Medical visual question answering (MVQA) plays a crucial role in the fields of computer-aided diagnosis and telemedicine. Due to the limited size and uneven annotation quality of the MVQA datasets, most existing methods rely on additional datasets for pre-training and use discriminant formulas to predict answers from a predefined set of labels. This approach makes the model prone to overfitting in low resource domains. To cope with the above problems, we propose an image-aware generative MVQA method based on image caption prompts. Firstly, we combine a dual visual feature extractor with a progressive bilinear attention interaction module to extract multi-level image features. Secondly, we propose an image caption prompt method to guide the model to better understand the image information. Finally, the image-aware generative model is used to generate answers. Experimental results show that our proposed method outperforms existing models on the MVQA task, realizing efficient visual feature extraction, as well as flexible and accurate answer outputs with small computational costs in low-resource domains. It is of great significance for achieving personalized precision medicine, reducing medical burden, and improving medical diagnosis efficiency.
In the process of guideline development and construction of clinical questions, it is necessary to guide clinicians to propose clinical problems into PICO (population, intervention, control, outcome) structured clinical questions. However, there are still unclear criteria to define and judge the appropriateness of the width of the PICO elements of a clinical question. Either too wide or too narrow can make the PICO question unsuitable to be a question for clinical practice guidelines to answer. We graded the clinical questions to be eight grades (3, 2, 1, 0, −1, −2, −3, mixed) according to the number of the PIC elements, which obviously needed to be adjusted to evaluate applicability of the appropriateness of the width of the clinical questions. Our work can provide methodological references for clinicians and guideline developers.
ObjectiveTo analyze responsiveness of Chinese version of Neck Outcome Score (NOOS-C) and provide a reliable measure to assess intervention effect for patients with neck pain.MethodsCross-cultural adaptation of NOOS was performed according to the Beaton’s guidelines for cross-cultural adaptation of self-report measures. Eighty patients with neck pain were recruited between September 2016 and May 2017. Those patients were assessed using NOOS-C and Chinese version of Neck Disability Index (NDI) before and after intervention. And 71 patients completed those questionnaires. The statistic differences of the score of each subscale and the total scale before and after intervention were evaluated by paired-samples t test. Internal responsiveness was determined by effect size (ES) and standardized response mean (SRM) based on the calculated difference before and after intervention. External responsiveness was analyzed by Spearman correlation coefficient.ResultsThe differences in symptom subscale, sleep disturbance subscale, participating in everyday life subscale, every day activity and pain subscale, and the scale between before and after intervention were significant (P<0.05) except for mobility subscale (P>0.05). The difference of NDI-C before and after intervention was –12.11%±17.45%, ES was 0.77, and SRM was 0.69. The difference of NOOS-C before and after intervention was 13.74±17.22, ES was 0.83, and SRM was 0.80. Spearman correlation analysis revealed that the relativity about NOOS-C and NDI-C before and after intervention were both negative (r=–0.914, P=0.000; r=–0.872, P=0.000).ConclusionNOOS-C’s responsiveness is good.
ObjectiveThe current medical questionnaire resources are mainly processed and organized at the document level, which hampers user access and reuse at the questionnaire item level. This study aims to propose a multi-class classification of items in medical questionnaires in low-resource scenarios, and to support fine-grained organization and provision of medical questionnaires resources. MethodsWe introduced a novel, BERT-based, prompt learning approach for multi-class classification of items in medical questionnaires. First, we curated a small corpus of lung cancer medical assessment items by collecting relevant clinical assessment questionnaires, extracting function and domain classifications, and manually annotating the items with "function-domain" combination labels. We then employed prompt learning by feeding the customized template into BERT. The masked positions were predicted and filled, followed by mapping the populated text to labels. This process enables the multi-class classification of item texts in medical questionnaires. ResultsThe constructed corpus comprised 347 clinical assessment items for lung cancer, across nine "function-domain" labels. The experimental results indicated that the proposed method achieved an average accuracy of 93% on our self-constructed dataset, outperforming the runner-up GAN-BERT by approximately 6%. ConclusionThe proposed method can maintain robust performance while minimizing the cost of building medical questionnaire item corpora, illustrating its promotion value of research and practice in medical questionnaire classification.
