Rapid Risk Assessment

Pablo Gomez

Outline

• Introduction
  • What is rapid risk assessment?
  • Integrating data from different sources
  • Sensitivity analysis
• Risk assessment in R
• Hands on practice
• More example

Posit cloud

https://posit.cloud

Introduction

Risk assessment

Rapid Risk Assessment?

Risk assessment is such a fundamental activity that a lot of agencies benefit from to inform their decision making process. In the context of animal health, risk assessment inform a policy regarding imports of animal products. There is plenty of documentation of how to do risk assessment and there are many different methodologies to implement it

The overall process of risk assessment can take several months. When we talk about Rapid risk assessment (RRA) we just refer to an accelerated process in which the In the absence of observational data or data collection, RRA leverages the lack of information with literature and expert opinion.

These are usually implemented in the initial stages of an adverse public health event, wheras formal risk assessments are produced at a later stage of an event, usually when more time and information is available.

In the lack of information to conduct a torough risk assessment,

The main objective is to develop an operational tool to facilitate rapid risk assessments for communicable disease incidents, drawing on the systematic methods used in evidence-based medicine or evidence-based practice where possible.

Integrating data from multiple sources

Observational/Experimental data

Fitting distributions

Summary statistics (min, max, mean, median)

Literature

Expert Opinion

• Focus groups
• Interviews
• Surveys
• Expert panels

Expert opinion

Expert opinion is a great alternative when there is a lack of evidence or time. Expert opinioin is also a great way to incorporate the opinions and concerns of stakeholders. There are multiple ways this can be incorporated into the data collection process.

Sensitivity Analysis

Current challenges with RA

One of the main challenges on the implementation, is that a risk assessment can take several months to perform. Outside academia, there is pretty much a monopoly on the tools used to develop the models. Lumivero, previously known as palisade, is the company that sells @Risk, which is a excel plug in to perform risk assessment. This plug in is very user friendly, most of people already know exel, so provides a realatively decent solution for performing risk assessment. The main issue with this, is that the price of the license can be pretty inaccesible to individuals and even to some health departments. There are some alternatives to this, but most of them rely on programming languages, which can represent a barrier.

quantrra: Risk assessment in R

To provide an alternative to this, we developed a platform specifically to implement risk assessment. The platform is implemented following the recommendations of the OIE and FAO manuals, and it is implemented in the form of an R package and an interactive user interface.

quantrra

# Sample from distributions
ra_sample(x = 'pert(0.01,  0.03, 0.64)', n = 10)

 [1] 0.11286461 0.06852999 0.30571981 0.23130342 0.05377351 0.10843924
 [7] 0.03413298 0.19453170 0.15181671 0.19920869

quantrra

ra_sample(x = 'pert(0.01,  0.03, 0.64)', n = 10e3) %>% 
  ra_plot_dist() # plot the distribution

quantrra

This platform facilitates to the user the translation of conceptual decision trees, into quantitative risk assessment models.

quantrra

# define a model table using one of the provided examples
m <- quantrra::OIRSA

ra_run(m = m, nsim = 10) # Run the model 

          P1         P2        P3        R1        R2        H1       H2
1  0.5213353 0.12729131 0.1744331 0.6353901 0.5117023 100.44826 16.08468
2  0.4182868 0.16656444 0.2498435 0.7495283 0.4372456  72.85584 34.56728
3  0.4910226 0.42013713 0.2590504 0.6077976 0.2958559  22.46179 32.95413
4  0.3355474 0.32046276 0.2244519 0.7015109 0.6258329  48.78371 40.31443
5  0.4305247 0.18676188 0.2873518 0.4876877 0.8201492  22.98473 11.00778
6  0.5902018 0.18723737 0.4043745 0.7484689 0.6104229  94.44883 12.21751
7  0.5398169 0.13835340 0.4702466 0.6481139 0.6234663  75.15061 37.95840
8  0.5532345 0.16148946 0.4261390 0.5585004 0.6990359 105.72870 42.26904
9  0.3448274 0.05702125 0.1805186 0.3849002 0.7271356  28.96834 43.93552
10 0.6737142 0.35788770 0.3772803 0.6353009 0.6151513  23.03876 23.06493
            P4        R3            X         Z          P
1  0.011575634 0.3251306 0.0037635926 116.53295 0.43858254
2  0.017407023 0.3277279 0.0057047675 107.42312 0.61282395
3  0.053441276 0.1798205 0.0096098382  55.41592 0.53253800
4  0.024135411 0.4390287 0.0105961369  89.09814 0.94409604
5  0.023104692 0.3999766 0.0092413374  33.99251 0.31413630
6  0.044686553 0.4568826 0.0204165074 106.66633 2.17775400
7  0.035120602 0.4040772 0.0141914331 113.10902 1.60517902
8  0.038071921 0.3904118 0.0148637285 147.99774 2.19979824
9  0.003549445 0.2798746 0.0009933995  72.90386 0.07242266
10 0.090967564 0.3908061 0.0355506831  46.10369 1.63901777

RA: Example

Demo

• Explain the problem with risk assessment
• How do we address this problem
• Examples:
  • ASF Carolina
  • ASF US
  • FMD Ecuador

Example of ASF in Spain

Few months ago we had a collaboration with the Visavet group from the Universidad complutense de Madrid, where we used QuantRRA to implement a risk assessment model to estimate the risk of introduction of ASF into spain via legal import of swine products This is the conceptual model that the student developed, Then we implemented it on QuantRRA to estimate the risk and we compared the results of QuantRRA with the ones obtained using the benchmark software @Risk. Most of the information used in the model was from literature and data regarding the imports of products into the country

Recap of some R concepts

R Studio

Comments

COMMENT AS MUCH AS POSSIBLE!

# This is a comment in R it will be only for the user
This is not a comment and will cause an error

What is the difference between line 1 and 2?

YES! the # character will make everything after it a comment in that line of code

10 + 10 # Everything after will be a comment
7 + 4

Comments are probably the most important feature of programming languages, commenting your code is not only for you (and feature you), it’s also to explain other people what you did and make it easier for them to replicate

Operators

Operators are characters with a specific function in R for example

3 + 3 # this is a sum operator

[1] 6

3 - 2 # this is a subtract operator

[1] 1

4 * 4 # This is a multiplication

[1] 16

Later we will see other kind of operators, but… DONT STRESS about learning everything.

Objects

Objects in R are containers for information, we can create objects with any names we want that start with a letter

myNumber <- 4
myResult <- myNumber * 5

Does anyone notice what operators are we using in those lines of code?

Hands on practice

More examples

Animal health

Food safety

Multilevel architecture

Multilevel architecture

Questions?

jpgo@ucdavis.edu