Ultrasonic medicine research design, basically the same as other clinical disciplines. Usually can be divided into two categories, one for prospective, one for retrospective. The former is the previous preliminary or not yet made the subject, to anticipate the study, such as the study of a new technology, a new treatment method for clinical diagnosis, treatment value,
The results can be of practical value, but also to draw the opposite conclusion. The latter is a retrospective summary evaluation of the work done in the past. The use of prospective studies is now taking on new practical significance.
In the prospective study of the subject, first of all, to set up the topic, and set up the topic must be new, it is best to check the new to determine the study to be done, whether others have been done, if many people have been done and has been a clear conclusion, they are only repeating the others of the project and no new significance. Of course, in the actual work found with the previous different results, this time in order to further in-depth study, can still set up the topic, the result may be a major breakthrough or invention. After the establishment of the topic, followed by a reasonable design of the subject, and absolutely must be rigorous.
After determining a specific research topic, the following points should be done and fixed.
Research subjects: to be constant. Including patients, control healthy people (volunteers), some kind of animal or other. In this column, it is best not to use the word "experimental" for human beings, and "experimental research" for animals or other categories. The age, sex, condition, etc. of the subjects should be consistent and fixed, especially in the patient group and the control group between the sex and age should be equivalent to the party.
Methodology: novelty. Tests and treatments should not duplicate everything used by others, and should be done by specialized personnel and machines. For example, skilled and new workers are not the same technical conditions, the results of the test will appear artificial differences. Another example of instrument performance is too different, the conclusion will have a certain difference, otherwise the conclusions are consistent, its credibility is worth considering.
Detection indicators: to be accurate. Choose objective indicators that are easy to observe and have a clear meaning. For example, to observe the time of fetal umbilical cord around the neck, preferably before delivery, if the time from delivery is longer, the results of the credibility is not as good as before delivery.
If the science is strong, the indicators of detection should have circumstantial evidence, such as detection and diagnosis of coronary artery stenosis, it is best to have a coronary angiography results of comparison. Diagnosis of malignant tumors of an abdominal organ should be confirmed by surgical pathology or needle aspiration cytology. These should be arranged in the research design.
Results of the study: scientific. It should be studied and observed according to the purpose of scientific research design, and the results should be concluded with statistical objective data as the best results of the research on this topic.
Statistical methods: the establishment of a control group: to be similar conditions. In order to make the results more convincing, there should be a control group, through the results of the experimental group and the control group comparison, distinguish the treatment factors and non-treatment factors on the research results of the gap. Treatment factors include testing of patients, treatment methods, dosage, etc., and non-treatment factors include social and environmental factors. The results of non-treatment factors are more scientifically valid when they are relatively consistent between the two groups, for example, when the age and gender of the two groups are the same, and when the testing period is the same, such as during a certain season.
Randomization: be objective. Randomization is the study of two different methods of testing, treatment, medication, etc. on the test or treatment results of the observation. Randomization is a method of maintaining a relatively balanced experimental and control group, i.e., applying lots, drawing lots, etc. This is not picking according to subjective wishes, but the subjects under study are randomly selected from the total, i.e. each subject has the same chance of being drawn.
Sample size: be large. The larger the sample size, the greater its reflection of the objective reality. Case group and control group samples are best in more than 30 cases or more. Such as common cases in the observation of drug efficacy, it is best