 598 questions from the Rosh In-Training Exam Question Bank. The subjects of the study included three versions of ChatGPT: the Default, a Custom, and ChatGPT-3.5. Accuracy, response length, medical discipline subgroups, and underlying causes of error were analyzed. RESULTS Custom did not demonstrate a significant improvement in accuracy over Default (p>0.05), though both significantly outperformed ChatGPT-3.5 (p0.05) and had a 99% predicted probability of passing while ChatGPT- 3.5 had an 85%. CONCLUSIONS The findings suggest that while newer versions of ChatGPT exhibit improved performance in emergency medicine board exam preparation, specific enhancements do not significantly impact accuracy. The study highlights the potential of ChatGPT-4 as a tool for medical education, capable of providing accurate support across a wide range of topics in emergency medicine.

  • Open Access

    Assessment of ChatGPT-4 in Family Medicine Board Examinations Using Advanced AI Learning and Analytical Methods: Observational Study

    by Anthony James Goodings, Sten Kajitani, Allison Chhor, Ahmad Albakri, Mila Pastrak, Megha Kodancha, Rowan Ives, Yoo Bin Lee, Kari Kajitani

    Background: This research explores the capabilities of ChatGPT-4 in passing the American Board of Family Medicine (ABFM) Certification Examination. Addressing a gap in existing literature, where earlier artificial intelligence (AI) models showed limitations in medical board examinations, this study evaluates the enhanced features and potential of ChatGPT-4, especially in document analysis and information synthesis. Objective: The primary goal is to assess whether ChatGPT-4, when provided with extensive preparation resources and when using sophisticated data analysis, can achieve a score equal to or above the passing threshold for the Family Medicine Board Examinations. Methods: In this study, ChatGPT-4 was embedded in a specialized subenvironment, “AI Family Medicine Board Exam Taker,” designed to closely mimic the conditions of the ABFM Certification Examination. This subenvironment enabled the AI to access and analyze a range of relevant study materials, including a primary medical textbook and supplementary web-based resources. The AI was presented with a series of ABFM-type examination questions, reflecting the breadth and complexity typical of the examination. Emphasis was placed on assessing the AI’s ability to interpret and respond to these questions accurately, leveraging its advanced data processing and analysis capabilities within this controlled subenvironment. Results: In our study, ChatGPT-4’s performance was quantitatively assessed on 300 practice ABFM examination questions. The AI achieved a correct response rate of 88.67% (95% CI 85.08%-92.25%) for the Custom Robot version and 87.33% (95% CI 83.57%-91.10%) for the Regular version. Statistical analysis, including the McNemar test (P=.45), indicated no significant difference in accuracy between the 2 versions. In addition, the chi-square test for error-type distribution (P=.32) revealed no significant variation in the pattern of errors across versions. These results highlight ChatGPT-4’s capacity for high-level performance and consistency in responding to complex medical examination questions under controlled conditions. Conclusions: The study demonstrates that ChatGPT-4, particularly when equipped with specialized preparation and when operating in a tailored subenvironment, shows promising potential in handling the intricacies of medical board examinations. While its performance is comparable with the expected standards for passing the ABFM Certification Examination, further enhancements in AI technology and tailored training methods could push these capabilities to new heights. This exploration opens avenues for integrating AI tools such as ChatGPT-4 in medical education and assessment, emphasizing the importance of continuous advancement and specialized training in medical applications of AI.


  • Open Access

    A Three-Year Retrospective Review of Gynecologic Oncology Referrals to the Specialist Palliative Care Team in a Tertiary Referral Centre: Population, Characteristics, and Outcomes

    by Anthony James Goodings, Mila Pastrak, Sten Kajitani, Elaine Cunningham, Hannah O’brien, Catherine Weadick, Karie Dennehy

    Objectives This study aims to characterize the patient population referred to the specialist palliative care service with a diagnosis of gynecological cancer. It also assesses referral frequency and response time in order to understand palliative care utilization in an acute hospital setting. Methods A retrospective chart review and database analysis was performed to extract data on demographics, cancer diagnoses, and referral reasons for patients referred to the specialist palliative care service over three years. The study focuses on identifying patterns in the characteristics of the referred patient population. Results Analysis of 162 patients revealed a distribution across cancer subtypes: 62% ovarian, 22% endometrial, 12% cervical, and 4% vulvar. A notable finding was that the outcomes for patients with ovarian cancer were more likely to be discharged home with or without community care (61.8%) compared to those with endometrial cancer (41.0%). A rapid response to referrals was observed, with 70% reviewed within three days and 98% within a week. This highlights the service’s efficiency and the demographic and diagnostic profile of the patient served. Significance of Results This study gives insight into the demographic and diagnostic profiles of gynecological cancer patients referred for palliative care, alongside demonstrating rapid response to such referrals. Despite the rapid assessment times, the research importantly identifies differences in outcomes among different cancer subtypes, with a particular emphasis on the variance in discharge destinations. These findings reflect both patient preference and medical need, demonstrating the role of tailoring palliative care approaches to meet the individual needs and desires of this diverse patient population. Characterizing these patient profiles and outcomes serves as a crucial step in enhancing the effectiveness of palliative care strategies, with the ultimate goal of improving the quality of life for patients facing gynecological cancer.