In the formulation of the clinical question of traditional Chinese medicine clinical practice guidelines, even if the intervention elements (intervention or control) have an appropriate scope, guideline developers are still faced with a variety of interventions. By analyzing the difficulty and necessity of priority selection of intervention interventions, we propose the approach of extending expert evidence to the process of priority selection of intervention interventions, and further provide the methodology of expert evidence data collection table design, application, data presentation and expert decision-making method to provide references and guidance for guideline developers.
Objective To verify the reliability and validity of a self-developed satisfaction evaluation questionnaire for outpatient department employees in public hospitals, and to provide suitable tools for conducting such surveys. Methods Two anonymous surveys were conducted on all employees of the Outpatient Department of West China Hospital of Sichuan University in July 2019 and November 2021, respectively. Questionnaire items were screened using methods such as item distribution, coefficient of variation, and decision value, and the reliability and validity of the questionnaire were evaluated using Spearman-Brown coefficient and Cronbach’s α coefficient, exploratory factor analysis, and confirmatory factor analysis. Results The final questionnaire retained 14 items, which could be divided into two dimensions: work conditions and interpersonal environment, and the overall fit index of structural equation model were as follows: χ2/ν=6.957, the standardized root mean square residual was 0.061, the root mean square error of approximation was 0.147, the goodness-of-fit index was 0.796, the adjusted goodness-of-fit index was 0.719, the normed fit index was 0.849, the relative fit index was 0.819, the incremental fit index was 0.868, the Tucker-Lewis Index was 0.841, and the comparative fit index was 0.867. The combined reliability of the two factors in the questionnaire was 0.94 and 0.91, respectively. The average variance extraction was 0.67 and 0.76, respectively, and the square root of the average variance extraction was 0.82 and 0.87, respectively, both of which were greater than the correlation coefficient of 0.71 between the two factors. The Spearman-Brown coefficient of the final questionnaire was 0.913, and the Cronbach’s α coefficients for the overall and two dimensions were 0.953, 0.937, and 0.910, respectively. Conclusion The reliability and validity of the satisfaction evaluation questionnaire for outpatient department employees in public hospitals are good and the questionnaire can be applied to practical surveys.
Recent studies have introduced attention models for medical visual question answering (MVQA). In medical research, not only is the modeling of “visual attention” crucial, but the modeling of “question attention” is equally significant. To facilitate bidirectional reasoning in the attention processes involving medical images and questions, a new MVQA architecture, named MCAN, has been proposed. This architecture incorporated a cross-modal co-attention network, FCAF, which identifies key words in questions and principal parts in images. Through a meta-learning channel attention module (MLCA), weights were adaptively assigned to each word and region, reflecting the model’s focus on specific words and regions during reasoning. Additionally, this study specially designed and developed a medical domain-specific word embedding model, Med-GloVe, to further enhance the model’s accuracy and practical value. Experimental results indicated that MCAN proposed in this study improved the accuracy by 7.7% on free-form questions in the Path-VQA dataset, and by 4.4% on closed-form questions in the VQA-RAD dataset, which effectively improves the accuracy of the medical vision question answer.
