Omaha System Partnership Research Overview

1
Omaha System Partnership Research Overview

• Madeleine Kerr, PhD, RN

2
Background
3
Purpose

• Describe completed and in-progress studies of the
  partnership research teams (41 to date)
• Studies are listed and available through links at
  omahasystempartnership.org

4
Investigators and co-investigators

• Multidisciplinary teams (n27)
• International teams (n8)
• Students (n25)
• Community partners (n22)

5
Subjects

• High risk families served by public health nurses
  (n22)
• Elders receiving home care services (n12)
• Mothers with intellectual disabilities
• Children with special health care needs
• Diabetics
• Firefighters

6
Settings

• Public health
• Home health
• Community advocacy organizations
• Community coalitions
• Nurse-managed wellness centers
• Workforce studies

7
Electronic health record data
8
Study Methods

• Descriptive (n37)
• Inferential (n26)
• Text mining (n3)
• Machine learning (n2)
• Data visualization (n1)

9
Machine learning example

• Data Management for
• Intervention Effectiveness Research
• Comparing Deductive and Inductive Approaches
• Purpose To use data mining techniques to create
  meaningful intervention clusters from structured
  Omaha System intervention data

Monsen, K.A., et al. (2009). Research in Nursing
and Health, 32 (6),647-656
10

• Method Intervention data from 2,862 clients from
  15 home care agencies managed using 3 deductive
  approaches and 1 inductive (data mining) method.
• Results The data mining approach generated more
  intervention groups (24) compared to action
  category, theoretical and clinical expert
  consensus approaches.

Intervention Scheme
11
Text mining example

• Informing Standard Development and Understanding
  User Needs with Omaha System Signs and Symptoms
  Text Entries in Community-Based Care Settings
• Purpose To study free text with Omaha System
  data, to improve use in computerized platforms,
  identify gaps, and propose improvements. Melton,
  G. B., et al. (2010). Proceedings of the 2010
  American Medical Informatics Association
  Symposium, 512-516.

12

• Method Free text data for other signs and
  symptoms from 2 years of client records analyzed
  by content experts into categories.
• Results Five categories 1) duplicate entries,
  2) multiple concepts, 3) medical diagnoses,
  4)interventions and 5) comments.

Signs and Symptoms
13
Descriptive inferential example

• Benchmark Attainment by Maternal and Child Health
  Clients Across Public Health Nursing Agencies
• Purpose To demonstrate benchmarking of public
  health nursing outcomes
• Monsen, K.A., et al. (2011). Public Health
  Nursing, 29(1), 11-18.

14

• Method MCH data from from 6 counties using a
  benchmark of 4 (1lowest, 5highest)
• Results All counties showed significant
  increases in client knowledge benchmark
  attainment 4 of 6 counties showed increases for
  behavior status.

Knowledge
Behavior
Status
15
Studies in progress

• Using Visualization Methodsto Discover Tailoring
  in PHN Intervention Data
• Evaluating Effects of PHN Home Visiting on Health
  Literacy (NI2012 AMIA-0421)
• Occupational Health Nursing Informatics Mapping
  Hearing Health Outcomes to the Omaha System
  (NI2012 AMIA-0413)

16
Questions?

• Omaha System Partnership omahasystempartnership.o
  rg
• Madeleine Kerr kerrx010_at_umn.edu