  • Open Access

    Safety of Conventional and Pulsed Radiofrequency Lesions of the Dorsal Root Entry Zone Complex (DREZC) for Interventional Pain Management: A Systematic Review

    by Mila Pastrak, Ognjen Visnjevac, Tanja Visnjevac, Frederick Ma, Alaa Abd-Elsayed

    Study Design Systematic literature review. Objective The goal of this systematic review is to assess the clinical safety and potential complications of conventional and pulsed radiofrequency ablations targeting dorsal root entry zone complex (DREZC) components in the treatment of chronic pain. Background There is a growing popularity for the use of radiofrequency ablation (RFA) techniques targeting DREZC components by pain management physicians for an increasing variety of indications. To date, we lack a systematic review to describe the safety and the type of complications associated with these procedures. Methods This was a systematic literature review. This systematic search was limited to peer-reviewed literature using “radiofrequency ablation” as a search keyword using PubMed’s database for manuscripts published between inception and December 2020. Abstracts that involved the application of radiofrequency currents, of any modality, to DREZC components for the treatment of pain were included for full-text review. Search was limited to original data describing clinical outcomes following RFA performed for pain indications only, involving the DREZC components outlined above, in human subjects, and written in English. The primary outcomes were complications associated with conventional RFA and pulsed radiofrequency ablation (PRF). Complications were categorized as type 1 (persistent neurological deficits or other serious adverse events, defined as any event that resulted in permanent of prolonged injury; type 2 (transient neuritis or neurological deficits, or other non-neurological non-minor adverse event); type 3 (minor adverse events (e.g., headache, soreness, bruising, etc.). Results Of the 62 selected manuscripts totaling 3157 patients, there were zero serious adverse events or persistent neurological deficits reported. A total of 36 (1.14%) transient neurological deficits, cases of transient neuritis, or non-minor adverse events like uncomplicated pneumothorax were reported. A total of 113 (3.58%) minor adverse events were reported (bruising, transient site soreness, headache). Conclusions This systematic review indicates that the use of RFA lesion of the DREZC for interventional pain management is very safe. There were no serious adverse effects with a sizable sample of randomized controlled trial (RCT), prospective observational, and retrospective studies.


  • Open Access

    Radiofrequency Ablation of the Superior Cluneal Nerve: A Novel Minimally Invasive Approach Adopting Recent Anatomic and Neurosurgical Data

    by O. Visnjevac, M. Pastrak, F. Ma, T. Visnjevac, A. Abd-Elsayed

    Objectives/Introduction Superior cluneal neuralgia (SCN) is a distinct cause of lower back and/or leg pain related to pathology of the superior cluneal nerve (SCn). SCN has been termed pseudo-sciatica and is an overlooked differential diagnosis when patients are otherwise presenting with low back and/or radicular pain. Radiofrequency ablation (RFA) is commonly used for denervation of the medial branches of the dorsal root for facet joint syndrome for sacroiliac joint; however, RFA has not been described to ablate the SCn for SCN. Herein, we present a novel interventional minimally invasive approach using RFA of the SCn for SCN in a series of 46 patients. Methods Institutional review board approved retrospective chart review was used to collect data for all SCn RFA cases from January 1, 2018, to February 8, 2021. Fluoroscopically guided SCn ablations were performed for patients with a positive “iliac crest point sign,” reproductive of their back and leg pain during physical examination. Sensory stimulation was utilized to confirm RF cannula-probe placement adjacent to the SCn, and motor testing was used to confirm no distal motor response prior to monopolar RF ablation with a Halyard RF Generator (100 mm curved 22G 10 mm active tip RF cannulae). Charts were reviewed for time of analgesia follow-up, duration and degree of analgesia, improvements in patients’ functional capacity, and changes in medication. Results Data were reviewed for 51 patients who underwent Scn RFA, 5 of which were lost to follow-up. The remaining 46 patients consisted of 29 women and 17 men with a mean age of 59.4 years; 78.3% (n = 36) had ongoing relief at a mean of 92.1 days follow-up, ranging from 13 to 308 days, with a mean of 92.3% analgesia (SD 15.0%). At a mean of 111.2 days of follow-up, ranging from 42–201 days, 21.7% (n = 10) of patients reported that their pain had returned and had 95% analgesia during that time period (SD 6.7%); 41.3% (n = 19) reported improved activity/gait, 17.4% reported improved mood (n = 8), and 8.7% reported decreased medication use (n = 4). Five patients had minor complications including bruising (1), 2–3 days of soreness on site (2), myofascial pain (1), and quadratus lumborus muscle spasm relieved with trigger point injection (1). Conclusions This is the first report of both technique and outcomes for radiofrequency ablation of superior cluneal neuralgia. This series suggests that RFA of the SCn is a suitable intervention for the treatment of SCN; 21.7% of patients reported a mean of 95% analgesia for a mean duration of 111.2 days, and the remaining 78.3% of patients reported ongoing relief with a mean of 92.3% analgesia at last follow-up (mean 92.1 days). There were no serious adverse events.