Objective To analyze the nurses' current view and perceptions of enhanced recovery after surgery (ERAS) by a questionnaire and to promote the clinical application of ERAS. Methods We conducted a questionnaire study for nurses who attended the First West China Forum on Chest ERAS in Chengdu during September 26-27, 2016 and 259 questionnaires were collected for descriptive analysis. Results (1) The application status of ERAS: There were 13.5% responders whose hospital took a wait-an-see attitude, while the others' hospital took different actions for ERAS; 85.7% of nurses believed that ERAS in all surgeries should be used; 58.7% of nurses believed that the concept of ERAS was more in theory than in the practice; 40.2% of nurses thought that all patients were suitable for the application of ERAS; (2) 81.9% of nurses believed that the evaluation criteria of ERAS should be a combination of the average hospital stay, patients’ comprehensive feelings and social satisfaction; (3) 70.7% of nurses thought that the combination of subjects integration, surgery orientation and surgeon-nurse teamwork was the best model of ERAS; 44.8% of nurses thought the hospital administration was the best way to promote ERAS applications; (4) 69.1% of responders believed that immature plan, no consensus and norms and insecurity for doctors were the reasons for poor compliance of ERAS; 79.5% of nurses thought that the ERAS meeting should include the publicity of norms and consensus, analysis and implementation of projects and the status and progress of ERAS. Conclusion ERAS concept has been recognized by most nurses. Multidisciplinary collaboration and hospital promotion is the best way to achieve clinical applications.
Objective To investigate the current development status of chest wall surgery at all levels of hospitals in Sichuan Province, as well as to provide evidence for the promotion of chest wall surgery. Methods We conducted a questionnaire study to investigate chest wall surgery at all levels of hospitals in Sichuan Province and to collect suggestions for chest wall surgery development from thoracic surgeons attending the meeting of the Sichuan International Medical Exchange & Promotion Association from September 2021 to January 2022. Results A total of 128 questionnaires were issued, with 97 (75.8%) of them being valid. According to the survey results, hospitals with grade A secondary or higher in Sichuan Province performed chest wall surgery. Chest wall surgery accounted for 14.3% of thoracic surgery, with 70.4% being chest wall trauma surgeries, 11.6% being chest wall tumor surgeries, 10.5% being chest wall infection surgeries, and 7.5% being chest wall deformity surgeries. Chest wall surgery accounted for 9.3% of thoracic surgery in the grade A tertiary hospitals, primarily for chest wall trauma and tumor; 23.1% in grade B tertiary hospitals, primarily for chest wall trauma and tumor; and 50.7% in grade A secondary hospitals, primarily for chest wall trauma and infection. Totally 96.9% of hospitals supported the establishment of a subspecialty in chest wall surgery. Suggestions for advancing chest wall surgery included: enhancing communication and cooperation (e.g. holding academic conferences, training courses), the establishment of the chest wall surgery association or consortium, and the formulation of regulations and guidelines or consensus, etc. Conclusion Chest wall surgery has been performed at all levels of hospitals in Sichuan Province. The relevant guidelines can be made based on the related academic associations, thus boosting the development of chest wall surgery in the future.
ObjectiveTo investigate the preoperative psychological state of patients with pulmonary nodules in order to make the content of the education more "individualized and humanized".MethodsWe conducted a consecutive questionnaire study for 107 patients who were planning to undergo pulmonary resection surgery from May 2018 to July 2018 in our department. There were 54 males and 53 females with an average age of 56.8±11.2 years. The questionnaire content included two parts: personal basic information and 20 questions about surgery, complications, follow-up and hospitalization expense.ResultsThere were 60.7% of the patients diagnosed with pulmonary nodules by CT scan during physical examination, and 52.3% of the patients had strong will to undergo pulmonary surgery to resect nodules; 64.5% of patients wanted doctors to tell them the extent of the disease and whether the tumor could be cured by surgery, and 30.0% of patients concerned whether chief surgeon would complete the whole surgery. The surgery risk and postoperative complications were ignored by patients easily (5.6% and 14.9% respectively). The hospital expenses were not the primary concern of patients. Only 1.9% of patients believed that doctors used nonessentials which deliberately led to increased costs. Network follow-up was accepted by most patients (94.4%).ConclusionIt will contribute to improve preoperative education rationality and effectiveness by understanding true psychological state of patients.