ChatGPT-3.5 Versus Google Bard: Which Large Language Model Responds Best to Commonly Asked Pregnancy Questions?

Large language models (LLM) have been widely used to provide information in many fields, including obstetrics and gynecology. Which model performs best in providing answers to commonly asked pregnancy questions is unknown. A qualitative analysis of Chat Generative Pre-Training Transformer Version 3.5 (ChatGPT-3.5) (OpenAI, Inc., San Francisco, California, United States) and Bard, recently renamed Google Gemini (Google LLC, Mountain View, California, United States), was performed in August of 2023. Each LLM was queried on 12 commonly asked pregnancy questions and asked for their references. Review and grading of the responses and references for both LLMs were performed by the co-authors individually and then as a group to formulate a consensus. Query responses were graded as "acceptable" or "not acceptable" based on correctness and completeness in comparison to American College of Obstetricians and Gynecologists (ACOG) publications, PubMed-indexed evidence, and clinical experience. References were classified as "verified," "broken," "irrelevant," "non-existent," and "no references." Grades of "acceptable" were given to 58% of ChatGPT-3.5 responses (seven out of 12) and 83% of Bard responses (10 out of 12). In regard to references, ChatGPT-3.5 had reference issues in 100% of its references, and Bard had discrepancies in 8% of its references (one out of 12). When comparing ChatGPT-3.5 responses between May 2023 and August 2023, a change in "acceptable" responses was noted: 50% versus 58%, respectively. Bard answered more questions correctly than ChatGPT-3.5 when queried on a small sample of commonly asked pregnancy questions. ChatGPT-3.5 performed poorly in terms of reference verification. The overall performance of ChatGPT-3.5 remained stable over time, with approximately one-half of responses being "acceptable" in both May and August of 2023. Both LLMs need further evaluation and vetting before being accepted as accurate and reliable sources of information for pregnant women.

Painful bruising: Gynecology, hematology, or just pill bias? A case report.

A 23-year-old woman, G0, presented to the emergency department with painful bruising of the legs shortly after starting an oral contraceptive pill. The presumed diagnosis was pill-induced ecchymosis, and she was instructed to discontinue the medication. Her bruising resolved. However, the working diagnosis was later questioned as the patient had used other oral contraceptive pills in the past without any adverse reaction. In addition, there is robust literature associating these medications with thrombosis, not bruising. The patient later disclosed that she had concomitantly started an oral hair supplement along with her oral contraceptive pill. Analysis of the supplement contents revealed that it contained extract of Aesculus hippocastanum, a herbal anticoagulant, making this a much more plausible explanation for the ecchymosis. She then resumed the original oral contraceptive pill alone without any reaction. The case highlights how cognitive bias resulted in a misdiagnosis. Specifically, this case introduces the concept of pill bias, as the patient's unexplained bruising was presumed to be a result of her use of an oral contraceptive despite the lack of evidence to support this claim. This bias has the potential to impact clinical decision-making and lead to clinical errors.

Monochorionic Diamniotic Vasa Previa Pregnancy: A Medical Student Perspective.

Monochorionic diamniotic twins and vasa previa are uncommon. We present a case that was followed from ultrasound diagnosis to delivery.

Low-grade fever, erythematous rash in pregnant woman · Dx?

A 31-year-old woman presented to her obstetrician's office at 16 weeks' gestation with a 2-day history of low-grade fever and an erythematous rash measuring 1 x 4 cm on her right groin. She had a medical history of a penicillin allergy (urticarial) and her outdoor activities included gardening and picnicking. What's your diagnosis?

Missing IUD Despite Threads at the Cervix.

Today, the intrauterine device (IUD) is by far the most popular form of long term reversible contraception in the world. Side effects from the IUD are minimal and complications are rare. Uterine perforation and migration of the IUD outside the uterine cavity are the most serious complications. Physician visualization and/or the patient feeling retrieval threads at the cervical os are confirmation that the IUD has not been expelled or migrated. We present a case of a perforated, intraperitoneal IUD with threads noted at the cervical os. Office removal was not possible using gentle traction on the threads. Multiple imaging and endoscopic modalities were used to try and locate the IUD including pelvic ultrasound, diagnostic hysteroscopy, cystoscopy, and pelvic magnetic resonance imaging (MRI). The studies gave conflicting results on location of the IUD. Ultimately, the missing IUD was removed via laparoscopy.

Use of an intubating stylet as a guide to complete uterine curettage complicated by uterine perforation.

Completion of uterine curettage may be challenging following uterine perforation even under sonographic and laparoscopic monitoring. This report illustrates the use of a flexible intubating stylet as a guide to place the suction curette into the uterine cavity when sonography and laparoscopy alone are not successful. Use of a malleable instrument such as an intubating stylet as a guide should be considered an option when insertion of the suction curette into the uterine cavity is complicated by anatomic variation and uterine perforation.

