Chapter 18 Sample Allocation Model (SAM)

IN WAREHOUSE DEVELOPMENT

18.1 What is the Sample Allocation Model (SAM)?

The Sample Allocation Model (SAM) integrates cost with the probability of disease introduction to pinpoint the best surveillance strategy across sub-administrative areas not known to harbor CWD. The SAM framework consists of two main components: posterior estimation and optimization. Posterior estimation is used to answer Question 1 while optimization is used to answer Questions 2-3 (below).

18.2 What Questions Does it Answer?

Question 1. What is the unseen infection state given one, two, or three years of historical sampling without finding a positive case? SAM considers prior years of sampling in conjunction with the unobserved spread of CWD to determine the probability that any given sub-administration area is presently disease-free. Results reveal sub-administrative areas that are ‘blind spots’ in disease surveillance, where insufficient testing has occurred and therefore CWD could be present and spreading silently. Alternatively, results can pinpoint sub-administrative areas where sampling has been sufficient through time such that CWD would have been detected by now, if it were present. [Mode 1]

Note: If you wish to compute the probability of disease freedom while explicitly considering the clustering tendancies of hosts, use the Probability of Disease Freedom Using Clustering Model.

Question 2. Given the answer to Question 1, where should I look for CWD this coming year, and how much effort should I use given my agency’s CWD surveillance budget? SAM considers the underlying status of disease in conjunction with introduction risk to strategically allocate samples in each sub-administrative area across an entire jurisdiction. Resulting allocation constitutes the ‘optimal control’ given a fixed surveillance budget. Optimal control means that the silent spread of CWD across the jurisdiction is minimized at first detection, given the capped amount of sampling dollars available to make that first detection. [Mode 2]

Question 3. Given the answer to Questions 1-2, what budget can improve this year’s surveillance program (i.e., reduce silent spread up to the moment of first detection)? SAM considers how optimal control may be improved with an increase in budget. Possible improvements include higher budget to achieve earlier detection or more strategic allocation relative to current practices to minimize cost while maintaining performance. Results of this ‘cost analysis’ are helpful in requesting the annual budget necessary to bolster your agency’s surveillance program. [Mode 3]

Note: SAM does not produce surveillance targets based on statistical confidence. If you wish to determine sample sizes to reach statistical assurance that CWD is absent in an area, use the Sample Size Quotas Using Clustering Model or the Efficient Sample Size Calculator.

18.3 Output Details

  • A map containing the probability that each sub-administrative area is disease-free given the look-back period of historical sampling.

  • A map containing the intensity of sampling needed in the coming season-year in each sub-administrative area to achieve optimal control.

  • A graph comparing optimal control under current and ideal budgets.

18.4 Abbreviated Tutorial

  1. Run the Sample Allocation Model (SAM) from the CWD Data Warehouse.
  2. Explore disease status (Mode 1).
  3. Explore disease status and obtain surveillance targets (Mode 2).
  4. Explore disease status, obtain surveillance targets, and run the cost analysis (Mode 3).
  5. Explore the model logs, input file, and output files used in the run.
  6. If the model did not run, check the model logs to understand required data that was missing.

18.5 Parameters Needed to Execute the Model

  • Model type: Select ‘Sample Allocation Model (SAM)’ from the drop-down list.

  • Reference name: Label the run.

  • (Optional) Applicable season-year: Label the season-year of the run to assist in documentation.

  • (Optional) Notes: Enter any remarks about the run.

  • Mode: SAM can be used for several sequential questions, which govern runtime (the time needed for the computer to run the model).

    Select Mode 1 to: Determine the posterior infection state only (i.e., to answer Question 1). [Shortest runtime]

    Select Mode 2 to: Determine the posterior infection state then allocate samples (i.e., to answer Question 2). [Moderate runtime]

    Select Mode 3 to: Determine the posterior infection state, allocate samples, then conduct a cost analysis to further improve optimal control (i.e., to answer question 3). [Longest runtime]

  • Total budget: Enter the total annual budget that can be spent on CWD surveillance across the administrative area in the upcoming season-year.

  • Look-back period: SAM can be used to compute probability of disease freedom using 1, 2, or 3 prior years of historical sampling data.

    Enter a look-back period of 0 to: Proceed by assuming all sub-administrative areas are disease free.

    Enter a look-back period of 1 to: Compute the probability of disease-freedom based on the prior year of sampling data. Use this option if introduction risk changed in any sub-administration area last year.

    Enter a look-back period of 2 to: Compute the probability of disease-freedom based on the two prior years of sampling data. Use this option if introduction risk changed in any sub-administration area two years ago.

    Enter a look-back period of 3 to: Compute the probability of disease-freedom based on the three prior years of sampling data. Use this option if introduction risk changed in any sub-administration area three years ago.

Note: Look-back period is limited to 3 historical years because SAM assumes that introduction risk is static. Changes in introduction risk include new CWD outbreaks in neighboring areas, additional avenues of anthropogenic prion introduction, etc.

  • Maximum harvest capacity: Enter harvest capacities specific to each sub-administrative area. If unspecified, SAM will use the same threshold of 1000 for all sub-administrative areas.

  • Annual growth rate: Enter the rate that governs the annual increase in CWD prevalence (once established) in a region. The default value (0.2) reflects estimates that prevalence rises from 0.5% to 1% over approximately five to seven years.

18.6 Details on the Theory

Wang J, Hanley B, Thompson N, Gong Y, Walsh D, Huang Y, Gonzalez-Crespo C, Booth J, Caudell J, Miller L, Schuler K. Strategic Allocation of Surveillance and Prevention Resources for Emerging Wildlife Disease. In peer review.

18.7 Code

The code is publicly available at https://github.com/Cornell-Wildlife-Health-Lab/sample-allocation-model.