Knowledge Discovery in Data Bases to Predict Improvement in Oral Medication Management for Home Healthcare Patients

2.50
Hdl Handle:
http://hdl.handle.net/10755/158397
Type:
Presentation
Title:
Knowledge Discovery in Data Bases to Predict Improvement in Oral Medication Management for Home Healthcare Patients
Abstract:
Knowledge Discovery in Data Bases to Predict Improvement in Oral Medication Management for Home Healthcare Patients
Conference Sponsor:Midwest Nursing Research Society
Conference Year:2010
Author:Westra, Bonnie, PhD, RN
P.I. Institution Name:University of Minnesota
Title:School of Nursing
Contact Address:308 Harvard St SE, WDH 5-140, Minneapolis, MN, 55455, USA
Contact Telephone:612-625-4470
Co-Authors:B.L. Westra, K. Savik, C. Oancea, School of Nursing, University of Minnesota, Minneapolis, MN; M. Steinbach, S. Dey, G. Fang, V. Kumar, Computer Science, University of Minnesota, Minneapolis, MN;
Problem: Annually, 29% of home healthcare patients experience an acute care hospitalization, often due to problems with managing medications. Framework: Donabedian's quality improvement model which is based on systems theory underlies the study. Purpose: The purpose of this study was to identify risk factors and clinician interventions from electronic health record data that predict improvement in oral medication management for home healthcare patients discharged and remaining in the community. Design & Methodology: We used a retrospective cohort design. OASIS assessment data, Omaha System Interventions, and medication data were abstracted and merged across 15 home healthcare agencies that used two different software vendors. Analysis: Predictive models were created to discover predictors for improvement in oral medication management using data mining techniques of discriminative pattern analysis and classification rules. Findings: The 1,688 cases represented predominately older Caucasian adults with two-thirds females whose services were frequently for post-hospitalization care. Improvement in oral medication management improved in 268 (16.1%) cases by discharge. Logistic regression produced three variables predicting a decreased likelihood of improvement and five associated with improvement. Discriminative pattern analysis resulted in two rules involving four variables that accounted for 90% of all cases for improvement or no improvement. Classification rules correctly classified patients likely to improve with a precision of .752 and recall of .94 while no improvement had a precision of .92 and recall of .69. The strongest predictor of improvement was the patient's baseline assessment; higher rates of dependence for managing oral medications at admission predicted higher improvement at discharge for oral medication management. Additional variables in the form of patterns and rules will be presented. Interpretation: Data mining techniques were useful to discover parsimonious patterns in the data and rules that are useful for clinicians to implement that Both methods had some overlap and some differences in predictive variables.
Repository Posting Date:
26-Oct-2011
Date of Publication:
17-Oct-2011
Sponsors:
Midwest Nursing Research Society

Full metadata record

DC FieldValue Language
dc.typePresentationen_GB
dc.titleKnowledge Discovery in Data Bases to Predict Improvement in Oral Medication Management for Home Healthcare Patientsen_GB
dc.identifier.urihttp://hdl.handle.net/10755/158397-
dc.description.abstract<table><tr><td colspan="2" class="item-title">Knowledge Discovery in Data Bases to Predict Improvement in Oral Medication Management for Home Healthcare Patients</td></tr><tr class="item-sponsor"><td class="label">Conference Sponsor:</td><td class="value">Midwest Nursing Research Society</td></tr><tr class="item-year"><td class="label">Conference Year:</td><td class="value">2010</td></tr><tr class="item-author"><td class="label">Author:</td><td class="value">Westra, Bonnie, PhD, RN</td></tr><tr class="item-institute"><td class="label">P.I. Institution Name:</td><td class="value">University of Minnesota</td></tr><tr class="item-author-title"><td class="label">Title:</td><td class="value">School of Nursing</td></tr><tr class="item-address"><td class="label">Contact Address:</td><td class="value">308 Harvard St SE, WDH 5-140, Minneapolis, MN, 55455, USA</td></tr><tr class="item-phone"><td class="label">Contact Telephone:</td><td class="value">612-625-4470</td></tr><tr class="item-email"><td class="label">Email:</td><td class="value">westr006@umn.edu</td></tr><tr class="item-co-authors"><td class="label">Co-Authors:</td><td class="value">B.L. Westra, K. Savik, C. Oancea, School of Nursing, University of Minnesota, Minneapolis, MN; M. Steinbach, S. Dey, G. Fang, V. Kumar, Computer Science, University of Minnesota, Minneapolis, MN;</td></tr><tr><td colspan="2" class="item-abstract">Problem: Annually, 29% of home healthcare patients experience an acute care hospitalization, often due to problems with managing medications. Framework: Donabedian's quality improvement model which is based on systems theory underlies the study. Purpose: The purpose of this study was to identify risk factors and clinician interventions from electronic health record data that predict improvement in oral medication management for home healthcare patients discharged and remaining in the community. Design &amp; Methodology: We used a retrospective cohort design. OASIS assessment data, Omaha System Interventions, and medication data were abstracted and merged across 15 home healthcare agencies that used two different software vendors. Analysis: Predictive models were created to discover predictors for improvement in oral medication management using data mining techniques of discriminative pattern analysis and classification rules. Findings: The 1,688 cases represented predominately older Caucasian adults with two-thirds females whose services were frequently for post-hospitalization care. Improvement in oral medication management improved in 268 (16.1%) cases by discharge. Logistic regression produced three variables predicting a decreased likelihood of improvement and five associated with improvement. Discriminative pattern analysis resulted in two rules involving four variables that accounted for 90% of all cases for improvement or no improvement. Classification rules correctly classified patients likely to improve with a precision of .752 and recall of .94 while no improvement had a precision of .92 and recall of .69. The strongest predictor of improvement was the patient's baseline assessment; higher rates of dependence for managing oral medications at admission predicted higher improvement at discharge for oral medication management. Additional variables in the form of patterns and rules will be presented. Interpretation: Data mining techniques were useful to discover parsimonious patterns in the data and rules that are useful for clinicians to implement that Both methods had some overlap and some differences in predictive variables.</td></tr></table>en_GB
dc.date.available2011-10-26T21:00:38Z-
dc.date.issued2011-10-17en_GB
dc.date.accessioned2011-10-26T21:00:38Z-
dc.description.sponsorshipMidwest Nursing Research Societyen_GB
All Items in this repository are protected by copyright, with all rights reserved, unless otherwise indicated.