This interview is in media cooperation with Frankfurt Forward (https://www.frankfurtforward.com/startups-of-the-month). We talk to MINDS-Medical (http://minds-medical.de/), the Startup of the Month July 2019. In this interview, we talk to Lukas Naab, the startup’s co-founder and CEO (https://www.linkedin.com/in/lukas-naab-42b6a423/). In the interview, we learn how Lukas went from studying medicine to political sciences and comparative religious studies, only be back with a startup in healthcare in 2016.
When I saw that some of the processes in the health care system suck, I just asked myself „how can we fix that”.
Learn more about our Affiliated Marketing here: https://www.startuprad.io/blog/affiliate-marketing-at-startuprad-io/
Minds Medical’s algorithm helps with unstructured medical data, in the coding of patient treatment and the invoicing of them. Their main tool is an OCR software, with an NLP-AI behind it. The team of MINDS-Medical estimates a 50%-60% productivity improvement in coding processes with their tool.
50.000 medical entries are recoded each day in the German health care system.
Minds Medical’s algorithm can also help with medical risk analysis for insurance companies. The tool can be applied in health insurance and life insurance. The MINDS-Medical team estimates a 30% efficiency improvement with their tool. The tool currently helps to improve the decisions with their tools by 10-15%.
10% of the hospitals in Germany are in danger of going bankrupt right now in Germany, part of the problem is the re-imbursement.
The startup is an Unibator Alumni (http://goetheunibator.de/?lang=en), which is the startup accelerator program of Goethe University in Frankfurt.
Digitization for us starts, when handwriting stops 😊
The startup already received seed financing in 2018 and is already talking to potential investors for a series A financing round end of 2020. If you want to talk to them, reach out to us.
During the interview Lukas and Joe are talking about:
• Zentrale Vergabestelle für Studienplätze (available in German only on Wikipedia): https://de.wikipedia.org/wiki/Stiftung_f%C3%BCr_Hochschulzulassung
• Appendicitis: https://en.wikipedia.org/wiki/Appendicitis
• DRG reimbursement (link not from the German System, but in English): https://www.verywellhealth.com/how-does-a-drg-determine-how-much-a-hospital-gets-paid-1738874
• OCR https://en.wikipedia.org/wiki/Optical_character_recognition
• NLP https://en.wikipedia.org/wiki/Natural_language_processing
• Berufsunfähigkeitsversicherung (yes, that is a proper German word) translates to Occupational Disability Insurance, here is an example in English, but again not covering the German system: https://www.investopedia.com/terms/o/ownoccupation_policy.asp
• Interview ottonova (audio only): https://www.startuprad.io/interviews/exclusive-first-english-interview-newly-founded-digital-health-insurer-ottonova/
• Interviews with the startups of Merck Accelerator Program (sponsored content): https://www.youtube.com/watch?v=ZHJfUAZyntU&list=PLUbiCAQc22WtH8HnTZ6FJxVw5rV1i8Ntj
When I saw that some of the processes in the health care system suck, I just asked myself „how can we fix that”.
Narrator: Welcome to startuprad.io. Your podcast and YouTube blog covering the German startup scene with news, interviews and live events.
Joe: This is Joe from startuprad.io your startup podcast and YouTube blog from Germany bringing you today the startup of the month in our media corporation with Frankfurt Forward, Minds Medical and Lucas. Hi, how are you doing?
Lucas: Hey how are you doing? I’m pretty good thanks for having me on the show.
Joe: You’re very welcome and thanks for making time on this lovely Saturday afternoon for our recording. Unfortunately, we are unable to bring you this live here but we are working on it. So, you are the startup of the month, congratulations.
Lucas: Thank you very much.
Joe: How on earth did you end up in startups? Because I did a little bit of stalking on your LinkedIn profile you studied Political science and now, you’re in healthcare startup can you help us to understand?
Lucas: Yeah, my CV is not the straightest of all CVs in that way. I started actually to study medicine back in the days when I graduated from high school and after that but realized quickly that medicine is not my field of expertise, so I went into political science and a comparative religious studies and did my master’s degree there. I worked at the university for quite a while and after that in school. But yeah I realized that the health care system could use some improvements especially in the administrative field and I’m a guy who always looks at processes and when I saw that some of the processes in the health care system just suck I asked myself “okay how can we fix that?’ and then I asked Matias my co-founder and a very good friend of mine and we’ve known each other for 20 years if we could build a solution for that and he is the technical guy, I am not and then we started Minds Medical. So, I that’s how I ended up in the startup world in 2016.
