Biometric Patient ID Technology with M2SYS President, Michael Trader Podcast Interview

Thought you might find this interesting


Biometric Patient ID Technology with M2SYS President, Michael Trader podcast interview and it’s ability to prevent duplicate MRNs, identify theft, and improve a hospital’s ROI

Biometric Patient ID Technology with M2SYS President, Michael Trader

Michael Trader, Co-Founder & President of M2SYS

In this podcast interview, HIT Consultant spoke with Michael Trader, Co-Founder & President of M2SYS for an insightful conversation about biometric patient identification technology. M2SYS’s Right Patient solution is the only multi-biometric patient identification system that uses fingerprint, palm vein, finger vein, iris and face recognition to ensure 100% patient accuracy, strengthen patient safety, and reduce hospital liability. The solution also has the ability to seamlessly interfaces with any EHR software so you can avoid keystroke error and maximize usability. As the healthcare industry moves closer towards an integrated delivery model, healthcare providers will begin to increasingly rely on tools such as these to achieve interoperability without compromising the potential risk of security breaches due to patient sensitive data.
During our conversation, Michael will describe the advantages of biometric patient ID technology including how it can help prevent duplicate medical records, identity theft, and ultimately improve data interoperability across health information exchanges and integrated delivery networks.

[See also: Biometric Patient ID Technology: Is It The Future of Patient Access?]

Key highlights of the podcast interview include:

  • Advantages/benefits of biometric technology for patient identification vs. traditional forms of ID
  • How biometric patient ID technology can prevent duplicate MRNs and medical theft
  • The ptoential ROI of biometric patient ID technology
  • Helping hospitals improve data integrity and data interoperability across health information exchanges and integrated delivery networks
  • Key considerations and concerns when testing biometric ID technology in healthcare organizations

This podcast is also available on iTunes

Thought you might find this interesting


Get in-depth healthcare technology analysis and commentary delivered straight to your email weekly

  • RT @ericpoje: Well done! #Biometric Patient ID Tech. audio interview w @M2SYS President, @michael_trader @hitconsultant

  • “The Future Patient Identification Technology Podcast with @m2sys President Michael Trader #HITsm #HCIT…

  • Your interview with Michael Trader from M2SYS featured several important misconceptions and exaggerations about biometrics that cannot go uncorrected.

    Before going to details, we need to understand two important characteristics of biometrics that are seldom if ever set out by vendors. Firstly, there is an inevitable trade-off between False Positives and False Negatives: the more *specific* the biometric system (ie less likely to confuse you with someone else = low False Positive Rate) the less *sensitive* it will be (ie more likely to reject you = HIGH False Negative Rate). This “Detection Error Tradeoff (DET) creates an inescapable tension between security and convenience. Both are obviously important to patient identification. M2SYS makes big claims about the importance of security and the elimination of duplicate records, but if the security of the system is tuned too high, it will become difficult to use because it will become “fussier” about detecting legitimate patients. Healthcare workers will not tolerate excessive False Negatives when it delays their access to records.

    The other important characteristic is a mathematical quirk called the “Birthday Paradox” which refers to the counter-intuitively high probability of matching pairs of people in large databases. In a group of just 25 people, the probability of any two of them sharing a birthday is surprisingly high (over fifty per cent) despite the fact that case by case the chance of you and I having the same birthday is much much less, at around 0.3%. The ‘paradox’ is explained by the fact that there are hundreds of paired combinations amongst the set of 25. Similarly, when a number of people are enrolled in a biometric database, even if the modality is very accurate, with say an error rate of one in a million (0.0001%) the probability of finding a False Match somewhere in the database grows exponentially: for 100,000 people, the chance of a match is 10%; for a million, it’s 63%.

    The Birthday Paradox effect becomes crucial in the “one-to-many identification” (1:N) mode advocated by M2SYS to prevent duplicate medical records. To achieve the results they claim, hospitals and HIEs will need to ensure that all patients are biometrically enrolled and that the templates are accessible at all points of care in the network so that each new patient presentation may be screened against the whole set. As biometric databases grow to hold millions and more, it is imperative that the vendors reveal what the true error rates and trade-offs are. In real life the performance of palm vein scanning is something like False Positive Rate = 0.01% and False Negative = Rate 1% [Reference International Biometric Group Fujitsu Palm Vein Testing, 2006]. For a database of 10,000 patients, the chance of at least one False Match in the set will be 63%. So I predict that as enrolments expand to a level sufficient to prevent duplicates, users will experience dozens of false matches every day. There will need to be a fall back protocol for resolving identity mismatches when an identification system shows there are several candidates in the database for a new patient when scanned. Despite this need, M2SYS actually claims that biometric identification can work when a patient presents with absolutely no identity documents. That has to be a significant overstatement.

