DreamIt All Access is a feature article series where we interview and feature the stories of startups that have gone through the DreamIt Ventures program with a focus on DreamIt Philly. DreamIt Ventures is a venture capital and seed accelerator firm focusing on early stage venture investing. Through its accelerator programs, DreamIt has helped entrepreneurs launch over 200 companies. You can learn more about DreamIt here. These interviews have been slightly edited from the original for clarity and length.
The first company I got a chance to interview through this feature series collaboration with DreamIt Ventures was Haystack Informatics. Haystack Informatics is a visual analytics platform that helps hospitals detect, investigate and report patient privacy breaches by insiders. Haystack was born out of a collaboration between The Children's Hospital of Philadelphia (CHOP) and DreamIT Health.
I got a chance to speak with the two co-founders of Haystack: serial entrepreneur Adrian Talapan and CHOP's Chief Medical Informatics Officer Bimal Desai on how everything started for Haystack.
![2016-03-07-1457364499-8660666-Haystacklaunch.jpg]()
The Haystack Informatics team during their launch
Q: What was the inspiration behind your idea?
Bimal: It basically started with a discussion I had with the privacy officer at CHOP. A patient's right to privacy is protected by federal law, so in a hospital, this means we have to monitor whoever accesses paper and electronic records. Hospitals typically split this task across the groups charged with information security and privacy/compliance, and every hospital in the country faces a huge challenge when monitoring privacy of electronic health records (EHR) since it's extremely hard to track all the different access events.
This is why we designed Haystack. I was experimenting with some open source visualization tools and it became very clear that you could analyze access log data to basically determine inappropriate behavior. When you analyze EHR access log records, the data organize into social networks. Patients and providers in these clusters of activity tend to work together, as you would expect in a care team. Our bet was that we could get enough information from EHR access logs to discern if a trusted insider was trying to snoop on a patient's record, if people were trying to commit medical fraud, or if people were trying to commit medical identity theft.
The entire application is very intuitive, very visual. It helps privacy officers detect and report privacy breaches a lot better than other tools for this.
Q: How did you guys meet each other?
Bimal: Adrian and I were introduced through Dreamit Health Philadelphia. Adrian has a very analytical and technical mind and is a fantastic entrepreneur. I am a practicing pediatrician and have a masters degree in biomedical informatics, with a decade of domain expertise dealing with healthcare data structures. We very quickly had a prototype and wireframe diagrams setup. Adrian and I hit it off pretty quickly. We had a similar vision that this had to be dead simple for the privacy officers. The two of us delved into this whole idea of robust visual analytics and this idea of taking very, very complex data and turning it into something that was intuitive, streamlined, and useful.
Adrian: Before I met Bimal, I was helping DreamIt look at possible ideas coming out of CHOP that could be turned into separate commercial entities. CHOP had an internal competition to surface potential ideas that could be turned into actual products. The premise was that there are so many world class people and researchers at CHOP but a lot of their research is understood in that specific academic environment. But what if there was something we could do and commercialize quickly, using a "start-up" mentality, so when we looked at these ideas, that's when I got to meet Bimal.
He had already started on the idea and he had a very slick Powerpoint deck that was very convincing on the topic of privacy. Up until that point, I didn't really think about that aspect of healthcare. But if you think about it, you go to a hospital now and your health data is basically open for any employee to see. The law says there should be protections in place, but that's hard to do.
![2016-03-07-1457364590-9554830-Haystackscreenshot.png]()
The first company I got a chance to interview through this feature series collaboration with DreamIt Ventures was Haystack Informatics. Haystack Informatics is a visual analytics platform that helps hospitals detect, investigate and report patient privacy breaches by insiders. Haystack was born out of a collaboration between The Children's Hospital of Philadelphia (CHOP) and DreamIT Health.
I got a chance to speak with the two co-founders of Haystack: serial entrepreneur Adrian Talapan and CHOP's Chief Medical Informatics Officer Bimal Desai on how everything started for Haystack.
The Haystack Informatics team during their launchQ: What was the inspiration behind your idea?
Bimal: It basically started with a discussion I had with the privacy officer at CHOP. A patient's right to privacy is protected by federal law, so in a hospital, this means we have to monitor whoever accesses paper and electronic records. Hospitals typically split this task across the groups charged with information security and privacy/compliance, and every hospital in the country faces a huge challenge when monitoring privacy of electronic health records (EHR) since it's extremely hard to track all the different access events.
This is why we designed Haystack. I was experimenting with some open source visualization tools and it became very clear that you could analyze access log data to basically determine inappropriate behavior. When you analyze EHR access log records, the data organize into social networks. Patients and providers in these clusters of activity tend to work together, as you would expect in a care team. Our bet was that we could get enough information from EHR access logs to discern if a trusted insider was trying to snoop on a patient's record, if people were trying to commit medical fraud, or if people were trying to commit medical identity theft.
The entire application is very intuitive, very visual. It helps privacy officers detect and report privacy breaches a lot better than other tools for this.
Q: How did you guys meet each other?
Bimal: Adrian and I were introduced through Dreamit Health Philadelphia. Adrian has a very analytical and technical mind and is a fantastic entrepreneur. I am a practicing pediatrician and have a masters degree in biomedical informatics, with a decade of domain expertise dealing with healthcare data structures. We very quickly had a prototype and wireframe diagrams setup. Adrian and I hit it off pretty quickly. We had a similar vision that this had to be dead simple for the privacy officers. The two of us delved into this whole idea of robust visual analytics and this idea of taking very, very complex data and turning it into something that was intuitive, streamlined, and useful.
Adrian: Before I met Bimal, I was helping DreamIt look at possible ideas coming out of CHOP that could be turned into separate commercial entities. CHOP had an internal competition to surface potential ideas that could be turned into actual products. The premise was that there are so many world class people and researchers at CHOP but a lot of their research is understood in that specific academic environment. But what if there was something we could do and commercialize quickly, using a "start-up" mentality, so when we looked at these ideas, that's when I got to meet Bimal.
He had already started on the idea and he had a very slick Powerpoint deck that was very convincing on the topic of privacy. Up until that point, I didn't really think about that aspect of healthcare. But if you think about it, you go to a hospital now and your health data is basically open for any employee to see. The law says there should be protections in place, but that's hard to do.