Clinical Information Processing Research Articles

Development of a liver disease-specific large language model chat interface using retrieval-augmented generation.

Large language models (LLMs) have significant capabilities in clinical information processing tasks. Commercially available LLMs, however, are not optimized for clinical uses and are prone to generating hallucinatory information. Retrieval-augmented generation (RAG) is an enterprise architecture that allows the embedding of customized data into LLMs. This approach "specializes" the LLMs and is thought to reduce hallucinations. We developed "LiVersa," a liver disease-specific LLM, by using our institution's protected health information-complaint text embedding and LLM platform, "Versa." We conducted RAG on 30 publicly available American Association for the Study of Liver Diseases guidance documents to be incorporated into LiVersa. We evaluated LiVersa's performance by conducting 2 rounds of testing. First, we compared LiVersa's outputs versus those of trainees from a previously published knowledge assessment. LiVersa answered all 10 questions correctly. Second, we asked 15 hepatologists to evaluate the outputs of 10 hepatology topic questions generated by LiVersa, OpenAI's ChatGPT 4, and Meta's Large Language Model Meta AI 2. LiVersa's outputs were more accurate but were rated less comprehensive and safe compared to those of ChatGPT 4. We evaluated LiVersa's performance by conducting 2 rounds of testing. First, we compared LiVersa's outputs versus those of trainees from a previously published knowledge assessment. LiVersa answered all 10 questions correctly. Second, we asked 15 hepatologists to evaluate the outputs of 10 hepatology topic questions generated by LiVersa, OpenAI's ChatGPT 4, and Meta's Large Language Model Meta AI 2. LiVersa's outputs were more accurate but were rated less comprehensive and safe compared to those of ChatGPT 4. In this demonstration, we built disease-specific and protected health information-compliant LLMs using RAG. While LiVersa demonstrated higher accuracy in answering questions related to hepatology, there were some deficiencies due to limitations set by the number of documents used for RAG. LiVersa will likely require further refinement before potential live deployment. The LiVersa prototype, however, is a proof of concept for utilizing RAG to customize LLMs for clinical use cases.

GastroBot: a Chinese gastrointestinal disease chatbot based on the retrieval-augmented generation.

Large Language Models (LLMs) play a crucial role in clinical information processing, showcasing robust generalization across diverse language tasks. However, existing LLMs, despite their significance, lack optimization for clinical applications, presenting challenges in terms of illusions and interpretability. The Retrieval-Augmented Generation (RAG) model addresses these issues by providing sources for answer generation, thereby reducing errors. This study explores the application of RAG technology in clinical gastroenterology to enhance knowledge generation on gastrointestinal diseases. We fine-tuned the embedding model using a corpus consisting of 25 guidelines on gastrointestinal diseases. The fine-tuned model exhibited an 18% improvement in hit rate compared to its base model, gte-base-zh. Moreover, it outperformed OpenAI's Embedding model by 20%. Employing the RAG framework with the llama-index, we developed a Chinese gastroenterology chatbot named "GastroBot," which significantly improves answer accuracy and contextual relevance, minimizing errors and the risk of disseminating misleading information. When evaluating GastroBot using the RAGAS framework, we observed a context recall rate of 95%. The faithfulness to the source, stands at 93.73%. The relevance of answers exhibits a strong correlation, reaching 92.28%. These findings highlight the effectiveness of GastroBot in providing accurate and contextually relevant information about gastrointestinal diseases. During manual assessment of GastroBot, in comparison with other models, our GastroBot model delivers a substantial amount of valuable knowledge while ensuring the completeness and consistency of the results. Research findings suggest that incorporating the RAG method into clinical gastroenterology can enhance the accuracy and reliability of large language models. Serving as a practical implementation of this method, GastroBot has demonstrated significant enhancements in contextual comprehension and response quality. Continued exploration and refinement of the model are poised to drive forward clinical information processing and decision support in the gastroenterology field.

Structural network disruption markers explain disability in multiple sclerosis

ObjectiveTo evaluate whether structural brain network metrics correlate better with clinical impairment and information processing speed in multiple sclerosis (MS) beyond atrophy measures and white matter lesions.MethodsThis cross-sectional study included...

