Israel Medical Association Journal

Background: Advancements in artificial intelligence (AI) and natural language processing (NLP) have led to the development of language models such as ChatGPT. These models have the potential to transform healthcare and medical research. However, understanding their applications and limitations is essential.

Objectives: To present a view of ChatGPT research and to critically assess ChatGPT's role in medical writing and clinical environments.

Methods: We performed a literature review via the PubMed search engine from 20 November 2022, to 23 April 2023. The search terms included ChatGPT, OpenAI, and large language models. We included studies that focused on ChatGPT, explored its use or implications in medicine, and were original research articles. The selected studies were analyzed considering study design, NLP tasks, main findings, and limitations.

Results: Our study included 27 articles that examined ChatGPT's performance in various tasks and medical fields. These studies covered knowledge assessment, writing, and analysis tasks. While ChatGPT was found to be useful in tasks such as generating research ideas, aiding clinical reasoning, and streamlining workflows, limitations were also identified. These limitations included inaccuracies, inconsistencies, fictitious information, and limited knowledge, highlighting the need for further improvements.

Conclusions: The review underscores ChatGPT's potential in various medical applications. Yet, it also points to limitations that require careful human oversight and responsible use to improve patient care, education, and decision-making.

Background: Carbapenem-resistant Acinetobacter baumannii (CRAB) is an important cause of nosocomial infections. Active surveillance for CRAB carriage to identify and isolate colonized patients is used to reduce transmission.

Objectives: To assess the rate and risks of clinical infection among CRAB-carrier and non-carrier patients.

Methods: Hospitalized patients from whom CRAB screening-cultures were obtained between January and June 2018 were identified retrospectively. All CRAB-carriers were compared to a convenient sample of non-carriers and were followed to detect development of CRAB clinical infection during admission.

Results: We compared 115 CRAB carriers to 166 non-carriers. The median age in the study group was 76 years (IQR 71–87) vs. 65 years (55–79) in the non-carriers group (P < 0.001). Residence in a nursing facility, debilitated state, and admission to medical wards vs. intensive care units were more frequent among CRAB-carriers (P < 0.001). Mechanically ventilated patients included 51 CRAB carriers (44%) and 102 non-carriers (61%). Clinical infection developed in 49 patients (17%), primarily CRAB pneumonia. Of the CRAB-carriers and non-carriers, 26/115 (23%) and 23/166 (14%), respectively, developed a clinical infection (P = 0.05). One-third of the ventilated patients were infected. Debilitated state and antibiotic treatment during hospitalization were linked to higher infection rates (P = 0.01). Adjusted analysis showed that mechanical ventilation and CRAB colonization were strongly associated with clinical infection (P < 0.05).

Conclusion: The rate of CRAB infection among carriers was high. Mechanical ventilation and CRAB colonization were associated with CRAB clinical infection, primarily pneumonia

Background: Physical restraints are a common, albeit controversial, tool used in the acute care setting. 

Objectives: To determine the prevalence of physical restraint use in an acute care hospital. Secondary objectives were to determine whether physical restraints are used more commonly in night shifts, identify patient risk factors for physical restraint use, and establish if staff-to-patient ratio correlated with physical restraint use.

Methods: An observational cross-sectional study was conducted over 3 months in 2013 in the medical, surgical and intensive care units in a mid-sized general hospital. All the physically restrained patients in each observation were added to the registry. At each observation one department was selected for comparison and all non-restrained patients were added to the registry.

Results: The study population comprised 2163 patients. Seventy-six were restrained and 205 were included as case-controls. The prevalence of physical restraint use was 3.51% (95%CI = 2.79–4.37%). Physical restraint use was more common in night shifts than day shifts: 4.40% vs. 2.56% (P = 0.03). Male gender, dependency, invasive ventilation, invasive tubes (nasogastric tube or urine catheter), and bedsores were all significantly correlated with restraint use. Staff-to-patient ratios were not significantly correlated with use of physical restraints.

Conclusions: Physical restraints are relatively common in acute care wards. The use of physical restraints seems to correlate with certain patient characteristics but not with staff-to-patient ratios, and seems more common at night. 

 

Objective: To study the attitude of physicians to obesity by examining discharge letters of overweight patients with ischemic heart disease.

Methods: We used the HOLEM database for this analysis. The HOLEM project was designed to study the NCEP (National Cholesterol Education Program) guideline implementation among patients with IHD[1] at hospital discharge. We documented the recording of risk factors and treatment recommendations for IHD by reviewing the discharge letters of 2994 IHD patients admitted to four central hospitals in Israel between 1998 and 2000. A follow-up visit was held 6–8 weeks after discharge, at which time the diagnosis of IHD was verified, risk factor status was checked, height and weight were measured and drug treatment was reviewed.

Results: Mean body mass index was 28.3 kg/m² and 32% were obese (BMI[2] ³ 30 kg/m²). Only 39.6% of the obese patients and 65.8% of the morbidly obese patients (BMI ³ 40 kg/m²) had "obesity" noted in their discharging letters, and weight loss recommendation was written in only 15% of the obese patients' discharge letters. Acute episodes like acute myocardial infarction and unstable angina did not influence the notation of obesity, and only BMI and the number of additional risk factors were positively correlated with the notation of this risk factor.

Conclusions: Despite the importance of obesity, weight status was not noted and weight loss was not recommended in most of the discharge letters of obese IHD patients.

Objectives: To examine whether the treatment recommendations given to ischemic heart disease patients at hospital discharge are compatible with the guidelines of the Israeli Medical Societies and the U.S. National Cholesterol Education Program for coronary artery disease prevention; and to study the effects of brief educational sessions on the adherence of physicians with the guidelines.

Methods: We included consecutive IHD[1] patients admitted to four central hospitals in Israel between 1998 and 2000. The study was conducted in two phases. In phase 1, we reviewed discharge letters to document treatment recommendations given to each patient. In phase 2 we educated the practitioners by reviewing the Israeli Medical Societies and the NCEP[2] guidelines and the quality of their recommendations in phase 1, after which we reevaluated the discharge letters.

Results: The study included 2,994 patients: 627 in phase 1 and 2,367 in phase 2. Of the patients who needed cholesterol-lowering according to their low density lipoprotein levels, 37.4% were not prescribed such drugs at discharge (under-treatment group). This proportion was reduced by education to 26.6% (P < 0.001) in phase 2. Of the treated patients, 65.6% did not reach the target LDL[3] goal in phase 1 (under-dosage group) as compared to 60.2% in phase 2 (P = 0.23). In phase 2 there was an increase in the percent of patients reaching LDL levels <130 mg/day (69.3% vs. 63.8% of patients prescribed medication, P = 0.01), but the percent of patients reaching LDL levels <100 was not different in phase 2 after adjusting for age and gender (the odds ratio for reaching target LDL was 1.16, with 95% confidence interval of 0.95–1.43).

Conclusions: Physician recommendations to IHD patients discharged from hospital were suboptimal. We documented a high proportion of under-treated and under-dosage patients. Brief educational sessions have a beneficial effect on the usage of statins; however, additional effort in guideline implementations is needed.