Chatting Beyond ChatGPT: Advancing Equity Through AI-Driven Language Interpretation.

Medical interpretation is an underutilized resource, despite its legal mandate and proven efficacy in improving health outcomes for populations with low English proficiency. This disconnect can often be attributed to the costs and wait-times associated with traditional means of interpretation, making the service inaccessible and burdensome. Technology has improved access to translation through phone and video interpretation; with the acceleration of artificial intelligence (AI) large language models, we have an opportunity to further improve interpreter access through real-time, automated translation. The impetus to utilize this burgeoning tool for improved health equity must be combined with a critical view of the safety, privacy, and clinical decision-making risks involved. Physicians must be active participants and collaborators in both the mobilization of AI tools to improve clinical care and the development of regulations to mitigate harm.

Uroperitoneum as an Atypical Source of Ascites: A Case Report.

Uroperitoneum secondary to spontaneous bladder rupture is a rare cause of ascites associated with significant morbidity and mortality. It can be difficult to detect and is often initially mistaken for other, more common etiologies. We present the case of a 56-year-old female with a history of cervical cancer treated with chemotherapy and radiation, radiation proctitis, and diabetes mellitus who presented with subacute onset abdominal pain and distension, urinary retention, and nausea. She had been diagnosed with cervical squamous cell cancer 12 years prior to presentation and was successfully treated with two months of chemotherapy and radiation, and a presumed recurrence five years later was treated to remission with chemotherapy. The golden-yellow appearance of her ascitic fluid during diagnostic paracentesis raised suspicion for urinary ascites that was confirmed by an elevated ascites-to-serum creatinine ratio and computed tomography (CT) cystography. Subsequent CT cystogram demonstrated leakage of contrast from the bladder with a 0.5 cm irregularity noted at the bladder dome, potentially representing the site of extravasation. A Foley catheter was placed at the time of admission with an immediate output of 1 L of fluid. Subsequently, her abdominal distension significantly improved, and her creatinine began to downtrend. Gynecologic oncology and urology were consulted and determined that she was not a candidate for surgical intervention given the significance of her bladder scarring. Positron emission tomography (PET)/CT was performed and revealed no active cancer. At the time of discharge, she had no episodes of emesis. Additionally, her creatinine had fallen to 1.0 mg/dl. She was discharged with a Foley catheter with plans to follow up with outpatient urology. While relatively uncommon, uroperitoneum should be suspected in patients presenting with new-onset ascites who have risk factors for spontaneous bladder rupture such as pelvic irradiation. Uroperitoneum has a significant rate of mortality and morbidity. Ascites urea and creatinine studies, followed by a CT cystogram if these studies are abnormal, should be performed in any patient with risk factors for uroperitoneum. Patients should be managed with the placement of a Foley catheter and urology consultation for surgical evaluation.

Characterizing trauma encounters among patients with language other than English preference at a level 1 trauma center's emergency department.

Objectives: Injury-related visits constitute a sizeable portion of emergency department (ED) visits in the United States. Individuals with language other than English (LOE) preference face barriers to healthcare and visits for traumatic injury may be the first point of contact with the healthcare system. Yet, the prevalence of traumatic injuries in this population is relatively unknown. Our objective was to characterize the prevalence and purpose of trauma encounters, and healthcare utilization, among a LOE cohort. Methods: We conducted a retrospective chart review of LOE patients who presented for a trauma encounter at a level 1 trauma and emergency care center between January 1, 2019 and December 31, 2021. LOE participants were identified by utilization of video-based language interpretive services. Variables evaluated included injury patterns and primary and subspeciality healthcare utilization. Quantitative analysis of categorical and continuous variables was performed. Results: A total of 429 patients were included. Most patients presented for one trauma encounter and the majority spoke Spanish. The most common causes of injury were motor vehicle collisions (MVCs) (28.5%, n = 129), ground-level falls (15.9%, n = 72), and falls from heights (14.2%, n = 64). Occupational injuries made up 27.2% of trauma encounters (n = 123) and only 12.6% (n = 54) of patients had a primary care visit. Conclusion: Our findings highlight the need for increased research and attention to all causes of injury, especially MVCs and occupational injuries, among those with LOE preference. Results reaffirm an underutilization of healthcare among this population and the opportunity for trauma encounters as points of access to care.

Health Status and Healthcare Utilization Patterns of Emergency Department Patients Who Prefer a Language Other Than English.

Immigrants, nearly half of whom prefer a language other than English (LOE), face structural barriers to healthcare. This subgroup is believed to be at increased risk for reduced access to quality healthcare, yet few studies have examined the health needs and utilization patterns of LOE-preferring patients who seek care in the Emergency Department (ED). Given that the ED is often an entry point to the health system, we sought to characterize the health patterns of this population in an urban ED setting. We conducted a retrospective chart review of the electronic medical records of 1,566 patients who utilized interpreter services in the ED or Urgent Care) at an urban safety net hospital. We found that LOE-preferring patients had high levels of chronic disease. We also found that the majority of these patients had not seen a primary care provider (PCP) within the study period. PCP visits were positively associated with ED utilization suggesting that those without a PCP are less likely to receive ED care. These findings point to a need for greater policy and community health solutions addressing the high burden of chronic disease and underutilization of healthcare for those with LOE preferences.

Availability of population-level data sources for tracking the incidence of deaths and injuries from road traffic crashes in low-income and middle-income countries.

INTRODUCTION: Tracking progress towards Sustainable Development Goal (SDG) 3·6 of reducing traffic deaths and serious injuries poses a measurement challenge in most low-income and middle-income countries (LMICs) due to large discrepancies between reported official statistics and estimates from global health measurement studies. We assess the extent to which national population censuses and health surveys can fill the information gaps. METHODS: We reviewed questionnaires for nationally representative surveys and censuses conducted since 2000 in LMICs. We identified sources that provide estimates of household ownership of vehicles, incidence of traffic deaths and non-fatal injuries, and prevalence of disability. RESULTS: We identified 802 data sources from 132 LMICs. Sub-Saharan African countries accounted for 43% of all measurements. The number of measurements since 2000 was high, with 97% of the current global LMIC population having at least one measurement for vehicle ownership, 77% for deaths, 90% for non-fatal injuries and 50% for disability due to traffic injuries. Recent data (since 2010) on traffic injuries were available from far fewer countries (deaths: 21 countries; non-fatal injuries: 62 and disability: 12). However, there were many more countries with recent data on less-specific questions about unintentional or all injuries (deaths: 41 countries, non-fatal: 87, disability: 32). CONCLUSION: Traffic injuries are substantially underreported in official statistics of most LMICs. National surveys and censuses provide a viable alternative information source, but despite a large increase in their use to monitor SDGs, traffic injury measurements have not increased. We show that relatively small modifications and additions to questions in forthcoming surveys can provide countries with a way to benchmark their existing surveillance systems and result in a substantial increase in data for tracking road traffic injuries globally.