Transcript The Poster Title (72 Point Font for 2 Lines, 96 Point for 1 Line) Use

Examining Front-Line Decision Making in State and Local Communicable Disease
Control Programs – A qualitative analysis of semi-structured interviews.
Laverne Snow, MPA1; Heidi Kramer2; Frank Drews, PhD2; Matthew Samore, MD1,3
University of Utah Departments of Biomedical Informatics1, Psychology2, and Internal Medicine3, Salt Lake City, Utah
Problem: Reportable diseases affect millions of Americans --- 76 million illnesses,
325,000 hospitalizations and 5,000 deaths from foodborne illnesses alone.
Introduction
Front-line professionals in state and local public
health agencies make important decisions for
communicable disease control, under complex
and challenging conditions. While front-line
decisions are often routine, many have significant
impact on the control or spread of serious and
highly infectious diseases.
Examining human factors and understanding
the underlying cognition of decision makers is
essential for developing useful and accepted
decision support tools.
Aims
• Examine decision-making of front-line
governmental communicable disease control
professionals to identify decision making
processes under condition of ambiguity,
uncertainty, high risk and time pressure.
• Develop recommendations for decision support
tools.
Materials and Methods
Semi-structured interviews
• 11 pre-pilot, 8 pilot and 42 formal interviews
• State or local health departments
• Epidemiologists, nurses, environmental health
scientists and infectious disease investigators
• Utah and Nevada
Interview protocol
Two part interview to examine decision making:
• Semi-structure questions on use of protocols and
ambiguity, focus on work and school exclusions.
• Enteric disease cluster scenario with multiple
causal pathways.
Data collection and analysis
• Two interviewers in each interview
• Interviews recorded, transcribed and individually
coded
• Qualitative research methods - modified
grounded theory
• Atlas.ti Knowledge Workbench software
This research was conducted under the leadership of
the Utah Center for Excellence in Public Health
Informatics, Decision Support for Infectious Disease
Epidemiology (DSIDE) Project
CDC Grant # 8P01HK000030
Findings
Commonly believed among public health disease investigators:








Public health agencies rely on protocols as an important strategy for outbreak investigations.
Disease cases are significantly under-reported.
Many individuals do not seek health care treatment when ill.
Public health agencies rely heavily on lab tests for case confirmation before taking intervening actions.
Physicians under report disease cases and frequently rely on laboratories to report lab results.
Physicians do not always order appropriate lab tests needed for case confirmation.
Case situations vary; therefore, protocols are underspecified by necessity.
Staffing, resources and training for communicable disease control are often under-funded.
Less understood:
 Protocols are perceived to be a substitute for experience and expertise.
 Front-line public health workers are gate-keepers for disease case consideration and investigation.
 Individual judgment is frequently required to navigate uncertainty of information and ambiguity in
protocols.
 Disease reporting often occurs because of personal and professional relationships.
 Appropriateness of tests ordered, specimen handling, timing of taking tests, delays in receiving lab
results, adherence to test procedures and test specificity and sensitivity create additional uncertainty
about laboratory data - - - factors not consistently considered.
Protocol challenges for making decisions on exclusions from work or school
Substituting protocols for experience and
expertise
“I would need to read protocols on that
because I don't have personal experience.”
“ I need to understand what that disease looks
like and at least have something written that
helps guide you in that….”
Unpredictability
“… with epidemiology you can't say that
exactly it's always going to go this way, things
change.”
Time Delays
“One of the more troublesome parts of doing
disease investigations is that … by the time
you find out, the people who were contagious
are no longer contagious.”
Protocol Ambiguity
“So I think what I have found with communicable
disease is there are always exceptions to the rules
and there is generally no black and white.”
“ You have to make your best scientific judgment
based on what you know about the situation… it is
almost never exactly what the guideline says.”
Discussion
• Trust in protocols as a substitute for expertise may be
creating a fragile foundation for disease investigations.
• Reliance on protocols comes from technical systems,
like aviation, where conditions are controlled and
predictable. Public health interface with a natural system,
like disease evolution, are not controllable or predictable.
• Reasons protocols break down in public health include:
* Infrequent and unfamiliar diseases
* Unpredictable time delays
* Missing and uncertain information
* Information overload
* Appropriate lab tests not taken
* Inconsistent judgment to resolve ambiguity
Vision for Future Research
Enhanced Decision Support Tools for Disease
Reporting and Investigations
 Electronic physician reporting using their
electronic medical records.
 Patient and school web portals for reporting to
local public health agencies.
 Automated case investigation forms populated by
electronic disease reports.
 Central knowledge base on PH grid with
reference information, case definitions, community
specific data, and directories.
 Automated line lists populated by electronic case
investigation forms.
 Algorithms to help identify common risks.
 Outbreak simulator for exploring “what-if”
scenarios.
Special thanks to those in Utah and Nevada Public Health
agencies for participating in this study.
Information Availability and Changing
Conditions
“…this day we may make this decision and then as
we get new and more and updated information then
within an hour or two we may make another
decision.”
Acknowledgements
Information Overload
“Well you walk into a situation and there is a pile of
50,000 pieces of data… is very difficult to make sure
that you identify the pertinent data out of that.”
Laverne Alves Snow, MPA
[email protected]
RWJF/ National Library of Medicine
Training grant # LM007124
Contact Information