IHI-Ching Hou

I-Ching Hou

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


I-Ching Hou has a total of 69 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Authors' Reply: Interpreting Agreement in AI-Assisted American Society of Anesthesiologists (ASA) Classification

    by Bo-Han Lin, Cheng-Wei Lu, I-Ching Hou, Tzu-Yu Lin, Chin-E Liu, Chong-Sun Khoi

  • Open Access

    GPT-4o–Powered Preanesthetic AI: Development and Validation Study

    by Bo-Han Lin, Cheng-Wei Lu, I-Ching Hou, Chong-Sun Khoi, Tzu-Yu Lin, Chin-E Liu

    Background: Accurate preanesthetic assessment is essential for perioperative risk stratification, but conventional tools such as the American Society of Anesthesiologists (ASA) physical status classification and postoperative nausea and vomiting (PONV) risk scores may be affected by subjective judgment, incomplete documentation, and fragmented clinical data. Large language models may support preanesthetic assessment by integrating structured and unstructured clinical information. Objective: This study aimed to develop and retrospectively validate a GPT-4o (OpenAI)–powered AI system for preanesthetic assessment. The primary validation focus was the agreement between AI-generated and clinician-assigned ASA physical status classifications. PONV risk stratification was evaluated as an additional clinically relevant performance outcome. A secondary objective was to assess the incremental contribution of National Health Insurance (NHI) cloud data to model performance. Methods: This single-center retrospective validation study included 600 adult surgical patients randomly sampled from 4404 eligible inpatient surgical patients between January and May 2025. The system processed structured and unstructured electronic health record data, with and without NHI cloud data, to generate preanesthetic assessment outputs. ASA agreement was assessed using Cohen κ with 95% CIs. PONV prediction was evaluated using sensitivity, specificity, overall accuracy, positive predictive value, negative predictive value, confusion matrix analysis, the 2-sided Fisher-Freeman-Halton exact test, and Cramér V. Results: Among the 600 patients, clinician-assigned ASA classifications were ASA I in 39 patients, ASA II in 498 patients, ASA III in 57 patients, and ASA IV in 6 patients. Agreement between AI-generated and clinician-assigned ASA classifications was high when NHI data were incorporated (κ=0.883, 95% CI 0.841‐0.946), whereas agreement was lower without NHI data (κ=0.518, 95% CI 0.412‐0.691). A total of 49 (8.2%) patients experienced documented PONV within 24 hours after surgery. Using the high-risk category as the primary test-positive threshold, the model achieved a sensitivity of 34.7% (95% CI 22.9‐48.7), a specificity of 99.1% (95% CI 97.9‐99.6), a positive predictive value of 77.3%, and a negative predictive value of 94.5%. The association between the model-assigned PONV risk categories and observed PONV outcomes was statistically significant ( P <.001, Cramér V=0.531). Conclusions: The GPT-4o–powered preanesthetic assessment system demonstrated feasibility and high agreement with clinician-assigned ASA classifications when NHI cloud data were incorporated. For PONV risk stratification, the system showed high specificity and a high negative predictive value but modest sensitivity, indicating that it identified a small high-risk subgroup with a high observed PONV incidence but did not capture all patients who developed PONV. These findings support the use of large language model–based decision-support tools as adjuncts to, rather than replacements for, anesthesiologist judgment.


  • Open Access

    Using drawings to examine the interactions between parents with advanced lung cancer and their children/adolescents

    by Mei-Ling Chu, I-Ching Hou, Chao-Chi Ho, Yen-Ya Chen, Ya-Ting Kuan

    Parental advanced lung cancer substantially affects parent-child relationships, particularly with children/adolescents. This study examined interactions, communication patterns, and coping strategies between middle-aged parents with advanced lung cancer and their children/adolescents. Purposive sampling recruited parents with advanced lung cancer and their children/adolescents. Data were collected through in-depth interviews and children's drawings using interpretative phenomenological analysis. 19 parents and 17 children/adolescents participated. Four themes emerged: (1) parental experiences and effects on parental roles; (2) effects on parent-child relationships; (3) children/adolescent's adjustment to Parental Cancer: physical and psychological; and (4) reconstruction of parent-child interactions and relationships. Families exhibited communication difficulties and emotional distancing, with children desiring love, support, and security. Middle-aged parents with advanced lung cancer experience emotional connection difficulties. Family-centred support is essential for strengthening communication, emotional bonding, and coping strategies.