Delivery route following elective induction of labor at term: analysis of 807 patients.

BACKGROUND: The purpose of this study is to compare mode of delivery for both nulliparous and multiparous women at term that underwent elective induction of labor to those who arrived in spontaneous labor. METHODS: Medical records of 807 deliveries were reviewed. There were 566 labor patients and 241 elective induction patients. RESULTS: Women who underwent elective induction of labor were more likely to undergo cesarean delivery compared to those women who arrived in spontaneous labor (41.1% versus 9.9%, P = 0.001). This was true for both nulliparous women (49% versus 31%, P < 0.0001), and multiparous women (22.7% versus 1.6%, P < 0.0001). The rate of operative vaginal delivery was also increased in the elective induction of labor group (8.4% versus 3.6%, P < 0.0001). Operative vaginal delivery was statistically significant in multiparous women (21% versus 4.1%, P < 0.0001), but not in nulliparous women (10.1% versus 9.8%, P = NS). CONCLUSION: Elective induction of labor at term is associated with an increased risk of cesarean section in both nulliparous and multiparous women. There is also an increased risk of an operative vaginal delivery in multiparous women who underwent elective induction of labor.

How do patients use email to communicate with their doctor?

OBJECTIVE: To evaluate unsolicited email sent from patients to a general obstetrician-gynecologist. STUDY DESIGN: A retrospective review was performed of 60 unsolicited emails sent to an obstetrician-gynecologist. Time sent, content, and physician response were evaluated. RESULTS: Most emails were sent during office hours. Fifty-four percent of the emails had content related to gynecologic issues. Twenty-seven percent of the emails had content related to obstetrical concerns. Nineteen percent of emails were nonclinical in nature. Fourteen percent of emails required a follow-up phone call. Twenty-nine percent of emails necessitated an office visit. CONCLUSION: Patients did not use email for emergency communication. The participating physician was pleased with the ease of responding to patients, especially while away from the office.

Clinical Presentation and Conservative Management of Tympanic Membrane Perforation during Intrapartum Valsalva Maneuver.

Background. Tympanic membrane perforation may occur when ear pressures are excessive, including valsalva maneuver associated with active labor and vaginal delivery. A pressure differential across the eardrum of about 5 psi can cause rupture; the increased intraabdominal pressure spikes repeatedly manifested by "pushing" during second-stage labor easily approach (and may exceed) this level. Material and Method. We describe a healthy 21-year old nulliparous patient admitted in active labor at 39-weeks' gestational age. Results. Blood appeared asymptomatically in the left ear canal at delivery during active, closed-glottis pushing. Otoscopic examination confirmed perforation of the left tympanic membrane. Complete resolution of the eardrum rupture was noted at postpartum check-up six weeks later. Conclusion. While the precise incidence of intrapartum tympanic membrane rupture is not known, it may be unrecognized without gross blood in the ear canal or subjective hearing loss following delivery. Only one prior published report on tympanic membrane perforation during delivery currently appears in the medical literature; this is the first English language description of the event. Since a vigorous and repetitive valsalva effort is common in normal vaginal delivery, clinicians should be aware of the potential for otic complications associated with the increased intraabdominal pressure characteristic of this technique.

Clinical and patient estimation of fetal weight vs. ultrasound estimation.

OBJECTIVE: To compare clinical and patient estimation of fetal weight to ultrasound estimation. STUDY DESIGN: Prospective study of clinical, patient and ultrasound estimation of fetal weight at term. RESULTS: A total of 200 pregnant women participated. There was no statistically significant difference between clinical and sonographic estimates of fetal weight: of the estimates, 64.0% were within 10% of the actual birth weight vs. 62.5% (P > .2). There was no statistically significant difference between patient and sonographic estimates of fetal weight: of the estimates, 53.5% were within 10% of the actual birth weight vs. 62.5%, respectively (P < .1). Senior resident clinical and sonographic estimates of fetal weight were superior to junior resident estimates: 75.2% of clinical estimates were within 10% of the actual birth weight vs. 59.2% (P < .03), and 73.1% of sonographic estimates were within 10% of the actual birth weight vs. 58.3% (P < .05). Nulliparous and multiparous patients were equally accurate in estimating fetal weight: 48.1% of estimates were within 10% of birth weight vs. 57.4% (P > .2). CONCLUSION: Sonographic estimation of fetal weight offers no advantage over clinical or patient estimation of fetal weight at term. Senior resident clinical and sonographic estimates are superior to junior resident estimates. Parity has no effect on patient accuracy in estimating fetal weight.