 Hi, everybody, we're back. This is Dave Vellante. I'm with wikibond.org and my co-host Paul Gillin and I have been going all day today. We're here at MIT at the Information Quality Symposium. It's a symposium that comprises the root, really, of the event, which was the Chief Data Officer forum that occurred on Tuesday, and a number of information quality practitioners. We're here at the Tank Center at MIT. It's a very intimate event, a lot of one-on-one conversations, a lot of hanging out, a lot of breakout sessions, keynotes and the like in the auditorium. And we've been covering it. This is the Cube. We go out to these events. We extract the signal from the noise. We bring you the best guests. Naim Hashmi is here. He's a chief research officer for information frameworks. He's involved in the MIT Sloan School CIO symposium, does a lot of CIO activity, focuses on analytics and big data, which is something that we're going to be talking about. Naim, welcome to the Cube. Thanks for coming on. Thank you, my pleasure. OK, so tell us, you have your hands and a lot of pies. So tell us a little bit about yourself and what you do. Right. Basically, I'm a nuclear physicist and health physicist. It's a weird combination. But having engaged with a lot in the industry has worked for digital for a long time, and doing a lot of innovative stuff, always been in the area where how we are going to manage information in the future. So looking forward 10 years, 20 years down the road, build the architectures and reference models, and then work with the vendors as well as the IT organizations how to roll out that kind of vision. So I'm sort of a vision type of person. So I started this after living digital. I started this business. And a lot of work has been publications. So I wrote about author four books. I'm working on the fifth book. And this fifth book is really interesting. And it talks about informatics and analytics, design strategies. The way I see it is in the health care today, there is a lot of talk about informatics and analytics. Well, you can define, I don't mean to interrupt you, but define informatics, because that's a word that's been around a long time. I'm not sure many people know what it means. Correct. It's very misleading. In fact, if you think of any kind of analysis in the health care business, it was classified as informatics. But it's really not informatics in that term. It's a lot to do with how if you look up, aside for the informatics world, it's informatics and ICS ticks, like analytics, analysis size, all of something. So anytime you see the informatics, it's informatics about something. So this informatics is, so when I start to write this book, I have to really spend a lot of time doing research. What's the difference between informatics and analytics? So the informatics have some attribute that you are basically discovering the information. You are synthesizing information. Where in the analytics world, you are analyzing the data. In the informatics world, you're also giving sort of a recommendations and the looking of some kind of pattern, especially in the health care world, disease or other area of planning. When you go back to the typical analytics world, there we're talking about very much of a structure data of qualitative and the predictive kind of modeling in that context. So that classifies. Plus also, it's a lot more of a informatics requires very powerful visualization type of piece because it's a very complex information, how you present that information, especially for the doctors. In this book, I also talk about how a clinician thinks and how a physician thinks. It's a very different approach. So where informatics really fit in from the physician perspective, where they're looking up some kind of pattern. So sometimes you're presenting the large information, set of information visually, and they can see certain type of trend is appearing. It was not intended to show that, but they could see this trend is going on. Where clinician, him or herself, they will be looking a lot more of an algorithmic. I need to run this ventilator. What steps I have to go through. So it's very much of a thought process is very different. And then comes up the analytics piece. How you optimize the clinical process? How you, if you want to, like in the ACU environment, you, how you do the profitability analysis? How you share the benefits? How you optimize this thing? How you, this is very interesting because we've been talking all day about healthcare and most of the conversation has been focused on electronic health records and gathering the data and simply getting the data into a database. You're talking about going beyond that to understand, to derive some greater meaning out of that data. How does that affect the quality of care? I mean, what kinds of analysis do you believe clinicians will be able to perform? As EHR goes mainstream, what kinds of, how will it improve their diagnostic ability, their treatments? There are several ways that actually we're just starting up that area. So it's, I would call it a virgin territory at this point. So a lot ahead of us, we will do a lot of learning, but I'll give you an example. Like just before, till last year, I was working at Forsenius Medical Care. I was their vice president for knowledge management. And when you talk about knowledge management, in this term, in this conference term, there's a lot more to do with exactly like what the CDO they're defining. So I'm responsible for all the data and the Forsenius is the world's largest dialysis company. So we treat about 60,000 patients daily and there are about 2,000 clinics throughout America and about 1,500, 1,600 individual EMR. And the patient moves from clinic to clinic. So how you make sure when the person comes to a clinic A, like sales person is today is getting treatment in Boston, tomorrow they're in New York, how the information have traveled. So we have, we build the architecture, that's I'll talk in the afternoon, use the big data platform to move all of this information from these clinical systems into this central hub and then analytics like the dose adjustment perspective. So we can see the latest result of the person and also look at the longitudinal query and build those analytics that how the dose should be adjusted. So when the clinician or physician comes up to give the treatment, they have the latest information and the recommendation because at night we