Facility Level Approaches To Infection Control Engineering FLAT ICE

1
Facility Level Approaches To Infection Control
Engineering(FLAT ICE)
2
The problem

• We confront a threatening array of pathogens
• Healthcare facilities amplify dissemination
• We have limited resources for containment
• We lack local guidance for containment

3
QuestionWe have a nephrology clinic in
downtown Minneapolis with 3 providers who each
see 20 patients per day. We have little time and
scarce resources, and are worried about the
rising prevalence of MRSA in our patients. What
can we do to protect our patients?
4
QuestionOur practice has a small network of 5
clinics in Southwestern Pennsylvania. Im told
there is a bad flu going around. Should we
designate one of our clinics as a flu clinic?
Is there anything else we can do?
5
QuestionWeve developed an exposure and
symptom driven, points based screening instrument
for influenza that can be implemented over the
telephone. What balance between sensitivity and
specificity should we use to minimize
dissemination in our clinics?
6
We will deliver software and analyses to help
local providers answer important questions such as

• What facility measures are most reasonable for
  me?
• How should high risk individuals be scheduled?
• How should I use pre-clinic screening
  instruments?

7
Our goal is to facilitate development of
infection control policies that are tailored to
local needs and resources

• During seasonal outbreaks
• In the face of endemic resistant organisms
• When confronted by novel pathogens

8
Clinical assessment, diagnostic testing, and
feasible containment options are interrelated
9
We address these interlocking aspects of
infection control using tools that are

• Monte Carlo realizations of Markov processes
• Temporally and spatially explicit
• Easily applied to different queue systems

10
Our analyses are based on the probabilities of
very simple events
11
Each of these probabilities is readily measured
in real world settings
12
We focus on practical, resource sparing
containment strategies
13
We build statistical tools that convert
pathogen characteristics and facility policies
into predicted dissemination rates
14
The anticipated effects of different containment
strategies on dissemination can be compared
(Computations adapted from Hotchkiss, Strike, and
Crooke, Emerging Infectious Diseases)
15
The resource costs of different containment
strategies can be examined
(Computations adapted from Hotchkiss, Strike, and
Crooke, Emerging Infectious Diseases)
16
The tools we will provide can inform local level
decisions regarding

• Resource-sparing containment strategies
• Nuanced responses to varying threat levels
• Simple, novel, and effective interventions
• Selection of diagnostic thresholds for screening
  tests

17
What we propose to do

• Calibrate models with prospectively collected
  data
• Observational data from medical clinics
• Data from mock caregiver/patient encounters
• Develop user-friendly software packages
• Generic prediction libraries for non-expert use
• User configurable software for expert use
• Identify optimal thresholds for diagnostic tests
• Incorporate more sophisticated biological details

18
What is a generic prediction library?A
digital database in which a computationally naive
provider can easily cross reference clinic
policies, pathogen characteristics, and
prevalence to identify potentially appealing
containment strategies
19
What is User configurable software?A software
package that allows individuals to construct
decision support tools that are fine tuned to
address specific facilities or networks of
facilities
20
Timeline

• Initial months
• Broad-based, generic recommendations
• Systematic investigation of diagnostic thresholds
  
• User-configurable software, Web distribution
• Begin to collect calibration data on specific
  pathogens
• Formative months to follow
• Calibrate and test models for specific pathogens
• Viral pathogens (Influenza, rotavirus, other)
• Bacterial pathogens (MRSA, VRE, C. difficile,
  other)
• Computational refinement
• Incorporate uncertainty in pathogen
  tranmissibility
• Address uncertainty in pathogen persistence
• Integrate with large-scale population models