The feasibility of a Bayesian network model to assess the probability of simultaneous symptoms in patients with advanced cancer.

Although patients with advanced cancer often experience multiple symptoms simultaneously, clinicians usually focus on symptoms that are volunteered by patients during regular history-taking. We aimed to evaluate the feasibility of a Bayesian network (BN) model to predict the presence of simultaneous symptoms, based on the presence of other symptoms. Our goal is to help clinicians prioritize which symptoms to assess. Patient-reported severity of 11 symptoms (scale 0-10) was measured using an adapted Edmonton Symptom Assessment Scale (ESAS) in a national cross-sectional survey among advanced cancer patients. Scores were dichotomized (< 4 and ≥ 4). Using fourfold cross validation, the prediction error of 9 BN algorithms was estimated (Akaike information criterion (AIC). The model with the highest AIC was evaluated. Model predictive performance was assessed per symptom; an area under curve (AUC) of ≥ 0.65 was considered satisfactory. Model calibration compared predicted and observed probabilities; > 10% difference was considered inaccurate. Symptom scores of 532 patients were collected. A symptom score ≥ 4 was most prevalent for fatigue (64.7%). AUCs varied between 0.60 and 0.78, with satisfactory AUCs for 8/11 symptoms. Calibration was accurate for 101/110 predicted conditional probabilities. Whether a patient experienced fatigue was directly associated with experiencing 7 other symptoms. For example, in the absence or presence of fatigue, the model predicted a 8.6% and 33.1% probability of experiencing anxiety, respectively. It is feasible to use BN development for prioritizing symptom assessment. Fatigue seems most eligble to serve as a starting symptom for predicting the probability of experiencing simultaneous symptoms.

Paper versus computer: feasibility of an electronic medical record in general pediatrics.

BACKGROUND: Implementation of electronic medical record systems promises significant advances in patient care, because such systems enhance readability, availability, and data quality. Structured data entry (SDE) applications can prompt for completeness, provide greater accuracy and better ordering for searching and retrieval, and permit validity checks for data quality monitoring, research, and especially decision support. A generic SDE application (OpenSDE) to support documentation of patient history and physical examination findings was developed and tailored for the domain of general pediatrics. OBJECTIVE: To evaluate OpenSDE for its completeness, uniformity of reporting, and usability in general pediatrics. METHODS: Four (trainee) pediatricians documented data for 8 first-visit patients in the traditional, paper-based, medical record and immediately thereafter in OpenSDE (electronic record). The 32 paper records obtained served as the common data source for data entry in OpenSDE by the other 3 physicians (transcribed record). Data entered by 2 experienced users, with all patient information present in the paper record, served as the control record. Data entry times were recorded, and a questionnaire was used to assess users' experiences with OpenSDE. RESULTS: Clinicians documented 44% of all available patient information identically in the paper and electronic records. Twenty-five percent of all patient information was documented only in the paper record, and 31% was present only in the electronic record. Differences were found in patient history and physical examination documentation in the electronic record; more information was missing for patient history (38%) than for physical examination (15%). Furthermore, physical examination contained more additional information (39%) than did patient history (21%). The interobserver agreement of documentation of patient information from the same data source was fair to moderate, with kappa values of 0.39 for patient history and 0.40 for physical examination. Data entry times in OpenSDE decreased from 25 minutes to <15 minutes, indicating a learning effect. The questionnaire revealed a positive attitude toward the use of OpenSDE in daily practice. CONCLUSION: OpenSDE seems to be a promising application for the support of physician data entry in general pediatrics.

Extracting data recorded with OpenSDE: possibilities and limitations.

PURPOSE: OpenSDE is an application intended to support structured data entry in a variety of settings, such as routine care and clinical research. The past years development has focused on data entry to support expressiveness and flexibility. The focus is now shifting to data extraction: what are the possibilities for extracting the data and does the adopted strategy pose limitations? METHODS: Data extraction is supported by presenting the concepts for extraction in the same tree structure as for data entry. Users can select all or a sub selection of these concepts for extraction. Selected concepts are extracted and converted to a table format that can be queried using conventional tools. RESULTS: The extraction tool (entity export) provides a successful technical solution for data extraction. Using the extracted data, however, leads to obstacles that are a result of a fundamental design principle of OpenSDE.

OpenSDE: a strategy for expressive and flexible structured data entry.

PURPOSE: This description focuses on the expressiveness and flexibility of OpenSDE: an application that supports recording of structured narrative data. METHODS: OpenSDE enables data entry with (customizable) forms based on trees of medical concepts. The relevant scope for data entry can be tailored per medical domain by construction of a domain-specific tree. OpenSDE is intended for structuring narrative data to make these available for both care and research. RESULTS: The OpenSDE application is currently in use at several departments in our academic hospital, including radiology, neurology, pediatrics, and child psychiatry. OpenSDE is available for all in open source.

Why are structured data different? Relating differences in data representation to the rationale of OpenSDE.

OpenSDE is an application that supports clinicians with structured recording of narrative patient data to enable use of data in both clinical practice and research. OpenSDE is based on a rationale and requirements for structured data entry. In this study, we analyse the impact of the rationale and the requirements on data representation using OpenSDE. Three paediatricians transcribed 20 paper patient records using OpenSDE. The transcribed records were compared; the findings that were the same in content but differed in representation (e.g. recorded as free text instead of in a structured manner) were categorized in one of three categories of difference in representation. The transcribed records contained 1764 findings in total. The medical content of 302 of these findings was represented differently by at least one clinician and was thus included in this study. In OpenSDE, clinicians are free to determine the degree of detail at which patient data are described. This flexibility accounts for 87% of the differences in data representation. Thirteen per cent of the differences are due to clinicians interpreting and translating phrases from the source text and transcribing these to (different) concepts in OpenSDE. The differences in data representation largely result from initial design decisions for OpenSDE.

OpenSDE: Row modeling applied to generic structured data entry.

Clinicians generally record medical narrative data, such as current complaints, physical examination, and progress notes, as free text in paper-based medical records. The medical narrative involves heterogeneous and detailed data that include the description of (multiple) occurrences of medical findings or symptoms that may progress over time. Structured, electronic recording of narrative data would facilitate the use of these data for research. The authors' OpenSDE application supports clinicians with the structured recording of narrative data in both research and care settings. Data entry is enabled using forms that are generated using domain-specific trees of medical concepts. For data storage, the authors have expanded the traditional row modeling methodology with additional columns that allow structured representation of medical narratives including descriptions of findings, multiple occurrences of findings, and the progression of findings over time.