Home » Interviews » Currently Reading:

HIStalk Interviews Carl Bertrams, SVP, HT Systems

March 25, 2011 Interviews 5 Comments

Carl Bertrams is SVP of HT Systems / PatientSecure of Tampa, FL.

3-25-2011 9-19-16 PM 

Tell me a about your background and about the company.

March 1 was my 22nd anniversary in this crazy business. I originally started out in more traditional management consulting, back in the day when information systems was mostly flowcharts. I think I learned programming on punch card decks, so that probably makes me sound really old.

After doing that for a while and really understanding process, I landed accidentally in healthcare in 1989 with a small company here in Chicago that did electronic billing for hospitals. I remember my first day. I came in and I really didn’t know a UB-82 from a hole in the ground. That’s when hospitals were moving from paper to electronic and business just took off, so it was a great way to get introduced to healthcare on the non-clinical side.

About the time that UB-92 came along, we hit the medical necessity market and really jumped on the Medicare fraud and abuse bandwagon for about eight years. We did a lot of cool technology around that when really nobody else was doing that. And then, most recently, kind of worked my way up the revenue cycle, spent some time in HIM. We sold one of our companies to 3M and got into transcription a little bit and ended up, finally, at the front of the revenue cycle river and patient access. 

HT Systems was started in 2005. The principals of the company literally have decades of experience helping hospitals improve revenue cycle efficiency one way or the other. Basically it’s about hooking specialized technology to the big vendor platforms. That’s what we’ve done on and off for 20 years. 

PatientSecure is really the coolest technology I’ve ever been associated with. It’s revolutionary way to positively identify the patients at any point in the access stream, whether it’s inpatient, outpatient, or emergency room. We do it using palm vein authentication technology. 

If you look at it at the 50,000-foot level, it’s really pretty simple. We create a one-to-one link between the patient and his or her medical record. We do that through the unique vein pattern in the palm of your hand. Every time the patient returns to the hospital or to the clinic, they simply put their hand on a scanner, and within a few seconds, their unique medical record is pulled up automatically in front of the registrar. It’s like doing a retinal scan in the palm of your hand. We don’t replace the existing ADT or registration system — we just make that process a lot faster and a lot more accurate.

When hospitals hear biometrics, they probably think of finger-type security for employee access to IT systems. Why is palm vein security better and how did you get the idea to move it out front to the patient?

Like a lot of good ideas, it started with some hospitals. Our alpha site is the Carolinas HealthCare System in Charlotte. It’s a very innovative group down there. They had been a long-time customer of ours.

When you think about fingerprints, that’s a good example you bring up. If I’m working for you and part of my job description is to punch in and do it with my thumbprint, that’s part of my job. But the experience that healthcare has had trying to have patients provide fingerprints, especially at the point of patient access, has not been that great. Carolinas had tried that and didn’t have success with it.

About that time – this was 2007 – they were looking with us at the Fujitsu PalmSecure device. It is not only significantly more accurate than a fingerprint, but doesn’t have the negative connotation that people associate with fingerprints, like law enforcement and all that. It’s contact-less and a technology for its time. Across the board, we have 99-plus percent patient adoption of the technology.

Have your clients found improvement in knowing that the person presenting an insurance card is really the person who’s entitled to the service?

This last year, the statistics I read said there were over ten million people in the United States who fell victim to identity theft. The fastest-growing form of that identity theft is medical identity theft. In 2005, medical identity was about 3% of the total, or a quarter million people. Last year, it was 7%. You’re talking about 700,000 cases of pure medical identity theft, and then maybe another half a million cases where people are complicitly lending their insurance card to their brother who lost his job or there’s some sort of minor conspiracy going on there between the patients.

This literally just shuts the door on that, but it also addresses the human error element. We’re putting the system in in Harris County in Houston, A Houston Chronicle story said there are 466,000 patients in their MPI that shared the same name with as least 24 other people in the system. You can imagine that whether you’re there with a stolen ID, or you just come in and say, “My my name is Jim Johnson” and there’s 37 other Jim Johnsons in the system, the chance for error at the front end is bigger than I think most people would think it is.

I know at my hospital we have that problem all the time, where either the patient gives the incorrect name or someone looks it up wrong, doesn’t find it, and enters the other name, and then they have to go back and merge the medical records. That’s a pain because not all systems, including the clinical ones, handle patient merges all that well. That’s pretty much eliminated, correct?

It is if you do it right on the front end. When the patient comes in for the first time, they’re in the hospital system, but not in the biometric system. You put your hand on the sensor. It’s going to say we don’t know you biometrically. At that point, the registrar does what they do every day — ask you for ID. Most of our customers will only enroll a patient if they present a valid photo ID. I find you in the system and do a one-time enrollment where I’m linking you to that medical record. 

From that point forward, when you walk in, you put your hand on the sensor. You’re basically finding yourself in the system. It’s virtually impossible for you to ever have a duplicate medical record downstream from that enrollment, and more importantly, to have a medical record overlay. That’s a much more serious situation where you’ve picked the wrong record of the same-named person and now you’ve laid their medical results and lab tests and blood type and all those things on top of it. In a good situation, that can just be a hassle for IT — like you said, merging the medical records — but in the worst-case scenario, you give somebody the wrong medicine or you kill somebody and the hospital is looking at a lawsuit that is hard to get away from.