Information processing in electronic medical records: A survey validation

The purpose of this study was to validate the Clinical Information Processing Instrument. This instrument attempts to examine information processing in electronic medical records (EMRs). We drew upon the theory of swift and even flow to guide survey development and construction. We used a mixed-methods approach to gather data from registered nurses. Nurses were invited to participate in focus groups, an expert panel, and the survey validation process. A semi-structured questionnaire addressed the following themes: method of communication, quality of information, and usability of the system. We conducted a confirmatory factor analysis using structural equation modelling. The Kaiser-Meyer-Olkin measure was greater than 0.7 (0.90), and the Bartlett's test of sphericity was significant (X2 =1519.03, df=105, P<0.001). The proposed structural equation model was analysed and revised to a final model that was statistically significant. The final survey, Clinical Information Processing Instrument, contained 18 Likert scale questions that supported the tenets of the theory of swift and even flow. The nurses perceived EMRs as efficient for medication management, time management, and communication. The Clinical Information Processing Instrument is a validated survey tool that assesses information flow in EMRs. The Clinical Information Processing Instrument was validated as an approach to analyse the utility of EMR in disseminating information among clinical staff. To increase the utility and meaningful use of EMR systems, it is important to consider factors that affect the distribution of information among clinicians.

Blink or think: can further reflection improve initial diagnostic impressions?

Experienced clinicians derive many diagnoses intuitively, because most new problems they see closely resemble problems they've seen before. The majority of these diagnoses, but not all, will be correct. This study determined whether further reflection regarding initial diagnoses improves diagnostic accuracy during a high-stakes board exam, a model for studying clinical decision making. Keystroke response data were used from 500 residents who took the 2010 American Board of Internal Medicine (ABIM) Internal Medicine Certification Examination. Data included time to initial response on each question, whether the answer was correct, and whether or not the resident changed her or his initial response. The focus was on 80 diagnosis questions that comprised realistic clinical vignettes with multiple-choice single-best answers. Cognitive skill (ability) was measured using overall exam scores. Case complexity was determined using item difficulty (proportion of examinees that correctly answered the question). A hierarchical generalized linear model was used to assess the relationship between time spent on initial responses and the probability of correctly answering the questions. On average, residents changed their responses on 12% of all diagnosis questions (or 9.6 questions out of 80). Changing an answer from incorrect to correct was almost twice as likely as changing an answer from correct to incorrect. The relationship between response time and accuracy was complex. Further reflection appears to be beneficial to diagnostic accuracy, especially for more complex cases.

Applying a generic function-based topical relevance typology to structure clinical questions and answers

This study investigates the manifestation and utility of a generic function-based topical relevance typology adapted to the subject domain of clinical medicine. By specifying the functional role of a given piece of relevant information in the overall structure of a topic, the proposed typology provides a generic framework for integrating different pieces of clinical evidence and a multifaceted view of a clinical problem. In medical problem solving structured knowledge plays a key role. The typology provides the conceptual basis for integrating and structuring knowledge; it incorporates and goes beyond existing clinical schemes (such as PICO and illness script) and offers extra assistance for physicians as well as lay users (such as patients and caregivers) to manage the vast amount of diversified evidence, to maintain a structured view of the patient problem at hand, and ultimately to make well-grounded clinical choices. Developed as a generic topical framework across topics and domains, the typology proved useful for clinical medicine once extended with domain-specific definitions and relationships. This article reports the findings of using the adapted and extended typology in the analysis of 26 clinical questions and their evidence-based answers. The article concludes with potential applications of the typology to improve clinical information seeking, organizing, and processing.

Personal Health System architecture for stress monitoring and support to clinical decisions

Developments in computational techniques including clinical decision support systems, information processing, wireless communication and data mining hold new premises in Personal Health Systems. Pervasive Healthcare system architecture finds today an effective application and represents in perspective a real technological breakthrough promoting a paradigm shift from diagnosis and treatment of patients based on symptoms to diagnosis and treatment based on risk assessment. Such architectures must be able to collect and manage a large quantity of data supporting the physicians in their decision process through a continuous pervasive remote monitoring model aimed to enhance the understanding of the dynamic disease evolution and personal risk. In this work an automatic simple, compact, wireless, personalized and cost efficient pervasive architecture for the evaluation of the stress state of individual subjects suitable for prolonged stress monitoring during normal activity is described. A novel integrated processing approach based on an autoregressive model, artificial neural networks and fuzzy logic modeling allows stress conditions to be automatically identified with a mobile setting analysing features of the electrocardiographic signals and human motion. The performances of the reported architecture were assessed in terms of classification of stress conditions.