Joe: You, went to study medicine. For everybody who’s not from Germany, there is nothing like pre-med here in Germany you start from the beginning and they teach you everything pressurized. It’s very tough. And many friends of mine who studied medicine usually compare it to memorizing a phone book.
Lucas: Yeah pretty much like that plus a lot of other stuff. Well, I got lucky to get a place in the university because in the German healthcare system or to get a place at the University for medicine, you need to be quite lucky. Some people wait for months and years and years for that. I got lucky to get in and then I forfeited my place four semesters later, but what can you do? Sometimes you see that it’s not your actual passion to do that and I left.
Joe: We may tell the people who are listening to this off viewing this on YouTube, IG TV or somewhere else and that in Germany there is something, okay pay attention it’s “Zentrale Vergabestelle für Studienplätze”.
Lucas: That’s a great word
Joe: Which means there’s a central place that hands students to the universities it’s not like in the US or in the UK where you apply to university or your college. Some of them are centralized and basically, medicine is usually done with a very high GPA. Wherein Germany you don’t know the curve so if you want to get an A or in Germany one, you have to know everything.
Lucas: Yeah, so sort of like that. But I think too if you mark the right universities to apply for them because in the ZVS that you can actually check some boxes where you want to go and if you are a little smart there then in some years you may be lucky that you can get in without knowing everything.
Joe: It’s the central administration for… the literal translation would be submitting students for whatever. So Central Administration whatever.
Lucas: That’s one of the cool things in the German language that you can just put words after words after words and then you have a very meaningful word for that. It’s hard to translate.
Joe: My all-time favorite is in English-speaking countries you have, on the product “best before” for example food in Germany there’s Mindeshaltbarkeitsdatum (MHD). Okay, your company is Minds Medical. You’re not doing one thing you’re doing two things. Can you tell us a little bit what you guys are doing and how it is different from what’s going on right now in the German medical field?
Lucas: Sure, we started our company to, as I said before, improve administrative work within hospitals. So, our product is a software-based on machine learning and natural language processing to read and understand medical files and the unstructured data in medical files for that matter. So, if you are in a hospital and you are treated then the doctors obviously, have to write down why you went to the hospital and what they did with you. And this usually is in the form of the medical health record within the hospital. By the way, about fifty thousand medical records are produced every day in the German health care system in hospitals so it’s quite a huge number of medical records that are produced. And obviously, it’s written as texts and after the patients are discharged, what do you do with the medical record? One thing, one part is to know what happened to the patients when he comes back to the hospital later but also for reimbursement purposes. Because the German healthcare system is based on the DRG system which is the diagnosis-related groups which are reimbursed from the payers in the healthcare system, the Fallpauschale the insurance companies. And the diagnosis-related group the name already says it a little bit, is based on the diagnosis you had when you came to the hospital. So poor souls have to read the medical files after the patient is discharged and to do a thing called ‘medical coding’ which is transferring medical contexts. Joe was admitted to the hospital with an appendix is for example and we removed it and stuff like that so the appendix gets an five-digit code and the removal gets some other code and that’s a fun task with about 80,000 different codes that people that are medical expert sand Rain Man have to get out of, of the code so it’s a manual process. And Minds Medical it’s partly automating that process by reading and understanding the medical record and knowing what the doctors wrote down there. So, it saves time in hospitals to make it short. It saves time for reimbursement purposes. And so, 20% of the doctors are doing this medical coding in Germany after the patients are discharged, which is a colossal waste of their expertise because they should be treating patients and not doing administrative work. Or in most hospitals, nowadays so-called [inaudible 00:07:51], coding professionals are used to do the medical coding in the hospitals and they just aren’t enough of these people. Because you need medical expert plus rain men to do that and many hospitals have a huge pile of patient files that are not coded and so we help with that.
Joe: Is it also the case that if they have unencoded files, they’re not getting paid for them?
Lucas: Yeah, true. So many hospitals have, I don’t know a couple of hundred files patient files lying at their desk and they don’t have the manpower to do the medical coding for the reimbursement. So, they’re not getting money from that. And in Germany, the rule is that after 30 days of the patient being discharged from the hospital, you have to hand in the medical coding and the DRG, group or the hospital that the patient is reimbursed for to the insurance company. So, there is a time and money issue with the hospitals. Ten percent of the German hospitals are in danger of going bankrupt at the moment in Germany and one of the reasons is that they don’t do the reimbursement in time.
