Skip to main content
    All Episodes
    Episode 091 · November 5, 2026 · 32m listen

    Why Medical Device Penetration Testing Is Different | Ep 91

    Key Takeaways

    Prioritize clinical impact — Christian explains how medical-device penetration testing evaluates...

    Prioritize clinical impact — Christian explains how medical-device penetration testing evaluates potential patient harm as well as technical security weaknesses.

    Scope testing from the device model

    Use the threat model and relevant hazard analysis to identify realistic attack paths rather than relying only on an IP-address inventory.

    Include the complete attack surface

    Consider software, connected applications, cloud services and applicable hardware interfaces such as USB, Bluetooth, JTAG and UART.

    Go beyond vulnerability scanning

    The episode discusses abuse and misuse cases, robustness testing, fuzzing and vulnerability chaining as ways to examine exploitable behavior.

    Make findings reproducible and traceable

    Christian discusses tester independence, test scope, methods, retesting and evidence linking findings to mitigations. His regulatory commentary is interview perspective, not a verified universal 180-day legal deadline; FDA guidance is nonbinding unless it cites a statutory or regulatory requirement.

    Episode Summary

    Host Christian Espinosa explains how medical-device penetration testing differs from traditional enterprise testing, emphasizing clinical impact, realistic attack paths and the complete device attack surface. He discusses threat models, relevant hazard analysis, software and hardware interfaces, abuse and misuse cases, robustness testing, fuzzing, vulnerability chaining and reproducible findings. The episode also covers tester independence, traceability and retesting. Regulatory statements are the host’s interview perspective, not a determination that a particular device meets FDA requirements or that a universal 180-day legal testing deadline applies.

    Chapters

    1. 0:06What is medical device penetration testing?
    2. 4:11like it's vulnerable
    3. 8:15process until we get the risk to an acceptable level in terms
    4. 12:19isolated from the division that is developing the product
    5. 16:23of data and random data to an input
    6. 20:28gives us the most accuracy and completeness
    7. 24:33the cloud. So we have to test all three of them. The
    8. 28:37I've talked about these indirectly, is that the test scope doesn't match

    Frequently Asked Questions

    Quick answers drawn from this episode.

    • Host Christian Espinosa explains how medical-device penetration testing differs from traditional enterprise testing, emphasizing clinical impact, realistic attack paths and the complete device attack surface.

    • Prioritize clinical impact — Christian explains how medical-device penetration testing evaluates potential patient harm as well as technical security weaknesses. Scope testing from the device model — Use the threat model and relevant hazard analysis to identify realistic attack paths rather than relying only on an IP-address inventory. Include the complete...

    • This episode covers Penetration Testing, FDA Premarket Cybersecurity, and Threat Modeling. It's part of The Med Device Cyber Podcast, hosted by Blue Goat Cyber, focused on practical medical device cybersecurity guidance for MedTech teams.

    • The episode also covers tester independence, traceability and retesting. It's most useful for medical device manufacturers, cybersecurity engineers, regulatory affairs professionals, and MedTech founders preparing for FDA review.

    • Prioritize clinical impact — Christian explains how medical-device penetration testing evaluates potential patient harm as well as technical security weaknesses.

    Listeners also asked

    Quick answers pulled from related episodes.

    • What does Episode 26 cover about "How to Measure Exploitability in Medical Device Cybersecurity"?

      How do you calculate cybersecurity risk when the thing you are measuring has an attacker making decisions on the other side? Probability works when there is historical data to lean on. Elevator cables wear out. Steel fatigues. Tornadoes follow patterns. A motivated attacker...

      From Episode 026 · How to Measure Exploitability in Medical Device Cybersecurity | Ep 93
    • What does Episode 25 cover about "How to Make Your MedTech Startup Investable with Gavin Leonard"?

      What actually makes a MedTech startup investable? Is it the technology, the clinical need, the founder, the pitch, or whether the company has simply done enough homework before walking into the room? Christian Espinosa speaks with Gavin Leonard of MedTech Syndicates about what...

      From Episode 025 · How to Make Your MedTech Startup Investable with Gavin Leonard | Ep 92
    • What does Episode 23 cover about "Why Medical Device Companies Need Tabletop Exercises with Daniel Hammond"?