  • Open Access

    Preliminary effects of an online diet, walking, and mindfulness-integrated lifestyle intervention for breast cancer survivors with overweight or obesity: A pilot randomized controlled trial

    by Yi-Fung Lin, Zerina Lokmic-Tomkins, King-Jen Chang, Mei-Lien Hu, Hsin-Chun Feng, Chin-Pao Cheng, Shiow-Ching Shun, Hsiu-Ju Chang, I-Ching Hou

  • Open Access

    Nurses’ Workarounds in Electronic Nursing Records for Pressure Injuries: A Mixed-Methods Study (Preprint)

    by I-Ching Hou, Shu-Min Chien, Chao-Ying Chang, Ya-Ting Kuan

    BACKGROUND Electronic Nursing Records (ENR) are essential tools for documenting pressure injury (PI) care and ensuring patient safety. However, nurses often adopt workarounds that deviate from standard procedures, which may affect documentation quality and workflow efficiency. Understanding the reasons and contexts of these workarounds is critical for improving clinical practice and electronic record systems. OBJECTIVE This study aimed to examine the prevalence, patterns, and motivations of nurses’ workarounds in pressure injury electronic nursing records (PIs-ENR) using a mixed-methods approach. METHODS In the first stage, 160 nurses from a regional hospital in Taiwan completed the PIs-ENR Workarounds instrument. Descriptive statistics and bivariate correlation analyses were conducted to identify relationships between demographic factors and workarounds. In the second stage, 18 nurses participated in focus groups and created empathy maps related to PIs-ENR workarounds. Qualitative content analysis was performed, and findings were mapped onto the revised Sociotechnical Electronic Health Record Workaround Analysis (SEWA) framework. Pain points and workflow challenges were further explored. RESULTS The mean PIs-ENR workaround score was 2.75 (SD = 0.36), indicating that workarounds frequently occurred. Age and years of experience were positively correlated with workaround frequency. Significant correlations were also observed among subdomains of PIs-ENR documentation. Qualitative analysis revealed that efficiency was the predominant rationale for workarounds, followed by social norms, task inference, and commitment to patient interaction. Key pain points included excessive treatment activities, limited staffing, unclear task division, frequent revisions of operating standards, and complex or non-user-friendly electronic systems. CONCLUSIONS Nurses commonly adopt workarounds in PIs-ENR to manage high workloads, prioritize patient care, and navigate complex documentation systems. These findings highlight the need to optimize ENR design, align workflows with clinical practice, and address staffing and administrative challenges to enhance documentation quality and patient safety.

  • Open Access

    Translation and validation of the traditional Chinese Mobile App Rating Scale (MARS-TW) for evaluating mHealth apps

    by Shih-Ting Chen, Yi-Ting Lu, Stoyan Stoyanov, Pei-Ching Liu, King-Jen Chang, I-Ching Hou

    Background: Mobile health (mHealth) applications are widely used in Taiwan, yet their quality remains unregulated. The Mobile App Rating Scale (MARS) is a validated tool for assessing mHealth app quality, but a Traditional Chinese version has not been developed and validated. This study aimed to translate and validate the traditional Chinese version of MARS (MARS-TW) and assess its reliability and validity using a cross-cultural adaptation framework. Methods: A three-phase cross-sectional study was conducted: (I) conceptual and item equivalence were evaluated by three experts; (II) semantic equivalence was assessed through forward and backward translation; (III) operational, measurement, and functional equivalence were tested on 44 mHealth apps related to cancer care (CC) and hypertension self-management (HTM), rated independently by two reviewers. Results: The expert review confirmed conceptual and item equivalence. Semantic equivalence was achieved through iterative refinements. Reliability testing demonstrated good internal consistency (Cronbach’s α ≥0.80) and inter-rater reliability [intraclass correlation coefficient (ICC) ≥0.76]. Convergent validity was supported by strong correlations within subscales [item-subscale (I-S) correlation >0.3, item-total (I-T) correlation >0.2], while divergent validity was acceptable (subscale correlations Conclusions: MARS-TW is a reliable and valid tool for evaluating Traditional Chinese mHealth apps. Future research should explore its applicability across broader health-related app categories.