Tumour node metastasis staging of bladder cancer: prognosis versus pitfalls

The WHO classification of urothelial cancer in 2004 has made changes based on the insights of molecular genetics, indicating bladder cancer with entities that are genetically stable versus those that are genetically instable. Seen as work in progress, the need of further validation is obvious. Clinical studies based on solid histological diagnosis are as necessary as the definition of more molecular features of bladder cancer. Solid histological diagnosis includes sufficient clinical information and adequate tissue processing. This combined with molecular data will lead to a more clear-cut distinction between benign and malignant and possibly to another change in terminology with higher concordance to other epithelial tumours. Whereas the identification of FGFR3 mutations has led to a better distinction of at least two pathways of urothelial carcinogenesis, additional multiparametric approaches may help improve the still inadequate search for urine and blood markers indicative of bladder cancer and/or its progression. Proteomic profiling, sets of epigenetic markers, and micro RNAs will be given as examples. Recent data mainly support the concept of the WHO 2004 classification of bladder cancer. We are optimistic that an even more clear-cut distinction between benign recurring, nonprogressing tumours and more aggressive tumours will enable us to focus and limit chemotherapy.

High-dose glycine impairs the prepulse inhibition measure of sensorimotor gating in humans

An impaired capacity to filter or 'gate' sensory information is a core deficit in cognitive function associated with schizophrenia. These deficits have been linked in part to N-methyl-d-aspartate (NMDA) receptor dysfunction. An association between high levels of glycine, a positive allosteric modulator of the NMDA receptor, and sensorimotor gating impairments (i.e. prepulse inhibition (PPI) deficit) have been reported in animal models of schizophrenia as well as patients with schizophrenia. This study examined the acute effects of modulating the glycine site of the NMDA receptor (with high-dose glycine) on sensory gating as measured by PPI. Sixteen healthy male subjects (final sample size of 12) participated in a double-blind, placebo-controlled, crossover design in which each subject was tested under two acute treatment conditions separated by at least a 5-day washout period; placebo and 0.8 g/kg glycine. PPI was recorded 45 min post treatment using electromyography of the eye-blink response. Relative to placebo, high-dose glycine significantly impaired sensorimotor gating as demonstrated by a decrease in PPI (t(11) = -2.983, p < 0.05). Administration of a high dose of glycine is associated with impairments in PPI supporting earlier observations in animals and patients with schizophrenia. This result, when taken together with findings in patients, suggests that high synaptic levels of glycine may have some clinically relevant detrimental effects and suggests a potential dissociation of clinical symptomatology and sensory information processing as a function of NMDA receptor modulation in schizophrenia.

Age-associated differences in cognitive performance in older community dwelling schizophrenia patients: Differential sensitivity of clinical neuropsychological and experimental information processing tests

Cognitive dysfunction is a common feature of schizophrenia and deficits are present before the onset of psychosis, and are moderate to severe by the time of the first episode. Controversy exists over the course of cognitive dysfunction after the first episode. This study examined age-associated differences in performance on clinical neuropsychological (NP) and information processing tasks in a sample of geriatric community living schizophrenia patients ( n = 172). Compared to healthy control subjects ( n = 70), people with schizophrenia did not differ on NP tests across age groups but showed evidence for age-associated cognitive worsening on the more complex components of an information-processing test. Age-related changes in cognitive function in schizophrenia may be a function of both the course of illness and the processing demands of the cognitive measure of interest. Tests with fixed difficulty, such as clinical NP tests, may differ in their sensitivity from tests for which parametric difficulty manipulations can be performed.

