HerdPlus: SeH for optimised sampling strategy
Calculate optimised sample size and cut-point and resulting
herd sensitivity (SeH) and herd specificity (SpH) for a specified
list of herd sizes, test sensitivity, test specificity and design prevalence.
Animal-level design prevalence can be specified as a fixed proportion, or as the equivalent of one animal per herd.
Sample size and cut-point are constrained to provide SeH less than or equal to
a specified maximum value and SpH greater than a specified minimum value.
Sample sizes are also constrained to be no more than the specified
value corresponding to each herd size.
A meta-simulation is also run, using the optimised SeH and SpH for each
herd size, to calculate the number of herds required to be tested to achieve a
specfied target system sensitivity (SSe). SSe achieved and the number of animals to be
tested from a specified number of herds selected at random from the population
are also calculated.
- Test sensitivity and specificity;
- Maximum desired SeH and minimum SpH;
- Mode of setting design prevalence, either as a specified value or as a proportion of herd size;
- Animal-level design prevalence (only required if option for specified design prevalence selected);
- Herd-level design prevalence;
- Target number of herds to be tested from population;
- Target system sensitivity (SSe) for population testing scheme;
- Number of iterations per simulation and number of simulations for meta-simulation of population testing;
- The desired precision of results (number of digits to be displayed after the decimal point); and
- Pasted data comprising columns for herd size and corresponding maximum sample size and counts or relative frequency of each herd size in the population. Include a row of column headings in the data.
Note:Large numbers of meta-simulations generally provide only minimal
additional benefit and may result in a system crash if used in combination with large numbers of iterations. A maximum of 100 meta-simulations is usually ample.
Results are calculated using the HerdPlus method (Greiner 2006?) and include:
- Table of optimised sample sizes and cut-points and corresponding SeH and SpH values for each specified herd size;
- Plot of SeH and SpH values for optimised sample size and cut-point against herd size;
- Summary table of meta-simulation results including median and 5th/95th percentiles of the numbers of herds required to achieve the target SSe and the achieved SSe and number of animals tested under the optimised scheme for the specified number of herds tested; and
- Detailed results of meta-simulations.
Developed by Matthias Greiner and Evan Sergeant