      What happens when a medical device company experiences a serious cyber incident? Daniel Hammond joins Christian Espinosa to discuss why incident-response decisions should be made before an organization is under attack. They explore the role of tabletop exercises, why ransomware...

      From Episode 023 · Why Medical Device Companies Need Tabletop Exercises with Daniel Hammond | Ep 90

    Share this episode

    Pre-fills with: "Prioritize clinical impact — Christian explains how medical-device penetration testing evaluates potential patient harm as well as technical security weaknesses."

    Show original episode notes
    What makes medical device penetration testing different from a traditional cybersecurity penetration test? In this solo episode, Christian Espinosa explains why medical device penetration testing has to be approached through the lens of patient safety rather than simply data protection. Christian breaks down the difference between vulnerability scanning and penetration testing, why the threat model and hazard analysis should drive the scope, and why hardware interfaces such as USB, Bluetooth, JTAG and UART need to be included alongside software, cloud and mobile components. He also walks through the testing methods the FDA expects to see, including abuse and misuse cases, robustness testing, fuzz testing, attack surface analysis and vulnerability chaining, as well as the importance of traceability, tester independence, retesting and reproducibility. IN THIS EPISODE: 00:06 What medical device penetration testing actually is 03:10 Vulnerability scanning versus penetration testing 05:14 Traditional IT testing versus medical device testing 07:19 Why hardware attack surfaces matter 11:44 Why penetration testing needs independence 14:56 Abuse, misuse and robustness testing 18:09 Vulnerability chaining and attack surface analysis 22:20 What the FDA expects in the final report 27:38 Why retesting and traceability matter 29:51 Common penetration testing mistakes that cause FDA problems The Med Device Cyber Podcast is brought to you by Blue Goat Cyber, cybersecurity experts providing essential security solutions for the medical device industry. Learn more: https://bluegoatcyber.com Schedule a Discovery Session: https://go.bluegoatcyber.com/meetings/blue-goat-cyber/discovery-session Christian Espinosa on LinkedIn: https://www.linkedin.com/in/christianespinosa/ Blue Goat Cyber on LinkedIn: https://www.linkedin.com/company/blue-goat-cyber/ Blue Goat Cyber on Instagram: https://www.instagram.com/bluegoatcyber/ Blue Goat Cyber on Facebook: https://www.facebook.com/bluegoatcyber/ Blue Goat Cyber on YouTube: https://www.youtube.com/@BlueGoatCyber/?sub_confirmation=1 Download MP3 — full interview
    What is medical device penetration testing? What does the FDA expect? And what makes medical device penetration testing different than what I like to call traditional penetration testing? That's the topic of today's podcast. I'm Christian Espinosa, founder of Blue Goat Cyber and the Med Device Cyber Podcast host. And this is a topic that comes up quite a bit. So by the end of this episode, you will understand what penetration testing is in context of what the FDA wants and with medical devices. Penetration testing, number one, is not a compliance exercise. It is about patient safety. With penetration testing, we try to use the same tactics, we do use the same tactics as an adversary, as a cyber criminal, as a malicious actor, as a nation-state actor, to try to break into a medical device. And then once we break into the medical device, from there we try to see what's the next thing we can break into and what damage we can cause to the patient. Can we cause a misdiagnosis? Can we take the device offline at scale? And what is the impact ultimately to the patient? With penetration testing, In cybersecurity for medical devices, it's about patient safety, not data protection. A lot of people still think it's about data protection. Not true with medical device penetration testing. So a couple things I want to talk about. First is a comparison and contrast between traditional IT enterprise penetration testing and medical device penetration testing, 'cause there's still a lot of confusion on this topic. And I understand there's a lot of confusion on penetration testing in general, but maybe I'll take one step back. So first off, penetration testing, as I mentioned, is trying to break into a device using the same tactics and techniques and procedures as the bad guys, except you hire the good guys like Blue Coat Cyber to do this, and they do it ethically and morally, and don't try to kill a patient. You know, obviously we do it in a controlled environment. And then we don't use it to our benefit. We work with a manufacturer to tell them what we found, how we found it, and give them a recommended course of action to fix what we identified, to plug the holes. And then we come back and do another round of testing to validate that the hole was plugged. This is like someone trying to test the security of your house. They walk around your house, they look at your windows, they look at your doors, They look at your cameras and they try to get in through the doors and the windows. And then they tell you which ones have vulnerabilities and what you should do to fix it. So the next time when a real burglar comes by, your house is more secure. That is penetration testing. And this is drastically different than vulnerability