The Role of Physician Characteristics in Clinical Trial Acceptance: Testing Pathways of Influence

ABSTRACT Eight videotaped vignettes were developed that assessed the effects of three physician-related experimental variables (in a 2 × 2 × 2 factorial design) on clinical trial (CT) knowledge, video knowledge, information processing, CT beliefs, affective evaluations (attitudes), and CT acceptance. It was hypothesized that the physician variables (community versus academic-based affiliation, enthusiastic versus neutral presentation of the trial, and new versus previous relationship with the patient) would serve as communication cues that would interrupt message processing, leading to lower knowledge gain but more positive beliefs, attitudes, and CT acceptance. A total of 262 women (161 survivors and 101 controls) participated in the study. The manipulated variables primarily influenced the intermediary variables of post-test CT beliefs and satisfaction with information rather than knowledge or information processing. Multiple regression results indicated that CT acceptance was associated with positive post-CT beliefs, a lower level of information processing, satisfaction with information, and control status. Based on these results, CT acceptance does not appear to be based on a rational decision-making model; this has implications for both the ethics of informed consent and research conceptual models.

Developing Clinical Reasoning Strategies

Incorporating decision support research and technology, such as computerized "best practice" clinical pathways, into patient care management is recommended or decreasing cognitive demands on nurses. Criticized for not representing "real-time" patient care, these standardized approaches may be used in concert with strategies for developing nurses' own clinical reasoning. Real-time clinical information processing activities are offered for staff educators to augment the development of clinical reasoning for both experienced and less-experienced nurses' decision making.

Wireless access to a pharmaceutical database: a demonstrator for data driven Wireless Application Protocol (WAP) applications in medical information processing.

BackgroundThe Wireless Application Protocol technology implemented in newer mobile phones has built-in facilities for handling much of the information processing needed in clinical work.ObjectivesTo test a practical approach we ported a relational database of the Danish pharmaceutical catalogue to Wireless Application Protocol using open source freeware at all steps.MethodsWe used Apache 1.3 web software on a Linux server. Data containing the Danish pharmaceutical catalogue were imported from an ASCII file into a MySQL 3.22.32 database using a Practical Extraction and Report Language script for easy update of the database. Data were distributed in 35 interrelated tables. Each pharmaceutical brand name was given its own card with links to general information about the drug, active substances, contraindications etc. Access was available through 1) browsing therapeutic groups and 2) searching for a brand name. The database interface was programmed in the server-side scripting language PHP3.ResultsA free, open source Wireless Application Protocol gateway to a pharmaceutical catalogue was established to allow dial-in access independent of commercial Wireless Application Protocol service providers. The application was tested on the Nokia 7110 and Ericsson R320s cellular phones.ConclusionsWe have demonstrated that Wireless Application Protocol-based access to a dynamic clinical database can be established using open source freeware. The project opens perspectives for a further integration of Wireless Application Protocol phone functions in clinical information processing: Global System for Mobile communication telephony for bilateral communication, asynchronous unilateral communication via e-mail and Short Message Service, built-in calculator, calendar, personal organizer, phone number catalogue and Dictaphone function via answering machine technology. An independent Wireless Application Protocol gateway may be placed within hospital firewalls, which may be an advantage with respect to security. However, if Wireless Application Protocol phones are to become effective tools for physicians, special attention must be paid to the limitations of the devices. Input tools of Wireless Application Protocol phones should be improved, for instance by increased use of speech control.

Conceptual models and integrative therapy: anorexia nervosa as a prototype.

To demonstrate how an integrative approach to psychiatric treatment can confer specific advantages. The impact of conceptual models on clinical information processing and therapeutic intervention, and the potential benefits resulting from the incorporation of multiple treatment modalities are outlined. These principles are applied using anorexia nervosa (AN) as a prototypical disorder. A brief, structured review of the broad range of models which have been applied to AN illustrates the impact of conceptualization on outcome. The opportunity for positive integrative effects in the treatment of AN are demonstrated. They are seen to be related to 6 factors: flexibility, patient-treatment matching, stage-treatment matching, facilitation, complementarity, and synergism. Integrative methods can address some limitations in clinical information processing associated with the exclusive application of a single conceptual model. Positive integrative effects are demonstrated in the context of AN; however, the approach is recommended for more general application.

The stigma of mental illness on children of color.

