i'd argue "think like a patient" and be scrappy, tenacious, seek dozens of opinions from different schools of thought and be laser focused on a tangible (business value or human health) outcome.... not IT concepts at all<p>security - do you know how many unencrypted phi emails i get from clinic staff despite all the expensive encrypted email and secure web tools? dozens a day. want to know where your doc really keeps your "shadow" emr record? not in epic - dropbox or google docs if you are lucky - a paper print out with handwritten notes behind their closet door most likely.<p>integration - even if you wanted to integrate with epic/allscripts/ge you need to sign contracts and pay through the nose --- just try licensing HL7 or FDB --- and even after you do, guess what - the clinician prints the record to pdf and emails it to their google account so now you have forked copies of the data anyway. oh and it gets stale really quick. and is anyone actually checking the accuracy of those meds in the epic record? especially when it forces the user to select a dosage even when the patient has no clue?<p>scalabilty - getting docs in an OR to agree on a simple thing is like herding cats...let alone departments...let alone campus leadership...let alone systemwide initiatives...state IHEs...fed registries...oh my! look at IHEs...10+ years of Frameworks(tm) and the best candidate i see is SMTP based...yep, email. definitely scalable! so i concede that argument (though i guess i would say that anything that is not email or maybe simple web pages is doomed to small scale use regardless of architecture)<p>so on the positive side, my suggestions from 8 years in healhtcare it markets: deliver value first, tech afterwards. streamline something, save time or money, save lives, substitute actions for inactivity, show data and peer review empirical evidence...and above all... let the data speak for itself.<p>(it also doesnt hurt to have deep pockets and a 10 year time horizon either)<p>your results may vary