    Now to pick up some specific misconceptions and exaggerations in the recorded interview.

    0’33”: The introduction mentions “100 per cent patient accuracy”. There is no such thing as perfect security. No discussion of biometrics should carry forward this misconception. In real life the accuracy can be surprisingly lower than 100%.
    4’28”: ‘Biometrics cannot be shared, forged or stolen”. This is simply untrue. Fingerprints for example are readily stolen and reproduced by silicone moulds to spoof even very high grade detectors (see the Mythbusters TV program on this). And if an attacker can gain access to the biometric templates, then they can forge matching synthetic traits by reverse engineering. This has been demonstrated with fingerprints, face and most recently iris modalities.

    5’57”: “Your biometric characteristics are absolutely unique”. This sort of unqualified impression of perfection is belied by the fact that all biometrics commit False Positive errors. No trait is perfectly measured, processed and matched, so ‘absolute uniqueness’ is a misleading choice of words. If a biometric 1:N identification system throws up a dozen candidate matches from a large database of patients, the user will be entitled to ask what “uniqueness” is supposed to mean.

    6’07”: Question: “It can definitely prevent duplicate medical records, correct?” Answer: “oh yeah, absolutely”. Again this is an unwarranted implication that biometrics work without error. With finite error rates, it is logically impossible for biometrics to “absolutely” prevent duplicates. Furthermore, with 1:N matching as advocated by M2SYS, the elimination of duplicates is based on an assumption that all enrolled templates are accessible at all sites. At 11’13” Mr Trader says it is a “fact that having biometrics … can prevent the creation of duplicate medical records”. Prevention is a big claim that can only be supported when every patient is enrolled and all templates are available for comparison at all points of care. With multi-million template databases, as discussed above, we really need the vendors to specify the real life Detection Error Trade-off to manage Birthday Paradox false matches.

    If the biometric system is not networked across all sites, then well organised medical fraudsters (the types of criminal that are supposed to be targeted by biometric security) will plan their doctor-shopping and will target sites that are not online with the identification system.

    20’06” “Only takes one duplicate in an HIE to poison every other hospital that is a member of that exchange”. Is this not a major exaggeration?

    24’15”: In the discussion of independent testing and certification, Mr Trader mentions NIST standards, and without being specific he says that independent testing has gone on for a long time, but he doesn’t actually answer the question about certification. The truth is that as yet there are still no accepted biometric performance testing standards and methods.

    26’55”: Mr Trader strongly urges that the biometric identification system retain raw images of the scanned trait, for audit purposes. He stresses at 28’35” that it is “extremely important” for hospitals to be able to forensically defend themselves when patients dispute their bills. I question this. Is there any research to substantiate the seriousness of this problem? For a patient that has been hospitalised for many days and has run up a bill that is big enough to be worth disputing, I would expect there is plenty of other routine evidence to be found amongst dozens of staff and procedures to establish the identity of the patient if such a case goes to legal resolution.

    Furthermore, the idea of saving raw images flies in the face of the common claim by biometric vendors that only templates are stored. In fact M2SYS itself has published a statement that raw images should not be stored! See

    “The M2SYS hybrid biometric platform does not store any biometric images and it is impossible to recreate the original biometric image if a hacker were to steal the biometric enrollment template. As we have stated before, biometric identity enrollment templates stored on a server or computers are not actually images at all. They are a mathematical representation of the data points that a biometric algorithm extracts from a scanned fingerprint, finger vein, palm vein or iris. The identity template is simply a binary data file, a series of zeros and ones. The algorithm then uses the template to positively identify an individual during subsequent fingerprint scans. No image is ever stored or transmitted across a network.”

    Thanks for the opportunity to comment at length like this. We would all agree that security and privacy of networked hospital records is a major challenge. I am keen that decision makers in the health sector have a proper understanding of the biometric option and that they do not run afoul of idealistic claims. There is no such thing as perfect security, and the subtle side effects of the failure modes of biometrics deserve careful attention.

    Stephen Wilson
    Lockstep Consulting, Sydney.