Joe: And I just looked it up. And for everybody like patients that means money if you don’t get paid for your patients. For example, I just realized that one day in the intensive care unit costs something like one thousand six hundred US dollars and so if you have a patient who’s like in there ten days that serious money already for a hospital.
Lucas: That’s serious money. The normal amount of money, the hospital gets for one patient is around two thousand eight hundred euros in Germany for the stay of the hospital. But the group you are diagnosed with basically gives a multiplier for that base number of two thousand eight hundred Euros. So, this multiplier can be as low as 0.4 or something so you get less money. If it’s a very easy case you’re just there for one day, they look at you and they say oh you’re fine you can go home.
Joe: You just ran into the door.
Lucas: Yes, you have a minor concussion or something are there for a day in the hospital then the hospital gets not enough money, but imagine they remove your lungs and you get a lung transplantation and you’re in the hospital for twenty days, the multiplier can go up as high as times fifty or something so it’s serious money to do the coding actually correctly
Joe: I see. And basically, do you help to get this data like it is now in the files? Are we talking about digital files or are we still talking about the handwritten files? And a disclaimer here; doctors in Germany are very well known for the bad handwriting because they have to do a lot of bad handwriting and even late at night and so it gets very soon, very bad.
Lucas: Well, digitization for us starts when handwriting stops. This is the level we are working with but we figured out… in the beginning, we thought okay everything’s going to be in databases in the hospital systems but we quickly… there was a rookie mistake so our software is now able to do a very good OCR so we can read any file, picture file that is sent or produced in the hospital. If you have a scan, we can read it and classify the written text and do the, do our natural language processing machine learning, find the common stuff in the text and then we can correctly code files.
Joe: And since you use the TLA, a three-letter acronym, OCR means optical character recognition so basically you scan something handwritten and then a computer can understand it and translates it into a usual word document or something like that.
Lucas: That’s correct. So, we can work with all the printed files and the hospitals or archive files that are printed out of PDF files and re-scanned and sent back to an archive system. We can all work with that now. But if somebody writes down something with a pen on a piece of paper, we are lost but luckily for most patient cases, that’s not super relevant in terms of reimbursement. So, we’re not a medical product, we’re just doing the administrative stuff on the backend so technically not a medical product, where this, in theory, might be needed to do as well. So luckily, we are not in that position.
Joe: I would say here’s something more like a big data company especially working with unstructured data.
Lucas: Yeah, basically, pretty much with a focus on the medical text. Because in every field there are domain-specific talk, domain-specific things that are easier for machines to read if you train them specifically for these tasks. So, if somebody gives us a newspaper, we would probably be a lot worse in discerning diseases that are written down there because it’s not just a… we’re not doing a, what’s it called? word finder or something like that, that’s not enough. Because if you have pain, that could be something very different than knee pain or acute knee pain or the patient doesn’t have acute knee pain, but he has something else that could be related to that. So, you need always need to look at the whole context and to just find a couple of words that might have medical meaning just doesn’t cut it.
Joe: I see. And usually, you taught your machine the Pschyrembel which is the medical handbook for Germany that explains doctor speech to normal people. For example, one of my all-time favorites is lumbago, you got a backache. You’re trained to understand, that, right? Lumbago?
Lucas: Yeah, we trained the system on the medical files that are already at our customers, at the hospitals for example or insurance companies. So, we can train them, train our system not only on the medical relevance but also on the specific dialect that is written in the hospital. Because you have a 40-year-old or 50-year-old medical professor who has taught everybody how to document this dialect will be different in the hospital then and the hospital right next door where a different guy from a different school of medicine, for example, has a different approach on documenting stuff. So, it’s a little trickier than just feeding the machine with lexicons.
Joe: I see. I do have a question concerning that. You are going to a hospital. At first, they say, “no way”, then they say “maybe” and then you get work with something like a pilot or something like this. Do you already work fully completely integrated into a hospital here in Germany?
Lucas: We are starting to integrate into the first three hospitals at the moment. We have a live product with our first insurance customer. We are doing something a little bit different than the medical coding for reimbursement purposes. So, as you let’s ask before we have two kinds of products that we are working with and when we figured out that we can actually contextually understand the medical records, what’s relevant in terms of pre-existing conditions with the patients, we figured out that medical risk analysis is the thing we can also do with that. So, if you want to have private health insurance for example, and you apply for that insurance you’re going to get asked some questions. What’s the last five-year medical history you have? If you don’t have any medical history, you’re fine. You get the insurance. But as soon as you have something that is not dealt with in five minutes checking a box, you just have to send all your medical records to the insurance company and people have to read it. It’s the same task as in the hospital, they have to understand what the applicant for the insurance had. And what we can do with our software is to yeah do the medical risk analysis for them because we can look for the diseases they don’t want to insure. And we can quicken the process in a way that it doesn’t take six weeks for the applicant to get an answer if he gets the life insurance or medical insurance, but it could go in minutes. Just send the file, scan the file, take a photo, get the idea, get the answer. And you can decide if you want to take that insurance or not based on the offer, they make you. At the moment, the insurance company is doing the same thing, but it just takes weeks and weeks and weeks.
