Overall do you envision that t

This is a very subjective question and there is never a one sizefits all answer to this discussion. If we ask whether statistics isnecessary as a subject, the answer is it depends on the context weare talking about.

For healthcare professionals it becomesnecessary to understand basic statistical conceptsand to be able to appraise scientific papers to see whether theyanswer the research problem in question, and to assess therobustness of results presented. Statistics isinteresting to see how results compare and lookingat the strength of the relationship (if any) that exists betweenthe two outcomes. Also they play a significant role in systematicreviews and meta-analyses of subjects.

At its simplest level, statistics is about summarising(descriptive statistics) and presenting data (inferentialstatistics) in ways that accurately reflect and convey theirmeaning. The next level is hypothesis-testing – the process ofsystematically answering a research question of importance. Orvalidating a status quo with an alternative.

Basic examples of issues important to healthcare professionalsand medical students include trying to ascertain whether a newlylaunched treatment or a drug treated to patient is better than anexisting drug or whether a possible risk factor.

For example, low birth weight, is associated with a particularoutcome, such as the likelihood of developing asthma. The essenceof this is to begin by assuming that the treatment or risk factorhas no effect; this is sometimes referred to as the nullhypothesis. The likelihood that any difference between groups hasarisen by chance is then calculated.

The level of difficulty depends on theexplainability of the system. Statistical jargon can sometimes beconfusing. Most books and research on statistics are written forthe specialist with masters in statistics, often being full offormulae and jargon. This is unfortunate because computer programsfor performing statistical tests are readily available, removingthe need to perform long, complicated calculations. The mostimportant part is the interpretability of the models and use it forrelevant test cases. Concepts that often perplex non-researchersshould be presented in a plain, straightforward way.

As a personal choice I would choose statistics as a subject asit gives an analytical approach to problem solving thus keepingirrationality out in our analyses.