scanning, which there's a lot of, you know, controversy about these two. Some people think they're the same. Vulnerability scanning simply identifies a vulnerability. That's it. It doesn't tell you what the risk is associated with that vulnerability. It's missing the impact. The penetration test takes it one step further. I identify a vulnerability and now I exploit the vulnerability because you can have a vulnerability that's super difficult to exploit. The penetration test tells you how easy it is to exploit that and gives you a, a true measure of the risk. And then also, When I'm able to exploit that vulnerability and get through the door, what's the next thing I can notice? A vulnerability scan doesn't allow you to see the next thing. It just identifies the first vulnerability. But often you get through one door, there's another door, you get to that door, and now you can have a greater impact. So this is what a penetration test does. It's more holistic and can measure impact and risk much better than a vulnerability scan, which is some, like somebody looking at your house and saying, that door looks like it's vulnerable. Someone may be able to pick the lock. The window looks like it's vulnerable. But on my house, where, or my condo where I live now, I'm on the second floor. Somebody would need like a 40-foot ladder to get to my window. So even though I may have the same vulnerability in my window versus the window on the first floor, from a contextual perspective, The risk is less and the exploitability of that window in my unit is less than the one downstairs. And that's what the penetration test will tell you. Because somebody actually tries to take advantage of that vulnerability. Traditional versus medical device penetration testing. Traditional, it's about the data. Can I protect the sensitive data? Can I protect the credit card information, the government secrets? The intellectual property, my protected health information, that is traditional penetration testing and what we look at. With medical device penetration testing, what harm can I cause a patient if I can attack this device? It's not about the data, it's about patient harm. Traditional, we scope the penetration test from an asset list or an IP range, a list of IP addresses or systems. For a medical device, we scope the test from the threat model and the architecture. So what are the threats that are applicable to this device that can harm a patient, and how do we validate those threats have been mitigated? That's what the penetration test is for. With normal penetration testing, we look at technical severity. I'm able to compromise the integrity of this asset. So what, what does that actually mean? In medical devices, we look at the context, not just the technical severity, but the context. So if I'm able to compromise the integrity of an in vitro diagnostic or IVD system and cause a misdiagnosis, what is the impact to that patient? That's what we look at. And to do this properly, We need the hazard analysis. So we don't go off of an asset list. We go off of the threat model, which I mentioned before, which gives us the threats that are applicable to the medical device. And to determine the impact, we have to look at the hazard analysis, which tells us if I'm able to alter the integrity, here's the impact to the patient. If I'm able to alter the availability, here's the impact to the patient. So a medical device penetration test, the risk is calculated in terms of patient safety. The attack surface is also different. For enterprise, we typically look at a web, network, cloud, endpoints. For a medical device, we have to look at the hardware as well, the device itself. A lot of medical devices are a physical device, like a wearable. Uh, it could be a glucose monitor, continuous glucose monitor, CGM. That often communicates with a mobile app, which often communicates with something in the cloud. So all three of those are the attack surfaces for a medical device. It could also be a patient monitoring system in a hospital. That patient monitoring system may have a USB port, may have Bluetooth, may have a specific RF frequency, may have DICOM, serial, all these different hardware interfaces into that device. So we have to test all of those plus any debug ports that are exposed, such as JTAG or UART, to be complete and accurate for a medical device. These are often never part of a traditional penetration test. Success for traditional looks like a, a plan that the client can go fix the items. With a medical device, we have to have traceability. We have the evidence of the penetration test, The client has to fix what we identify, then we have to do another test, often another round, to make sure their fixed action worked and to make sure they didn't introduce any new vulnerabilities. And we have to continue that process until we get the risk to an acceptable level in terms of patient safety. That often takes multiple iterations. Each iteration has to be tracked and logged, as well as each fixed action. That provides the traceability. So at the end of the day, we have to show, here's what we identified initially on the threat model, here's what the penetration test validated or didn't validate, here's the next step and the next step, and ultimately what's the residual risk to the patient. Along with that, with the traditional penetration testing, it's often complete when the report is delivered. Medical device penetration testing is not