The other benefit would be that most every provider organization has multiple venues of care, whether it be clinics or physician practices that are owned or affiliated, plus their own inpatient facilities. If you were connected to the same system, or maybe even if not, you could enroll the patient once and be sure that no matter where they show up, you know who they are.

That’s one of the cool things about the way we set up the technology. You can have an unlimited number of unique identifiers associated with the same single biometric.

Duke is a good example. They have GE Centricity at the clinics, they’ve got Siemens Invision at one hospital, they’ve got Meditech at another, they’ve got a homegrown at the big university, they have their own EMPI. I could be five or six different numbers within the system. By putting my hand on the sensor, it knows who I am, and it’s smart enough to know that, “Oh, I’m in Durham, pull me up in Invision. Oh, I’m in Raleigh, pull me up in Meditech.” 

Just as you were saying, a lot of mistakes happen when somebody gets registered at the physician’s office or at the clinic, but is registered differently at the hospital. That’s one of the places where the mistakes happen. By having this cross the whole enterprise platform, you tend to eliminate that mistake.

You mentioned that your hardware is from Fujitsu. How are you adding value to that? What is your secret sauce that brings you into the picture as part of the value chain?

The Fujitsu device is a near-infrared camera — great technology. With biometrics, you need to very aggressively manage the biometric database. It isn’t one plus one equals two. There are a lot of moving parts.

We wrote the algorithms, the search algorithms. We make it incredibly fast and easy for you to be found in the database, even if you come into the emergency room unconscious. If you were previously enrolled, they’d be able to bring this to the bedside and know who you are, as opposed to treat you as John or Jane Doe. 

Our secret sauce is really those proprietary algorithms and the edit engine that we wrote. I think that makes us a really comfortable partner for our hospitals. We have decades of experience being under the hood of all these different HIS and PMS platforms. We know the workflow. We know how these things operate. We’re very comfortable in all these different platforms. We’re not just technology guys come in and selling something slick to the hospital. We know hospital revenue cycle and bring a technology that absolutely shows them an ROI, but makes it easy to adopt both by their staff and by the patients.

If I’m a hospital and I’m interested in your solution, what’s involved with implementing it and how do you price it?

The pricing model is enterprise-driven, so there’s a one-time software license fee. The enterprise could be that I’m a 200-bed community hospital and that’s the start and end of it. It could be that I’m 17 hospitals across three states with 57 clinics and 20 owned physician practices.

There’s an implementation fee and that goes up or down based on how many different interfaces we need to write and how many different points of entry that we’re actually going to roll this out to. Our implementation fee is all-inclusive of the interfaces, the on-site implementation, and the user training. We sit there with the hospital staff while they actually enroll patients and answer those questions that come up.

It’s a pretty light install. It all happens behind the hospital’s firewall. We operate on SQL Server. It can be a virtual server. It’s a very small footprint. Carolinas, with almost two million patients in the database — they’re probably a couple of gigs of storage. It’s amazingly small of a footprint that drives this whole engine.

For that 200-bed hospital that you mentioned, how long would it take to implement and roughly what would the cost be?

The implementation time is a pretty standard 60 working days, two to three months from the time we say let’s go, have a kickoff meeting, and figure out where in the workflow they want to insert this. We do a lot of the interface work off site — dial into their test system — and then we put the technology on site and do the training. 

From start to finish, a hospital is normally going to be live in a couple or three months at the most. If they want to be more aggressive, it can be shortened sometimes.

In terms of a ballpark figure, if I’m a 200 bed hospital and have 15, 20, or 30 points of entry that I want to cover, you’re probably talking about $100,000 to $150,000 as a one-time cost with an annual maintenance fee beyond that. We also have a model where if a hospital doesn’t want to lay out upfront capital, they can spread the whole thing out over three years and there’s no money up front and we don’t tag on any interest.

We try not to nickel and dime. The one thing I’ve learned in twenty-some years of hospitals is give them a price and let them budget it and be done with it. If hardware breaks, we replace it. We extend the warranty on the hardware for as long as somebody’s a customer. If your interface needs to be tweaked, if you want a custom report, all that’s included. The only time that you’d be looking at additional fees was if you took out Meditech and put in McKesson, where you have to totally rewrite the feeds. Other than that, it’s pretty straightforward.

Your website mentions that Japanese banks are already using the palm vein scanning and also that standardized test companies are moving in that direction. Do you see other potential uses in healthcare, for instance, anything related to patient safety?

We’re meeting with some folks around the country who want to look at this for e-prescribing. You could certainly put this in the nursery and control who’s coming in and out. We’ve had hospitals that want to use it also as a vendor identification system. For us, we’ve started in patient access, but we certainly see a lot of other use cases. Once you’ve got the technology, extending it to another place in the system is a minor cost.

Any concluding thoughts?

