Data Analysis of Consumer

1
Data Analysis of Consumer Patient Health
Questions Using NLM Medical Subject Headings
2
Sandy Oelschlegel, Martha Earl,Jenny
ColePreston Medical Library, UT Graduate School
of MedicineKnoxville TN
3
Purpose

• Reporting on the use of Medical Subject Headings
  and an Access database to analyze consumer health
  questions received by the Preston Medical Library
  from 2002-2004.

4
Background

• The Preston Medical Library has provided a
  Consumer Patient Health Information Service
  since 1993.
• An average of 60 questions are answered monthly
  on a variety of topics.
• A form is filled out for each question received.
• Statistics have been kept on the number of
  questions, but no other information on the form
  has been analyzed

5
What Did We Do?

• In 2003 a database was designed by a library
  student employee.
• The database included fields for the subjects of
  the questions.
• National Library of Medicine Medical Subject
  Headings (MeSH) were used to populate these
  fields.

6
What Did We Do?

• Approximately 480 individual consumer requests
  were entered In the database
• Requests were from Knox and surrounding counties

7
Results/Outcome

• The plans for the database were for the
  Librarians to analyze the headings for
• collection development
• staff training
• web site development
• All for the Consumer Patient Health Information
  service

8
(No Transcript)
9
Results/Outcome

• It was also be used to correlate questions to
  demographics by looking at zip codes.

10
(No Transcript)
11
Results/Outcomes

• Data analysis resulted in plans for further
  development of the database and the submission of
  a grant proposal

12
Results

• It was apparent that the database needed to be
  changed to separate the consumer name and
  addresses from the health question
• Several other changes needed to be made in order
  to enter data (for example multiple questions
  from a single consumer).
• Data integrity protection needed to be put in
  place

13
Results

• In addition
• librarians began to consider whether the data
  might prove that consumer calls increased when
  outreach training was done in a geographic area
• And whether the topics of the calls received
  would correlate with the major disease states
  of East Tennessee

14
SC/MLA Research Grant Proposal

• Research Methodology The research project will
  be a retrospective look at data already collected
  and will serve to prove two hypotheses.

15
Our Research Questions

• .. whether outreach programs conducted by
  Preston Medical Library have an impact on the
  subsequent number of consumer health questions
  received from that geographic area, and to prove
  the theory that questions are most often related
  to the major disease states of East Tennessee.

16
Our Hypotheses

• Outreach increases the number of questions.
• Consumer questions we receive are most often
  related to the following disease states (heart
  disease, cancer, stroke, chronic lower
  respiratory disease, accidents/unintentional
  injury).

17
What Will We Do?

• Revise secure the database, increase data
  integrity features
• The data will be analyzed relative to the
  hypotheses 

18
Data Analysis Relative to Hypotheses

• Correlating the frequency of questions to zip
  codes and comparing the results to the frequency
  and location of outreach programs.
• Correlating the frequency of MeSH topics to the 5
  Major Disease States of East Tennessee region

19
Questionssoelschl_at_mc.utmck.edumearl_at_mc.utmck.edu
Jcole_at_mc.utmck.edu