Children and adolescents who have been designated as mentally ill, mentally retarded, juvenile delinquent, poor, homeless, or sexually abused at some point in their lives experience a great deal of difficulty in removing such labels. Society has persisted in using these labels even when the designated person is engaged in many of the things taken for granted among "normal" persons of their age and culture. The author discusses the stigma that results from unreliable, invalid, unstandardized identification of psychiatric disorders, particularly among children from ethnic minority groups. Research is needed that improves clinical information processing, particularly among the multicultural population of the United States. For too long majority-based norms have been the focus of research, with little consideration given to the creative adaptations that have occurred among ethnic minority groups.

A prototype radiation oncology clinical information processing system on a personal computer (Network)

Software has been developed that facilitates prospective computer entry of the clinical contents of a radiation oncology chart in a way that assists in retrospective analyses. This software, which has been operational since 1988, now contains detailed clinical information on over 500 patients treated since that time, and limited information on over 3000 children treated since 1970. The system has been programmed using the FoxPro database management system, and may be run on an IBM® compatible or Macintosh® personal computer either alone or in a networked environment. Information is transcribed as routine patient dictations that require no extra personnel for entry. Standard printouts are the conventional notes compatible with the existing paper chart. Exportation of patient records is facilitated by means of built-in facsimile. The software has been designed to generate detailed reports automatically, including the statuses of selected patients or groups, lists of those who are behind in follow-up, and reports of specific events (e.g., mucositis) that occur before, during, or after therapy using free-text searches. Complete reconstruction of the patient's clinical radiation oncology chart is also included. Other abilities include calculation of Kaplan-Meier survival, relapse-free survival, and local control rates, statistical comparison between survival curves, and radiation dose-response. Exportation of data to statistical and graphics packages is also possible. For pediatric patients, a program to predict stature loss has been incorporated. Use of this system can enhance one's ability not only to follow patients but also to monitor and report their outcome continually. With the need to report institutional experiences and increasing demands to monitor quality assurance, such a system can be of great benefit.

Conservation and formal thought disorder in schizophrenic and schizotypal children

AbstractWe studied the conservation scores of 21 schizophrenic and schizotypal children (aged 6;7–12;5 years) and 21 yoked normal mental age matches and their relationship to the Kiddie Formal Thought Disorder Scale measures. Like the normal children, the schizophrenic/schizotypal subjects were able to recognize the invariance of two dimensional space, number, and substance, but not of weight. Unlike the normal children, the schizophrenic/schizotypal children were poor conservers of continuous and discontinuous matter, and this was related to their illogical thinking scores. Age and IQ accounted for the conservation competence of the normal children. The conservation skills of the patients, however, were associated with the severity of their formal thought disorder scores, not with their age or IQ. The developmental approach employed in this study enabled further clarification of the cognitive parameters of formal thought disorder in middle childhood. We discuss the developmental, clinical, and possible information processing implications of these findings.

Connecting Clinical Information Processing With Computer Support

Clinically oriented computer systems require compatibility with the information processing that typifies clinical work. This paper summarizes some of the research that can be applied to such processing and uses the summary to suggest likely errors. Alternative means to correct for such errors are also summmarized leading to an emphasis on decision support systems that integrate human and computer information processing strengths. An example concerning long-term mental illness is used for illustration.

Computational and noncomputational clinical information processing by computer.

Computers have been used in their computational as well as noncomputational modes for the scientific study of the clinical decision making. As computational tools, computers serve as storage devices for hospital file data so that these data can be analyzed statistically for a large variety of epidemiological and diagnostic purposes. As noncomputational devices, computers are functioning as knowledgable medical teachers that interface with students who can query the system. In this mode, computers are sometimes used as consultants to clinicians who need information on technical topics. Noncomputational computers are also used to simulate the inferences of clinicians. Byproducts of these simulations are computer programs that become formal statements or theories of the decision processes under investigation.

Clinical Information Processing by Computer

Clinical Information Processing by Computer Get access Clinical Information Processing by Computer. By Kleinmuntz Benjamin. xii and 399 paperback. pp., $5.95. New York: Holt, Rinehart, and Winston, Inc., 1969. Clement J. McDonald Clement J. McDonald Bethesda, Md. Search for other works by this author on: Oxford Academic Google Scholar American Journal of Clinical Pathology, Volume 54, Issue 1, 1 July 1970, Page 154, https://doi.org/10.1093/ajcp/54.1.154 Published: 01 July 1970