run the algorithm. So when the patient gets the treatment, so they get the right input that this is the recommended value but up to the physician to really override because that's the decision, that's the regulatory thing. I cannot really automatically change the dosage. Someone have to sign. So what medications I administer, they will go back. So this is just one example. I could see the usage of Watson. Watson as a service. Yeah, I was just gonna say. It's a differential diagnosis because not every IT organization or the healthcare organization will have the capability or the resources to build that kind of knowledge service or the diagnostic service where they are looking up information from all the publication, published material, getting up all the from the practice information and doing variant tests. That will be part of, I will call again, informatics, not analytics, but a lot more do informatics. They are synthesizing the lot of information, giving some recommendation to the decision maker, in this case, physician, and they will decide what benefits and are, which treatment is good and are. So I think those kind of things will start to emerge more and more. And you're familiar with the Kaggle content, the crowd source. So where would you put that kind of activity? Do you know about this, Paul? No. So Kaggle is essentially the crowd source, the crowd source healthcare data and have a contest. I think the move winner actually is lucrative, the million dollars to the winner to develop, do predictive analytics and develop algorithms on which class of patients are gonna be most susceptible. Exactly. I think last year I was also one of the judge for Massachusetts Health Data Palooza. So they were 21 really innovators. They were presenting it. And the winner was one of the young lady from Harvard. The young entrepreneur started the firm and what she was doing was looking at, she took her own depression scenario and converted into the innovation. What they were, she mentioned that because of, she's taking the birth control, it was affecting their, it was causing depression. So she started a firm, take up all the medications data and look, scan through all the social networks, see how people are describing their depression, especially in the women and then map out what kind of medication birth control will be better for them. So to me, these kind of things are really coming around. The major challenge comes up is once you are getting this information from disparate systems, especially the healthcare, how you really make sure the proper context and the content is done properly. Healthcare is such a very specialty oriented. So term are very specific to one, what they call practice. It's the same term on the other one but have very different meaning. So those kind of things, it will require a very much of a semantic and what they call ontology properly, especially in the healthcare perspective. I think those things will be coming more and more and I see in the future, actually in 2003, I published one of the cover story in the Intelligent Enterprise Magazine and that was on the BI on sale, Business Intelligence on sale. And that was the same thing in the future when you're going on, it will be a lot more to do with few powerful services that are available and you subscribe to this part of the clinical process. So the clinical processes themselves have to be more of a customizable. Today's vendors, I don't think their systems are really flexible enough because to me, their systems are nothing more than changing their paper forms to electronic form. I call it forming without reforming. So not understanding how even you automate these things or electronize these things. It's a great point. So how do you make those things together? So essentially they're taking the same business process and they're pushing it to electronics and of course that business process was developed for paper. Right. And it was developed because paper's so inflexible. Correct. So they're taking that process and say, okay, let's make this electronic. Exactly. It's not really only the process level. Also the content, the paper, exactly the same form electronically and now in paper form, you can write and write, there's no validation. It's like when TV came along, right? Well the radio guy said, why would anybody want to watch a bunch of guys doing radio? That's true. That's true. So I think in my experience when I was at Forcinius, we learned that too. When we were implementing the EHRs, the migrating thousands one to the newer systems and it was inflexible. A lot of data quality issues were coming up because there's no validation and verification on this field because those fields are originally designed from the form and you do the same form up there and since there is no control there, what based on the rule is certain value comes up what should be the next one or valid one? So how should we think, I mean thinking of us as patients, how should we think differently about how our health information should be used in the future? As we get out of, as Dave mentioned, it's kind of paving the CalPath approach right now where you're just taking data and putting it into a different form. What you're talking about is it's going to be used differently and what should we expect as patients? I think to me there's a big responsibility on the patients and to me it's a healthcare itself is such a fragmented industry. Pares are fragmented, providers are fragmented and patient have their own ego. So once nobody's trying to work together, so they will be all with that kind of issue comes up. To me, I think patients also have to be open because there is, I think there's a concern for privacy but there's too much concern for privacy. Once that comes up, then this information sharing becomes an issue. But information sharing becomes an issue, you can't have a longitudinal view of the information. So I think some of the healthcare reforms they're coming up, they're going in the right direction. We need to learn a lot through that. I think there's a intent, intent is really right because sometimes when you do the meaningful use, once you know I'm going to the EHR giving incentive, that's a good move but it's still at the early stage to change the culture. Because again, it's not really only the treatment, physicians are taught differently, nurses are taught differently and they're always, they have to live because of regulated industry, they