In healthcare today, there’s a lot of cool technology, as we saw at HIMSS. But for those of us in the revenue cycle — the non-clinical side of healthcare — the bottom line is the bottom line. CFOs are tired of hearing about this fluffy, feel-good kind of ROI. You’d better be able to show them that you actually are reducing costs, or you’re solving a problem and improving quality and patient safety, really prove it. We feel this technology does that every day.

Patient access is the filter at the front of the revenue cycle. If you get it right there, everybody else’s job downstream is a lot easier. If you screw it up at the front, you know what they say about stuff running downhill. We help the hospital get the very first job done right, and that’s identifying the patient. If we can do that, the ROI is undeniable. 

HT Systems is in a great space in the market. We love what we’re doing. We also really love the fact that there’s vehicles like HIStalk out there to help us get this message out and to give us feedback from the field, from the vendor community, and from the hospital community. 

It’s exciting time for us. We think we’re just at the beginning of a big set of waves that are going to come down. Other than that, we just looking forward to keep telling people about what we’re doing.

There is one last thing I would like to say. I’d really like to let our Fujitsu partners and friends over in Japan know that we’re thinking about them and praying about the situation over there. They’ve got a tough road to go, but it’s a great culture and a great spirit, and I’m sure that they’re going to ultimately recover from this as strong as ever. Our thoughts and prayers are definitely with them.

HIStalk Featured Sponsors


Currently there are "5 comments" on this Article:

  1. Has this technology been tested anywhere else? False positive rates? False negatives? If it is so good, I would have expected law enforcement to be using it but have not seen that.

  2. Good question. Yes. The banking industry in Japan uses palm vein scanning in conjunction with the smart ATM card for identification on ATM machines where the scanner is built right in. Also, the organization that administers the Graduate Management Admission Test (GMAT) uses palm vein scanning as part of its registration process. You are enrolled when you go into the test site and then if you need a bio-break during the 4 hour exam, you have to authenticate yourself back in to prevent “ringers’ from coming in and taking the test for you.

    Since vein pattern recognition requires blood flow and therefore is not suitable for forensic analysis, the scan data has no practical use by law enforcement for crime investigations. You can’t leave your vein pattern at the scene of a crime.

    While many people may not be familiar with this technology, it is way past the “proof of concept” stage. PatientSecure has been live in the U.S. healthcare market since 2007 and over the course of the last 4 years our hospital customers have successfully enrolled/authenticated nearly 5,000,000 patients with thousands more added each day.

  3. If you only have to put in your birth date then scan, how would your system stop a person from enrolling multiple times just under different birth dates?

  4. The same person cannot enroll multiple times with different birth dates because the system will simply not allow the same biometric template to be duplicated and stored in the database.

  5. This technology has never been tested by NIST, who test biometric algorithms like this for free, with the caveat that the results are public. The only publicly available independent testing of the technology is the International Biometric Group’s CBT several years ago, which showed Fujitsu Palm Vein technology had a serious problem with “drift” (undesireable changes in the biometric landmarks measured due to time passing), such that performance was not as good after the 30 days as it was the same day. This is the reason that you have to enter your birth date – it is not accurate enough to reliably search millions of enrollments and find a person without the “hint” of their birthday to limit the false matches on a 1 to all search. Only as an exception can a patient be searched among all the patients, and the claim that “the system simply will not allow the same biometric template to be duplicated and stored” is a calculated and deceptive answer, since literally duplicating a template would be detected, since it’s identical. A second template from the same person, however, has the possibility of the feature drift causing it to not be detected as a duplicate, and being allowed into the database. Be careful of crafty answers that dodge the real issue.

    Interestingly, the usage for the GMAT at Pearson is quite unambitious, since it only confirms that a person is the same person who entered the testing room the same day, and only 90 days of history is searched for duplicates, on the premise that the professional test takers would have taken the test in the last 90 days too. That 90 day search took 30 days offline to complete. Compare that the the use by Prometric of fingerprint technology from LexisNexis and BIO-key for the MCAT, CPA, and other exams, and they perform 5 years of real-time lookback against all test takers whenever someone signs in. That is real commercial ready search power, and explains why the FBI also uses BIO-key technology.

    Another thing to think about – Fujitsu is the sole maker of the underlying technology. There is no Palm Vein competitive market – only Fujitsu. So, you pay four hundred dollars per scanner (monopoly prices), and Fujitsu might decide one day to no longer manufacture palm vein readers, just like they decided to stop making fingerprint scanners and software back in 2006, leaving their customers high and dry.

Founding Sponsors


Platinum Sponsors


















































Gold Sponsors













Reader Comments

  • ex-epic: It's been awhile since I left Epic in 2012, but reading these stories has brought a rush of memory around the company cu...
  • Former Cult Member: The culture of fear is real at Epic. There have been multiple instances during staff meetings (sometimes lovingly called...
  • TenuredEpic: So you "would struggle to return to campus work as an Epic employee if I were high risk and otherwise fastidiously isola...
  • ex-HHC: Great take. It's amazing how many errors there are in address information in EHRs , which makes geocoding quite a challe...
  • Gonepostal: re: "Politico also reports that two senators have introduced legislation that would make the post office’s address mat...

Sponsor Quick Links