complete until the risk is at an acceptable level, which is considered controlled risk for patient safety. A lot of medical device manufacturers make the mistake of hiring a generalist penetration testing firm to do their medical device penetration testing. This doesn't work for all the reasons I just mentioned, uh, and the skill set's very different. The risk matrix is very different, like I talked about. So you need— it's important when you're making these decisions to hire a firm that specializes in medical device cybersecurity and medical device penetration testing. It is a very different skill set. They're sort of loosely related, but they're also drastically different. They're both under the umbrella of penetration testing, but it's like an Audi and a Ford Pinto. They're both under the umbrella of a car, but they're very different. An Audi is probably not a good example, maybe a Porsche and a Ford Pinto. So a few things we have to look at. We talked about the hazard analysis. A hazard analysis is simply the effects that something, some sort of stimulus could cause to a patient. Hazard analysis has things like Electrical shock. If the medical device has a problem and it shocks the patient, what is the impact? If the device goes offline during the middle of a procedure, what's the impact? So on that hazard analysis, we have to have things like integrity and availability because these are the effects we can cause from a cybersecurity point of view and a penetration testing point of view. So if we can alter the integrity, what's the impact to the patient? If we can alter the availability, what's the impact to the patient? And that's how the risk is derived. I talked about traceability quite a bit. Whenever we perform a penetration test for a medical device client, we ask for a few things. One is the hazard analysis so we can calculate risk properly. Risk, as I mentioned, is in terms of patient safety, which is based on ISO 14971. That's a standard that talks about severity of harm to a patient if something goes wrong. We ask for the threat model. We have to have traceability, which means the threats that are probable or threats that are realistic for the medical device need to be put on a threat model and a threat table. And our penetration test has to validate the controls that they work properly that the medical device manufacturer has put in place. So if you have a control that secures Bluetooth Low Energy, as an example, then we have to go test and make sure that control is working properly. If not, then the control has to be updated. And in some cases, you may have to disable Bluetooth Low Energy if you simply cannot fix the way we're able to get into the device. The penetration testing must also be done by an independent third party. The FDA states that there has to be independence. There are some companies that are large enough to get away without using a third party. If you're a very large company, uh, like some of the strategics in our industry, you may have a division that is completely isolated from the division that is developing the product. In that case, you can argue independence because they don't have insight. Or visibility into each other, those two divisions. But typically 99% of organizations need to find a third party such as Blue Coat Cyber to perform the penetration testing. The whole idea is we don't want the same people that have developed the product to test the product. Some of the items the FDA is looking for are number one, to prove the design works. Cybersecurity should be designed into the product, not bolted on later. It should be designed into the product. So part of what we test for are the security requirements. The security requirements should be designed into the product, and we test that those requirements that were designed and developed are working properly, such as authentication, such as encryption. We want to test that it was developed properly. We also want to test threat mitigation. So what threats can happen and what are we doing to mitigate them? A threat could be a door to my condo, like over there to my left, there's a door. There's a lock on the door, deadbolt lock. Does that deadbolt lock actually work? I think it does, but the penetration test would validate it. Someone could come in, see if they can stick a credit card in there. A special tool and see how easy it is to get through the door. Of course, as I mentioned, they'd have to take a large ladder to climb up to the second floor and a lot of people would probably notice them. But regardless, if I care about my security, I want to have that door tested because a threat is somebody coming through that door. That is an entry point into my environment. Just like on a medical device, we test all the entry points into the software. This is any type of interface, any type of interface to the software, I should say. Touchscreen, USB, network connection, Bluetooth, NFC, RFID, those are all ways into the software. Those are all things that have to be tested. The FDA asks for a lot of specific tests. I'm going to run through these. They're labeled vulnerability testing, and this is per ANSI 62443-4-1. The first one is abuse and misuse cases or malformed inputs. Software developers create the device with use cases, which is how it's supposed to be used. What we need to test from a penetration point of view with medical devices is what happens if we try to use it in a manner it wasn't designed to be used. This is