  • Open Access

    A Quality Survey on the Pressure Injury Nursing Information System in One Taiwan Regional Hospital: Based on Information System Successful Model

    by Shu-Min Chien, Chao-Ying Chang, I-Ching Hou

    Pressure injury nursing information system (PI-NIS) supports nursing care process. This study validates the quality of PI-NIS with system successful model through cross-sectional survey in one reginal hospital of northern Taiwan. The results showed the information quality, system quality, service quality, intention to use and user satisfaction were determining factors to the net benefits of using PI-NIS. The quality of PI-NIS was acceptable in this study.

  • Open Access

    Can “ChatGPT Prompt Formula” Support Patients with Breast Cancer in Information Needs Inquery?

    by I-Ching Hou, Yu-To Chen, Ya-Ting Kuan

    ChatGPT with prompt engineering had high potential to support information needs in clinical settings. The perfect prompt formula was recommended to include the components of task, context, persona, exemplar, format. This study conducted five types of ChatGPT prompt formula as patients inquires. The results showed the answers can provide the useful information for those unknow medical terminology. The impressive answer included the empathy content that may ignored by real person. The answers with designed format with precise exemplar such as question list when visit clinic provided good demonstration and benefit to the patients. Generally, ChatGPT can increase the quality of answer with advanced prompt formula. It is recommended to conduct more empirical validation. Strategy for patients’ education may contain these advanced ChatGPT prompt components in the future.

  • Open Access

    Mobile Health Technology for the Surveillance of Surgical Site Infections Among Women Undergoing Cesarean Section: A Scoping Review (Preprint)

    by Ya-Ting Kuan, Hsiu-Ju Chang, Chi-Wen Chen, Hao-Chih Tai, I-Ching Hou

    BACKGROUND With a global increase in cesarean section rates, the incidence of surgical site infections (SSIs) can increase due to short postoperative hospital stays and limited access to follow-up care. OBJECTIVE This study aimed to evaluate the current status and utilization of mobile health (mHealth) interventions for the surveillance of surgical site infections in post-cesarean section mothers. METHODS A systematic scoping review was conducted using the PubMed/MEDLINE, Ovid MEDLINE, Cochrane Library, and CINAHL databases covering the years 2015–2025. Nine studies were included. RESULTS A total of 2,891 women who had undergone cesarean section were included. Effective post-cesarean recovery requires targeted support, particularly within the first 30 days. It is critical to conduct systematic follow-up during this period, with an extension to 90 days for women identified with infections. The themes for the mobile health technologies used by women who were discharged after cesarean sections included the types of mobile health tools, concepts, and outcomes. The subthemes of the mobile health tools included mobile devices (n = 8, 34.8%), mobile applications (n = 7, 30.4%), smart phones (n = 7, 30.4%), and electronic tablets (n = 1, 4.3%). The subthemes of the concepts of mobile health technology use included image-based wound assessment (n = 6, 27.3%), SSI monitoring (n = 4, 18.2%), post-discharge follow-up (n = 4, 18.2%), remote assessment (n = 3, 13.6%), predictive value (n = 2, 9.1%), screening algorithms (n = 1, 4.5%), communication (n = 1, 4.5%), captured and transit wound photographs (n = 1, 4.5%). The subthemes of outcomes of mobile health technology use included support surveillance (n = 8, 72.7%), high level of satisfaction (n = 1, 9.1%), informed decision-making on preventative strategies (n = 1, 9.1%), and reduced SSI rates (n = 1, 9.1%). The use of mHealth technologies is primarily focused on the development and evaluation of post-cesarean section image-based records to improve SSI detection and empower women to actively participate in post-discharge surveillance. CONCLUSIONS Healthcare providers can utilize mHealth technologies to improve women’s engagement in post-cesarean section wound care and enhance image-based postoperative follow-up, thereby improving the overall quality of care.

  • Open Access

    Correction: Development of a Therapeutic Video Game With the MDA Framework to Decrease Anxiety in Preschool-Aged Children With Acute Lymphoblastic Leukemia: Mixed Methods Approach

    by Dai-Jie Yang, Meng-Yao Lu, Chi-Wen Chen, Pei-Ching Liu, I-Ching Hou

    [This corrects the article DOI: 10.2196/37079.].