have to live within their own. Well, and it is cultural. Look at the backlash from prison. This is a manifestation of the country's opinion right now but you got one side and the other side with people defending it, people attacking it. And so privacy is one of those touches. We had Scott McNeely on the cube recently about a year ago now. There's no privacy, just get over it, which is kind of what you're saying. Maybe not that much. That's true. In fact, I went to this HEMS conference in New Orleans. I was mentored and also do their award and also the BI and the informatics track. When you enter in their hall, what they call exhibit hall, they have about thousands of vendor, 1,000 plus vendor and you can smell in the room when you open the door, smell of analytics. And everyone, and actually 90% of those analytics are not analytics. I wrote one paper as search for a health IT. I call them arithmetic, not even analytics. Analytics means? What do you call them? Arithmetics. It's analytics or arithmetic. And the CMS meaning for use, the major PQRS and all the way. When you think of, it's nothing more, it's not intelligent, it's not analytics. It's to do with what's the numerator and what's the denominator, give me the ratio. So it's simply arithmetic. I couldn't agree more. Right, it's the basic arithmetic. It's a basic arithmetic. We're going to make it pretty. All right. Not even pretty, just give the number. Not even more. Everybody, every vendor is going on the bench over the bandwagon, we're analytics. So that's why I wrote this series of those analytics. Analytics washing. All right. I call it the snake oil, you know. And actually they are moving more towards future, especially with the like Watson kind of thing. In the book I also will talk about how the translation informatics comes, translation informatics is basically coming from the bioinformatics and medical informatics and all the therapies and things after trial gets into the practice. Now when you go to the practice, because these trials and the effectivity was basically on a controlled trial basis, but not in a practical, when the patient comes, look at a lot of economic factor, a lot of disparities comes along, how a patient behavior come into play. That time the actual, what the translation of research came from the lab is implemented into the clinical practice. It doesn't really give the same results outcome than really what it is. That's what last year healthcare and administration people they have launched CER, Comparative Effective Research Project, meaning that look up how different clinical practices they do the outcome. And that's become also dealing with the social data, taking the feed from there and refining why a certain type of therapy really worked or not with this kind of community, with this area, with this kind of physician, with these kind of comorbid conditions, while the bioinformatics pure science was saying this, but actually it's this. So how you marry those together? And that's where you know some kind of Watson type of area, the technologies. And I know I have talked with few people when they're starting up that area. But on top of the other area, which is my interest these days is called psyche mining. Psyche mining. Psyche mining. Okay, what's that? I published that in 2004. And that psyche mining is how people think, why the things, what influences them, and especially in a cyberspace perspective, different part of the world, people think differently. The concept of reward is differently. The concept of respect is differently. So how you develop distributed autonomous agent technology that becomes like a virus and attaches to the mobile and on and on on both end of the communication or the end of the communication. And also senses, like if I'm making a gesture, because now there's all the mobile devices, it's sophisticated, the cameras and everything. So you're doing the gesture mining as well. You are seeing the object and you're creating some emotion, that brain waves kind of thing. And actually I had four PhD students, they were doing some research, I was advising them, I formed Center for Knowledge Engineering. It's an interesting concept, but I see that a lot into, especially when you're moving more towards the behavioral and mental health, that will be really key. And right now I'm working with six Harvard professors. They're launching a company called I Hope. And that will be the behavior and health therapy perspective. So developing those kind of technologies. Do you know Jeff Hammabarker? Jeff Hammabarker is one of the, he's a work at Facebook, he's one of the founders of Cloud. Oh yeah, sure, sure. So he has a famous statement in the big data world, this is the best minds in my generation are trying to figure out how to get people to click on ads. Sure, sure. He's also a doctor and he's now at Mount Sinai, he's trying to apply his knowledge elsewhere. But I wonder are things like clicking on ads, do they fit into that psyche mining with maybe turning into behavioral economics? Yeah, in fact, it is there. Actually in 2006, no, 2005, I advised one of the, talked to one of the company. You know, they are the one of the largest pay for performance, you know, the pay for click, you know. So how many time you click? So how do you identify, how do you define, like if I'm a market, I'm broadcasting in the advertisement and every click will consume my one cent. So the competitor, they will have some robots, you know, clicking on to exhaustive funds. They have a Google can fund those robots, right? Right, so, but the thing is now, how you identify those, and also, Algorithmically, how do you? Exactly. Separate those. So not only that, how do you make the impressions? What time, if you're selling a BMW or you're buying a car or buying a BMW, your decision making process is different. So based on the context, how do you really engage them, so behavioral marketing to the contextual marketing, switching back and forth? So a lot of people talk about real time and the way we define real time on theCUBE is typically before you lose the customer, maybe you would define it before you lose the patient. So, all right, we're out of time. Great conversation, really appreciate you coming on to theCUBE, it was a pleasure meeting you. My pleasure. All right, keep it right there, everybody. Paul Gillan and I will be back, we're live from the MIT Information Quality Symposium. This is theCUBE. Thank you.