the abuse and misuse cases. If there's an interface that expects DICOM images, as example, what if we try to send it something different? What does the interface do? Does it accept the data? Does it crash the device? Does it give us a command line where we can elevate our privileges? That is testing it for things it's not designed to do. That's also a malformed input. The other requirement is robustness. So if I'm able to attack a device, a medical device, which they're gonna be attacked persistently when they're deployed in a hospital or healthcare delivery organization, these are considered hostile environments. If I'm able to attack it persistently, does it fail safe? Does it have resilience? Is it robust in its design and in its ability to withstand the hostile fire it's going to be under? And when I say fail safe, if it does crash or something goes wrong with it, does it do it in a safe manner versus an open manner, which means people can get into it. So that is robustness. The other item, and a lot of people ask about this, the other requirement is fuzz testing. Fuzz testing is basically sending a lot of data and random data to an input. So this is against a protocol, an interface, an input on a web application, on an API. So the whole idea is we have a device called a fuzzer that generates random characters of small links, long links, mid, mid-sized links. It just generates all these random things and various patterns to see how the device responds to it. This has been a way a lot of attacks have been successful in the past. Sometimes even accidentally hit a certain command, a keyboard sequence on Windows and that got them into a command line, or they sent a certain size packet to a device and that caused the device to crash. These are all ways of fuzzing. We're just doing this in a very regimented and prescribed manner to see how the medical device responds. So any input it receives, we want to fuzz the data to those inputs to make sure it's doing some sort of input validation and bounds checking. So it's not accepting large amounts of data, which is bounds checking, and it's validating the data to say, if I don't need to accept special characters, I'm gonna reject that input. The other requirement is attack surface analysis. I kind of alluded to this, but this is all the ways somebody can get into the software. And this is ways people may not even document, such as the USB port that supposedly was only used for maintenance. Or if I can easily pop the COVID off of a device, there may be a JTAG port exposed. That's a way into the software. So we have to look at all of those surfaces because those are attack surfaces that somebody could use. And there's various ways to mitigate those, such as sealing the enclosure so it can't be opened for the JTAG and UART ports, as example. The other requirement is vulnerability chaining. This means exactly what it sounds like. If I'm able to break into one vulnerability or exploit one vulnerability, let's say to get in the front door of a medical device, What's the next one that I can see beyond that? And can I take advantage of that one? And then what's the next one? So I'm chaining these vulnerabilities together in order to achieve my objective or my exploit, which means 3 seemingly low or medium vulnerabilities may actually be critical depending on what I can get to. This is like somebody breaking into my condo, through the door I mentioned over here, and then I have a safe. My safe has a simple passcode of 1234. So now they break in the door, they're like, oh, there's a safe, there's a vulnerability to the safe, it's a lame password. They get in the safe and they're able to steal my Rolex watch. I don't really have a Rolex watch in the safe, but if there was, you get the point. And I wouldn't really use 1234 either. But that's the point, I'm able to chain multiple vulnerabilities together to get to the loot, which is the Rolex watch. A lot of people like to use the term loot in, in cybersecurity, but it's really, in terms of medical devices, to cause some sort of harm to the patient. That's the true risk we've been talking about. The other part of vulnerability testing that's required by the FDA is what's called closed-box known vulnerability scanning. This is where we scan the device as shipped This is also known as black box penetration testing. So with our penetration testing, we do black box, which means we don't really know much about the device. We receive it, we try to hack into it. Then we do gray box, which means we know a little bit about the device, user-level knowledge, we have the instructions, and then we try to break into the device based on that. And then we do white box as well, which gives us the most accuracy and completeness. Which is where we have the software requirement specifications, or maybe the system architecture diagram, or maybe a data flow diagram. With medical devices, since we're talking about patient safety, we really should be doing white box penetration testing. Another part of the vulnerability testing that the FDA requires is called SCA, or software composition analysis. Technically, this is not under the umbrella of penetration testing. It's just a different type of testing you should do. Which is looking at the third-party libraries and your SBOM. I've talked about the SBOM in other podcasts. The other item they talk about is static and dynamic code analysis. Again, this is not part of penetration testing. The static