Joe: I see. Before you shifted into the insurance arm of your company, I was curious. Do you have any guesstimates of how much will you contribute to the bottom line of the hospital’s you’re working with?
Lucas: Yeah, we have a pretty good idea. With our current set of technology, we can automatically and for reimbursement purposes, soundly do about forty to sixty percent of all stationary patient cases to automatically code them so that the hospital gets the money for that. So, it’s about a fifty percent increase in productivity just with the software. What happens then? You have people who are used to reading all the files that come on their table, but they have a little more time at the moment, which helps because before that, they were super overworked and didn’t finish with all the files. And now what they can do is go back to the doctors and help them write in the actual file because they know what’s relevant. So, the medical file, the structure of the medical file gets better and the content of the medical file gets better. And every person in the hospital who is doing that, medic coding professionals who do that contribute about 150,000 a year plus to the revenue of the hospital. If you know that many hospitals operate on about two hundred thousand profit a year, it’s a lot of money.
Joe: I see. And for the insurance part, health insurance here in Germany, because we are only talking right now about you and Germany, there’s that Gesetzliche Krankenversicherung so it’s health insurance by law you can get in. And they’re about a thousand providers, some really big nationals, some small local that you can hop in between. And then if you earn enough money, you can go to a higher level and that’s called private health insurance. And to get in there, you either have to be admitted as a kid, because your parents have been in private health insurance or you have to earn enough money to have the option. No obligation, but the option to change.
Lucas: The public healthcare system is quite good in Germany and we’re lucky that we have it that basically everybody has health insurance. But, if you want that little extra or you make enough money or you have been in the private health insurance and want to change, then you have to do this application process with the medical backup check for example. And this is basically where we come in. But also, on life insurance, if you have an accident or want to insure yourself against some diseases you get and then you get paid some money if you get the diseases. There’s also this health checkup needed or Arbeitsunfähigkeitsversicherung I forgot what’s that in English.
Joe: Yeah, let me quickly check on the internet. Yeah that was also going through my mind and then I was going “oh, I don’t know the English word”
Lucas: We realized quickly that the health insurance system generally is very tribal with its very specific kind of words.
Joe: Oh yeah. Occupational disability insurance.
Lucas: Now you know.
Joe: So now you know guys. So, remember Arbeitsunfähigkeitsversicherung 🙂
Lucas: So, for that kind of stuff. Or we can do claims management. If you went to the doctor off private health insurance the doctor does something and he writes down the doctors’ letters and the insurance is left with the Doctor’s letters to pay for that, to pay for the doctor service. Then, they also have to read it. So medical risk analysis at the point of sale, medical underwriting and automated claims management. This is the stuff that we do for insurance companies.
Joe: When I was reading about the insurance companies and do you believe there are some people who are really going to like it because they’re healthy and they get lower premiums but they are also are people who are really going to hate it because they are sick and they usually could not get into the insurance. How do you deal with this ethical dilemma?
Lucas: Well, for us, we help in a process that is already in place. So, if you want that insurance you need to go through the backup check. And it doesn’t… the question is “is the decision the insurer is going to make to admit you to the insurance or not data-based or is it somebody who forgot to drink his coffee in the morning, is very distracted because he is stressed at home, reads through your file, sees something and says well I don’t give a crap he gets the insurance or not and gives you a big fat ‘no’ on the admission form or if the decision is data-based? And we see that the diagnosis we make out of the document is generally 10 to 15 percent better than the humans understanding the text.
Joe: Is it also that you contribute then 10 to 15 percent of the bottom line to the health insurance company or are you looking for more when your product? …
Lucas: Well, the market for medical risk analysis and stuff like that in Germany is about 5 billion I believe. So, it’s a huge market. We’re looking for about 30 percent in cost reduction in the process. We do not have enough data at the moment to say if the bottom line for the insurance company is 15 percent 20 percent or something like that. But what our software enables the insurers to do is scale sales. Because if you get admissions from potential customers, you have to read them. You have to check if the data says I can insure you or not because every wrong decision you make goes directly to your bottom line as an insurance company. And with our software, you can scale the admittance of the right customers and that’s super interesting for most insurance companies.
