PDHA Model User's Guide

Ben Saylor
Institute of Social and Economic Research
University of Alaska Anchorage
August 30, 2007

Introduction

This document describes how to run the PDHA village-level simulation model, which is a discrete event model. It individually models each person in a simulated population of a large village (about 1000) over a period of 20 simulation years. As time progresses, an individual's decay rate, which may be reduced by any care they have received, determines the frequency of periodic occurrances of new decay. Separate counts of decayed, missing, and filled teeth are tracked for each individual. When an individual makes the decision to seek care, it begins a cycle through the flowchart (see the Technical Appendix), in most cases receives care, and then after some amount of time returns to the starting point to begin the cycle again.

The model is, by necessity, a greatly simplified picture of reality, and does not take into account every possible factor that could affect the oral health of a population. Its purpose is to aid in developing an understanding of the differences between different staffing and policy scenarios in terms of clinical outcomes. The numbers it produces as output should not be taken as accurate quantitative predictions, but serve to give a general sense of how outcomes differ among possible scenarios.

Requirements

Using the Model

The model is a standalone Java application, but reads input from and writes output to Excel spreadsheets. It consists of the following files which should be contained in the same folder:

Setting the Input Parameters

The model has about 70 user-adjustable parameters. These are contained in the Excel file DHAmodel_input.xls in the worksheet named "Global". Each parameter is on a single row, with its name in column A (the names should not be changed), its value in column B, and a short description in column C. The default values have been carefully chosen, but any can be changed. For example, to run the simulation with no DHA in the village, set the value of the parameter "PDHAenterYear" to a value greater than the "stopTime".

The majority of the parameters describe the benefits received by patients from various activities and treatments. The names of these parameters begin with "bft_". Most of these describe reductions to the rate of tooth decay (abbreviated "DRR" for "decay rate reduction" in the description field) and are expressed as a percentage.

It is not necessary to alter the other worksheets in DHAmodel_input.xls (they contain village population by age and birth, death, and migration rates).

Running the Simulation

Before running the model, save any changes to the input file and close it. If the input file is open when the model is running, the parameters may not be read correctly.

To run the simulation, double-click on the file DHAmodel.bat. Two windows should appear on the screen: a black command window, and a window showing the model animation, which looks like this:

DHA model screenshot

To quit the simulation before completion, close the command window; until this window is closed (or stopTime is reached) the model will continue to run.

The animation shows the current state of the model as it runs.

At the top of the window are the simulation controls - the play, pause, and restart buttons, the speed slider, and the settings button. The simulation will automatically play when started using the DHAmodel.bat file and can be paused and resumed at any time. Restarting the simulation has not been thoroughly tested and is not recommended. The speed slider controls the ratio of simulation time units (in this case, years) to real seconds. The computer running the model may not be able to run it as fast as the desired ratio, and in this case, it will simply run it as fast as possible. The "VT" (virtual time) setting at the rightmost end of the slider causes the simulation to run as fast as possible.

Below the simulation controls is a graphic showing the current oral health profile of the population by age. The topmost portion is a scatterplot in which each simulated individual in the population is represented as a dot whose x coordinate represents its age in years, and whose y coordinate represents its oral health by some measure. The green box to the right allows the user to select the outcome measure, which can be DMFT, Filled & Sound Teeth, T-Health, or Caries-Free (defined as 1 when DMFT = 0, 0 when DMFT > 0). The check box labeled "Show Scatterplot" below the green box toggles the display of the scatterplot. It is unchecked by default, and the simulation runs somewhat faster with the scatterplot disabled.

Below the scatterplot is a set of green bars representing the average of the selected outcome measure for all individuals currently in each age group (0-4, 5-9, 10-14, 15-19, 20-49, and 50-100). A number above each bar shows this average value. In the case of the Caries-Free measure, the bars show the percentage of the population in each age group who are caries-free.

To the right of the bar chart is a box showing the percentage of children 6 and under with ECC (early childhood caries) or Severe ECC. In this model, a child with a nonzero value for deft is considered to have ECC. Children under two years of age with non-zero deft are considered to have severe ECC. Children from 2 to 6 with a deft value greater than 5 are considered to have severe ECC. Any child with severe ECC is also counted toward the Severe ECC percentage.

Below the green bars are blue bars showing the average percentage reduction to decay rate experienced by the population in each age group. These bars extend downward - more blue means a greater reduction to decay rate. A 100% reduction corresponds to a zero chance for new decay, while a 0% reduction corresponds to the baseline decay rate with no intervention.

The rest of the window displays the values of various model state variables.

The current simulation time is displayed in the lower right corner of the animation window. The value is in years since 2000. When this timer reaches the value set by the stopTime parameter (20 by default) the simulation will stop, the output file will be written, and the animation window and command window will close.

Interpreting the Output File

The output file will only be generated when the model has been run to completion. The presence of an existing output file is required for subsequent output to be generated; please do not delete the file.

The "noDHA" worksheet in the output file contains the model output for the "no DHA" scenario. The model output is written to this worksheet if and only if the PDHAenterYear parameter has been set to a value greater than stopTime (i.e. the PDHA never enters the village). Otherwise, the output is written to the next worksheet, called simply "DHA". When the output is written to one of these sheets, output previously written to the other is preserved. This allows comparison of the "DHA" and "no DHA" scenarios. These two worksheets contain time series data of average DMFT and F&ST for two age cohorts: people who were 0-5 years old at time 0, and people who were 20-25 years old at time 0. These series are graphed in the worksheets "DMFT chart" and "FST chart".

The "kids" worksheet contains time series data for children 5-18, this time not following the same people through the entire simulation, but keeping the age bracket constant. The measures included are percentage caries-free, percentage who have received clinic-based care within the past year, percentage who have received outreach-based care in the past year, SiC (significant caries index), percent who received toothbrushing prophylaxis in the past year, and percent who received root planing in the past year. The "adults" worksheet contains the same measures for ages 18 and over. Most of these series are graphed in the preceeding worksheets.

In all of these series, the column marked x is the simulation time, and the column marked y is the the measure being recorded. Data points are recorded at quarterly increments of time.