This was a quasi-experimental design. This type of research is meant to find a cause and effect relationship when there are independent and dependent variables. The goal is to find out if a specific intervention affected the outcome (Burns & Grove, 2007). In this particular case, the goal would be to find out if the $25 incentive improved the number of people at risk for HIV that went for testing and follow up. The study is an experiment as there is a control group used to measure against.
There were 372 patients referred during the three periods that were tracked. During the control periods in which the incentive was not offered, 20 of 252(8%) patients completed HIV counseling and testing while during the incentive period 27 of the 120 (23%) patients completed testing and counseling. There was a statistically significant difference in the groups that showed in race and ethnicity. This had to be accounted for by using a multivariable logistic regression model. The result was that there was still a statistical difference in the two groups. The conclusion from this was that there is a increased proportion of patients who complete testing and counseling when a small financial incentive is used. However, there was not a significant answer as to whether there were more positive patients found in this manner and whether this program would be worth its cost. There will need to be further clinical studies to determine these answers.
The study was performed over three consecutive 6 month periods resulting in two control groups and an experimental group. This might have been a better experiment if it had been conducted over the full year and 1/2 randomly assigning patients to the different groups. In setting up the groups the way they did for this particular clinical experiment, they did not realize their bias in one group toward one ethnic group until it was too late. Had they randomly assigned