part is not. Penetration testing does exercise dynamic code analysis, which means we're trying to break into the code or analyze the code as it is running. That's dynamic code analysis, and we do that through penetration testing. So a lot of people ask the question, does your penetration testing include DAST, or Dynamic Application Security Testing, which is the same as dynamic code analysis? The answer is yes, because we are testing the code as it is running with specific tools that fulfill the requirements for dynamic code analysis. Static code analysis cannot be tested during penetration testing, because the code is not running, the application is not running. It's just the code itself, which requires a different type of tool to analyze. There are 5 things in addition to what I talked about, 5 elements that the FDA wants to see in a penetration test report. They want to see all the tests I just talked about, so that gives you the accuracy and completeness. They want to see risk calculated in terms of patient safety. Which means you need to have the hazard analysis. They want you to understand the context of the device, how it's used, which the threat model can tell you, along with the indications for use. So we have to have the— all that to figure out the true risk in terms of patient safety and do all the tests we just talked about. The 5 elements they also want to see— these are not related to the tests, these are just 5 elements— are the independence of the penetration testers, So third-party independence and the technical expertise of the testers. So are they competent to perform the penetration test against the medical device? There are various ways to measure the competence. That could be experience. These testers have tested hundreds of medical devices. They're very familiar with ISO 14971 and AAMI SW96. They follow a specific methodology that the company has. Like we do at Blue Goat Cyber. There's also certifications, and this is where it gets a little muddy. There are a lot of cybersecurity certifications, but most of them, well, I would say all of them, pertain to traditional IT penetration testing or traditional IT cybersecurity. Just because somebody has a CISSP doesn't mean they know anything about medical device cybersecurity. Same thing with Security Plus, as an example. It means they know a little bit about cybersecurity, but not specifically medical device cybersecurity. So, the technical expertise should primarily be driven by the training, by the understanding of the regulated industry, by the risk methodology, and by their experience. The other item of the 5 elements the FDA wants to see in the report is the scope of the testing. The scope should be the entire medical device that is being submitted to the FDA. So whatever is part of the 510, De Novo, or PMA, or IDE submission, that all needs to be tested. We've seen a couple problems with this come up in the past where clients have asked us just to test the physical device and not the mobile device or the cloud, but their submission for the 510 is the physical device, the mobile application or the mobile device, and the cloud. So we have to test all three of them. The scope of the testing has to match the scope of the submission for the device. The other item is the duration of the testing. The duration of the testing— how long did it take from start to finish to do the initial test? How long did it take to do the retest? And overall, what was the length of the testing? This is very important. And another item to add to this is the final round of testing needs to be done within 6 months of your submission date, within 180 days. The fourth item is the testing methods employed. I talked a lot about already using abuse and misuse cases, robustness testing, fuzz testing, attack surface analysis, vulnerability chaining, closed box testing or black box. So we have to make sure that all of these are articulated in the report. So we have completeness of testing in the report. There are really two things we want to have. We want to have completeness and accuracy. The third thing we want to have is traceability, which I've talked about as well. We— that's why we require the threat model. For the penetration test to have that traceability between our test case and the threat ID to show from a traceability perspective that this threat that we think we mitigated with a specific control was actually tested with this specific penetration test case. And the fifth thing that the FDA wants to see are obviously the results, the findings, and any observations. We include all these in our penetration test report, obviously. The observations are important because some things you may observe may not be a finding, but may be worth bringing to the attention of the client. Like during the test, you notice periodically the system experienced some sort of anomaly. It wasn't necessarily a cybersecurity vulnerability, but every now and then it had the weird anomaly. It's worth Writing that observation down so the client knows about it and can fix it. And also the FDA wants to see those observations that come up as part of the testing. The other item that is extremely important, and we do this obviously at Blue Goat Cyber, is reproducibility. For every test we do, we tell you the exact version of the tool, the exact steps we followed for every step along the way, and how we accomplish the objective. This makes our tests very reproducible. So if there ever is an