Discussion:
Research Design
Assume that you are going to conduct an experimental study related to your research question. What statistical procedures would you need to use to accept or deny your hypotheses? Support your answer with references to statistical manuals. You will need to convince the professor that your choices are the best and that you know what you are talking about. The Kanji book will be a great place to begin. This should probably not be more than three to four pages in length.
Be sure that you provide an appropriate statistical test for each of your hypotheses
Statistical Tests Sample By Robert Ward
In Partial Fulfillment Of The Requirements For BA 595
In order to answer the "Should we implement an electronic medical record system?" it is necessary to test each of the hypotheses.
The following are hypothesized and the chosen tests are included.
1. The new system will enable them to search the patient history more efficiently.
To test this hypothesis, a record will be kept of the time that each employee spent searching patient history for six months during the search for information on a possible system. This data will be averaged across all employees to produce a mean search time. This will be compared with the search time required for six months and twelve months after the system was installed. A Z-Test will be used to compare the means. (Kanji, 1999)
2. The new system will save money in the long run.
Comparison of records for the six months and twelve months prior to implementation will be made with the records for six months and twelve months after the installation. Data will be calculated on the total cost of ownership of the paper based system with the total cost of ownership of the electronic system.
3. The new system will prevent inadvertent errors of patient care.
Quality assurance records for the twelve months prior to implementation will be compared with the twelve months following to establish a mean error rate. These will be evaluated using a Z-Test will be used to compare the means. (Kanji, 1999)
4. The current staff will be able to learn the new system quickly.
The staff will have been assumed to have learned the system when they spend no more than the same time managing patient records with the new system than with the old. By using a log of time spent in patient record keeping, it will be possible to determine when the mean time to manage patient records is equal. The comparison will be for a period of one year prior to implementation to one year following implementation. A simple mean will suffice for this hypothesis.
5. The new system will enable the practice to file medical claims more accurately.
A comparison of the Quality Assurance data for one year before implementation until one year after will indicate the mean error rate on medical claims. By using a Z-Test, we can determine the significance of any difference in the means. (Kanji, 1999)
6. The new system will enable the practice to file medical claims more easily.
A simple survey will be administered to claims filers one year prior to implementation and one year following implementation. A mean value will be computed for the before and after surveys. These will be evaluated using a Z-Test on the means. (Kanji, 1999)
For each of the tests, a confidence level of 95% will suffice to rule out the null hypothesis.