Joe: I see. Everybody who’d like to learn more, we do have a little audio interview from Auto Nova, Germany’s only digital health insurer. Go down here in the show notes, you’ll find a link to the interview. And there in the interview, I believe I promised an update so maybe I’ll reach out to Auto Nova today. Talking about that, people don’t see it through the window in the back but actually, you are located here in lovely Frankfurt.
Joe: Since we are talking as media partners of Frankfurt, what does Frankfurt right now mean for you? Why did you set up shop here?
Lucas: Well I’m originally from Frankfurt. So, I’m one of the very few founders I am a “Frankfurter Bub” as we say it. So homegrown Frankfurt dude. And my co-founder Matthias and I, we were located in the area here in our previous work lives. And when we decided to start the company, we thought “well we might as well stay here in Frankfurt”. Frankfurt is not particularly known for its intro tech or health tech companies. But I think Frankfurt, it’s a great opportunity for startups in that area because we are a B2B business, we have to visit our customers face to face. So, Frankfurt has the big train station that goes all over Germany, the airport and everything and is very well connected. Also, to the finance industry. So, we’re right in the middle there so this means a lot to us to be here. But also, Frankfurt is home, just plain and simple. My wife and I live here, my family lives in and around Frankfurt. Same with Matthias my co-founder. So, we enjoy it here. We have a big network of friends and family and that’s super helpful. If you want to hang out and want to meet your friends and then you don’t have to travel hundreds of miles to be home sometimes. Also, we happened to, well at least I and Matthias as well, happened to study in Frankfurt when we started our careers. And that’s why we got admitted also in the Uni Batwa which is the startup incubator of the University of Frankfurt which we were alumni from. So that’s also a big thing for us, to be part of the Frankfurt ecosystem in startups because we are alumni of the incubator program of the Good University. And for us, after having all these benefits from the Good University, from the incubator program and all that, it’s natural for us to stay close and to give back to the ecosystem if we can. That pretty much sums it up.
Joe: What came to my mind is Frankfurt may not be known for its medical and data startups, but they should. I do see quite a surge there. And also, Merck is running a very good accelerator program down in Darmstadt. Disclaimer; they are a sponsor of mine and that’s why I know a lot of those startups.
Joe: Well one more question.
Joe: If your life would be described by the name of a book or a movie or a combination of both, what would it be and why?
Lucas: Life or a book?
Joe: Yeah, a movie or a book.
Lucas: A movie or a book
Joe: Yeah, the title of your life.
Lucas: We may have to cut the interview here later on because I need to think for a minute.
Joe: I always say it’s between Pinky and the Brain and Forrest Gump.
Lucas: Since I’m a big fan and also a long journey, maybe the Hobbit.
Joe. The Hobbit, yeah.
Lucas: So, when Bilbo the hobbit stumbles into Gandalf and he just picks him up on a journey and takes him to all these crazy places he had no idea that they even exist. That pretty much sums it up and fighting the German health care system for better operability, better processes that certainly feels like facing Smaug.
Joe: That is freaking good. I’ve looked in my show notes and there’s one thing we talked about before. You raised a seed ground in 2018 and you’re now looking for potential investors for series A maybe in 2020.
Lucas: Right. Pretty much. What we’re doing at the moment is doing our proof of markets and have working products as well in hospitals and insurance companies and build a couple more products at the moment and then we’re looking at the end of next year probably for serious A-round to scale that up and be a little quicker about what we’re doing at the moment. But then we have all that proof and money in the back, and we could just grow normally and slowly, or we could just pick up a little pace. And that’s what I want to look for series A by the end of next year maybe 22, beginning of 23 or something like that. The healthcare system always takes longer than you think.
Joe: Ok great. So, the only thing for me left to say is thank you very much I enjoyed this interview quite a lot. And for all the listeners and the viewers, we took out approximately 30 minutes of thinking time for you, right?
Lucas: Yeah approximately. Maybe, an hour and a half as well. I don’t know.
Joe: Thank you very much.
Lucas: Thank you, Joe, and have a great day I’m looking forward to seeing the interview later.
Joe: You too. Bye-bye.
Lucas: Thanks, man.
Narrator: That’s all folks. Find more news streams, events, and interviews at www.startuprad.io. Remember, sharing is caring.