Dataset Information

Automatic quality of life prediction using electronic medical records.

ABSTRACT: Health related quality of life (HRQOL) is an important variable used for prognosis and measuring outcomes in clinical studies and for quality improvement. We explore the use of a general pur-pose natural language processing system Metamap in combination with Support Vector Machines (SVM) for predicting patient responses on standardized HRQOL assessment instruments from text of physicians notes. We surveyed 669 patients in the Mayo Clinic diabetes registry using two instruments designed to assess functioning: EuroQoL5D and SF36/SD6. Clinical notes for these patients were represented as sets of medical concepts using Metamap. SVM classifiers were trained using various feature selection strategies. The best concordance between the HRQOL instruments and automatic classification was achieved along the pain dimension (positive agreement .76, negative agreement .78, kappa .54) using Metamap. We conclude that clinicians notes may be used to develop a surrogate measure of patients HRQOL status.

SUBMITTER: Pakhomov S

PROVIDER: S-EPMC2656045 | biostudies-literature | 2008 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Automatic quality of life prediction using electronic medical records.

Pakhomov Sergeui S Shah Nilay N Hanson Penny P Balasubramaniam Saranya S Smith Steven A SA

AMIA ... Annual Symposium proceedings. AMIA Symposium 20081106

Health related quality of life (HRQOL) is an important variable used for prognosis and measuring outcomes in clinical studies and for quality improvement. We explore the use of a general pur-pose natural language processing system Metamap in combination with Support Vector Machines (SVM) for predicting patient responses on standardized HRQOL assessment instruments from text of physicians notes. We surveyed 669 patients in the Mayo Clinic diabetes registry using two instruments designed to assess ...[more]

PMID: 18998862

Similar Datasets

Project description:BackgroundRisk scores can be useful in clinical risk stratification and accurate allocations of medical resources, helping health providers improve patient care. Point-based scores are more understandable and explainable than other complex models and are now widely used in clinical decision making. However, the development of the risk scoring model is nontrivial and has not yet been systematically presented, with few studies investigating methods of clinical score generation using electronic health records.ObjectiveThis study aims to propose AutoScore, a machine learning-based automatic clinical score generator consisting of 6 modules for developing interpretable point-based scores. Future users can employ the AutoScore framework to create clinical scores effortlessly in various clinical applications.MethodsWe proposed the AutoScore framework comprising 6 modules that included variable ranking, variable transformation, score derivation, model selection, score fine-tuning, and model evaluation. To demonstrate the performance of AutoScore, we used data from the Beth Israel Deaconess Medical Center to build a scoring model for mortality prediction and then compared the data with other baseline models using the receiver operating characteristic analysis. A software package in R 3.5.3 (R Foundation) was also developed to demonstrate the implementation of AutoScore.ResultsImplemented on the data set with 44,918 individual admission episodes of intensive care, the AutoScore-created scoring models performed comparably well as other standard methods (ie, logistic regression, stepwise regression, least absolute shrinkage and selection operator, and random forest) in terms of predictive accuracy and model calibration but required fewer predictors and presented high interpretability and accessibility. The nine-variable, AutoScore-created, point-based scoring model achieved an area under the curve (AUC) of 0.780 (95% CI 0.764-0.798), whereas the model of logistic regression with 24 variables had an AUC of 0.778 (95% CI 0.760-0.795). Moreover, the AutoScore framework also drives the clinical research continuum and automation with its integration of all necessary modules.ConclusionsWe developed an easy-to-use, machine learning-based automatic clinical score generator, AutoScore; systematically presented its structure; and demonstrated its superiority (predictive performance and interpretability) over other conventional methods using a benchmark database. AutoScore will emerge as a potential scoring tool in various medical applications.

Project description:AimsAssess the relationship between timely treatment intensification and hemoglobin A1C (HbA1C) control quality-of-care performance measures, i.e., HbA1C levels, among patients with uncontrolled type 2 diabetes.Materials and methodsElectronic medical records and diabetes registry data from a large, accountable care organization (ACO) were used to isolate a sample of adult patients with type 2 diabetes who received at least one oral antidiabetes agent and had at least one HbA1C level measurement ≥8.0% (64 mmol/mol; i.e., uncontrolled diabetes) between 7/1/2011 and 6/30/2015. Treatment intensification status was evaluated for each patient during a 120-day treatment intensification window following the index HbA1c measure. Two-level hierarchical generalized linear models, with patients aggregated at the physician level, were used to assess the association between treatment intensification and achieving HbA1C quality performance measures.Results547 patients met study selection criteria and 480 patients had at least one HbA1C test after the treatment intensification window and were used for the statistical analyses. About 40% of patients who had uncontrolled diabetes received treatment intensification during the 120-day window. Greater index HbA1C, greater patient body mass index, and fewer unique pre-index oral antidiabetes agents were significantly associated with greater likelihood of receiving timely treatment intensification. The odds of receiving treatment intensification were about 1.8 times higher (P = 0.0027) among patients with poor index HbA1C control (HbA1c level >9.0% [75 mmol/mol]) compared to other patients (index HbA1c 8.0% - 9.0%). Hispanic patients (compared to White patients) were significantly more likely to exhibit poor control after treatment intensification (odds ratio [OR] 2.91, P = 0.0304), underscoring the difficulty of controlling diabetes in this vulnerable group. In contrast, being male and being treated primarily by an internist (compared to primary treatment by a family medicine specialist) were both significantly associated with achieving superior control (HbA1c level <8.0%) after treatment intensification (OR 0.53 [P = 0.0165]; OR 0.41 [P = 0.0275], respectively).ConclusionsTimely treatment intensification was significantly associated with greater likelihood of patients achieving superior HbA1C control (<8.0%) and better HbA1C control quality performance for the practice. Even in an ACO with resources dedicated to diabetes control, it is incumbent upon clinicians to readily identify and open dialogues with patients who may benefit from closely supervised, individualized attention.

Dataset Information

Automatic quality of life prediction using electronic medical records.

Publications

Automatic quality of life prediction using electronic medical records.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets