Excel Statistical data mining of 200+ independent variables, 3 dependent ones

Determining which of my 6+years of daily tracked 200+ lifestyle variables (exercise, treatments, supplements, liquids, food, prescription drugs) affects my cancerousness, body fat, and blood pressure.
Cancerousness is measured by 4 molecular cancer markers (max. 46 for one, over 6 years) at sporadic (mainly monthly) intervals, with a possibility of there being daily surrogates of these amongst my body composition/urine biochemistry metrics (see http://www.bladdercancerfight.blogspot.com/#!http://bladdercancerfight.blogspot.com/2013/12/my-potential-cancer-marker-surrogates.html)
Body fat is measured daily by a Tanita InnerScan body composition monitor (6 years, 2,000+ records)
Blood pressure measured daily, 2+ years, 750+, average of 3 over 10’ each time.
I presently have used a set of Statistica multiple regression analyses, but feel this has given unsatisfactory results. However, a more expert eye would provide a useful evaluation of this for me. If I am right, then this alone may be enough. But I suspect a time-series multiple regression analysis or entirely different approach (AI?) would be better.
So I need someone who is already expert in multiple regression analysis and alternatives which may be better. It is probable that an expert in SPSS or similar (Python?) would be needed.
I am in the UK; can be contacted via Skype (Zarathrustra) for an initial chat and display of my Excel spreadsheet. My work can be seen at Http://www.bladdercancerfight.blogspot.com/