issue and somebody wants to reproduce our finding, we know exactly which version of the software we tested or version of the product. We know what tool we use. We know exactly step by step by step what we did in our test cases. And we know exactly how we arrived at our finding. So our tests are very reproducible. And whenever findings are discovered, we work with our clients to remediate the finding. We do not do the remediation ourselves. A lot of clients will ask us to do it, but that removes our independence. We will give them guidance on how to remediate what we discovered, and then we have to do a retest. And at BlueGoat, we include unlimited retest. Some clients only need a couple rounds of testing. Some clients need dozens of rounds of testing. Uh, for a very complex system, we often have many, many findings, and when they fix one thing, they may break something else or introduce new vulnerabilities. So we have to continue that testing until all the fixes are confirmed and the risk is reduced to an acceptable level to a patient. A few areas where we've seen submissions fall apart because a client chose the wrong penetration testing firm, more of a traditional penetration testing firm. Then they received deficiencies. Then they came to us to have us redo the penetration testing to help fix their deficiencies. The common things we've seen, and I've talked about these indirectly, is that the test scope doesn't match the threat model. So that means there's no traceability. That means we have a gap in the threats we've identified and the tests we performed. If you think about it from my condo, I've got windows and I've got doors, but if I say this door, right here has a threat that somebody could go through it because the lock may not be secure, but I think the lock is secure, but then I never test that door and I have no traceability that the door, that door was ever tested. That probably makes you suspect about how secure my condo is, right? So every threat that's identified needs to be evaluated and tested. The other part is there is no tie to the hazard analysis. If you do not tie your impact to ISO 14971 or the hazard analysis for your penetration test, you have no context and you cannot do the penetration test properly because you're leaving out what matters most, which is the impact to the patient. So we have to know if I'm able to break into this device and affect the integrity of this device, What is the cause to the patient? What is the severity of harm to that patient? Without knowing that, you cannot properly calculate risk. Otherwise, you're just guessing. And the hazard analysis is something the client or the manufacturer needs to provide to the penetration testing company. Other areas, the report is often missing the duration or the independence, or the things that are required that I mentioned by the FDA, those top 5. Something else we see quite a bit, the testing is on a different version of the product than the product that is submitted to the FDA. The product that is submitted as part of the 510, de novo, PMA, the exact product is the one that needs to be tested, not an iteration later on. Again, if you make any changes, you have to redo a penetration test. The final thing is, and we see this quite a bit, is we will see companies decide to go ahead and close a penetration test finding. They'll say this is a false positive, as example, but there's no rationale, or the rationale is very weak. So if you're going to say something is closed or it's a false positive, you need to justify it. And that needs to be extreme justification, not just, this is not applicable to our build. It needs to be justified why that's a false positive or why you're closing that finding. So I hope you learned something about medical device penetration testing. It is drastically different than traditional penetration testing. As I mentioned, a few of the key elements I've talked about, it has to have traceability and tie back to the threat model. We have to look at the hazard analysis. We obviously want to do white box testing whenever possible because we're talking about patients. We're not talking about data. You can recover from your data being stolen. You cannot recover if somebody hacks into a drug infusion pump and increases the flow rate of a drug that kills you. So the risk needs to be tied to that patient. And this is very important. I mean, we're talking about patient lives here and patient safety. So we can't just treat it like a traditional penetration test. We need to be as accurate and complete as possible. So thanks for tuning in. I hope you learned something about medical device penetration testing and what the FDA is looking for. If you ever need penetration testing for your medical product, you know where to find Blue Goat Cyber at bluegoatscyber.com, and we're happy to help.

    Hosted by

    Explore every episode in the topics covered here.

    More from your host

    Other episodes diving into Christian's areas of focus.

    Episodes covering similar ground - including Pen Testing, FDA Premarket, Threat Modeling.

    Why this matches shares the Pen Testing, FDA Premarket, and Threat Modeling topics.

    Why this matches shares the Pen Testing, FDA Premarket, and Threat Modeling topics.

    Why this matches shares the Pen Testing, FDA Premarket, and Threat Modeling topics.

    Why this matches shares the Pen Testing, FDA Premarket, and Threat Modeling topics.

    Listen to this episode