Many statistics analysis have been found in many clinical researches,some may undermine clinical significance of some research investigation and other may over-emphasize the outcome .How do we draw a line to how much impact statistic analysis would add clinically to patient outcome and community health care?
Each research design has a corresponding appropriate statistical test for data analysis.What creates the confusion is not selecting the most appropriate tests. Some good journals will help to minimise this by having a in-house statistician contribute to manuscript review.
A statistical program in my opinion is just a tool that delivers at the behest of the user In itself it has no power except that which you give it. It cannot unilaterally alter the results of any study. Errors occur when the analysts uses the wrong tool for the right data. Secondly regardless of the result from data analysis data interpretation still lies squarely at the foot of the researcher. In essence a good knowledge of the different statistical tools and their appropraiteness is key to great research.
It is very important how your results agree with other statistical data existing in the field considered. You should compare your results with the results of such organizations as WHO, USAID, NIH, CDC and others. In other words, you should be sure that your research may be well integrated into the whole statistical picture of healthcare community.
Before we change practice or introduce a new health program, we need some evidence that the change will provide health benefits. Over-reliance on statistical significance was a problem until the last decade or so. The emphasis now is on confidence intervals rather than p-values, and on effect size. This latter especially helps balance statistical significance with clinical significance.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis