Fortnightly Healthtech Update #17

Lengthy article on the evolution of mobile phones for digital phenotyping (ie. continual data collection of the individual). 

CareSignal (formerly Epharmix) gets funding from customers and others for remote patient monitoring. The beauty is, it’s a relatively low-tech solution that can be used by many diverse patient populations with different conditions. See studies on COPDdiabetes, and mental health. Often, simple really is better.

The direct primary care model continues to evolve with Clove Health claiming to be the first legally incorporated as a public benefit corporation getting to work in Florida. More on that style  of incorporation hereCitizen Health is heading down a similar path.

I touched on the ongoing battle over patient access to their own data. Apparently the federal government has started hitting providers with fines if they make it difficult for patients to get their data.

From Imperial College London, using sound to detect vital signs allows a device to penetrate layers of clothing.

Medicare builds on the mixed success of ACOs to create the Direct Contracting model for the next stage of value based care. There’s also the Primary Care First model, now due sometime next year. Also, a fresh study of Medicare ACOs finds that much of the cost variation comes from the use of out-of-network primary care docs.

More Medicare, the CJR program was controversially introduced as a hip and knee replacement bundle a few years ago. Controversially because it was a mandatory bundle introduced because providers didn’t sign up for the similar voluntary program with sufficient enthusiasm. Now CMS wants to extend it for another 3 years, and add in outpatient settings too. Which makes sense because it will drive the cost down, all other things being equal.

A new design to make better, cheaper biosensors for fluid analysis. Quite honestly, it works in ways I don’t fully comprehend, but hopefully it means something to some of you…

More healthcare reform, CMS picks 205 EMS services for the experimental ET3 model. The intent of ET3 is mostly to help people with chronic conditions get treated in their homes, avoiding the personal discomfort and stress of perpetual trips to the ED. It also should save Medicare money on avoidable ED visits too. To do that, ET3 allows EMS to get reimbursed for other services, not just transport. This potentially opens up another route to market for medtech vendors in applications like remote patient monitoring.

Wearables and machine learning start to show real promise: PhysIQ and VitalConnect study shows promise for predicting hospitalizations for heart failure patients. We’re already seeing machine learning breaking into imaging. I think continuous patient monitoring also has real potential. There’s a big need to determine baseline vital signs for individual patients, rather than just using generic values. I think machine learning might have the potential to usher in more adaptive algorithms that can help to reduce the long-running over alarming problem.

Apple leans hard into atrial fibrillation. A new collaboration with J&J makes the Apple Watch available for $49 to seniors who take part in a study. This could be a win for everybody: Seniors who avoid a potential stroke through the early detection of afib, cardiologists who get more business, and obviously Apple. Presumably J&J has some meds for managing afib in its portfolio too. And Medicare could win, since stroke care is reckoned to cost $40bn a year.

Sepsis has been a silent killer for a long time. A new study finds a stunning growth of 40% among Medicare beneficiaries who were hospitalized. Digging deeper into the economics, the same study finds the total cost of treating sepsis to be much higher than previously thought – over $23bn for inpatient care alone.

At the other end of the age scale, Children’s Hospital of Philadelphia has developed a new algorithm for detecting pediatric sepsis.

A new potential tool in the opioid crisis, a wearable to detect opioid induced respiratory depression (OIRD) in the community. Developed by Altair Medical in Scotland, I’m interested to see how this develops and works in practice. There are definitely challenges in current opioid addiction treatment that this could address. (For podcast fans, Freakonomics has a good two-parter on the crisis and current approaches). Also, equal parts tragic, genius, and sheer practicality, a number of communities are training kids to administer Narcan. Note, OIRD is still a major problem for post-surgical patients too, where hopefully the patients location and time to rescue are more predictable, so potentially more effective overall.

Simple early warning scoring is still an undervalued (and under adopted) practice for catching patient deterioration early. One hospital system in the UK dropped cardiac arrests by 75% by capturing vitals at the bedside with iPads and calculating the early warning score automatically. 

An unfortunate reminder that technology alone can’t change anything: WellSpan Health York Hospital gets cited for failure to respond to a patient in distress. It seems even thought the patients falling heart rate and oxygenation levels were noted for at least 20 minutes, nursing staff failed to intervene.

Fortnightly Healthtech Update #16

Walmart moves into radiology, opening a clinic in Calhoun Georgia with x-ray facilities. Which is a nice lead in to will big retail disrupt healthcare more than big tech?

Profusa’s Lumee tiny oxygen sensor gains a CE mark. Intended for conditions such as peripheral artery disease, the device operates for 28 days, clinical study here.

Apple highlights it’s healthcare presence in its earnings release, claiming 100% of F500 companies in the healthcare sector use Apple. Which sounds impressive, but when you think about it, that’s a small number. You might even be able to count those companies on the fingers of one hand. 

Medicare’s bundled payment isn’t perfect, it rewards doctor’s who take lower risk patients. Really would be nice if it worked the other way around perhaps.

The legal fight over price transparency continues. Meaningful transparent hospital pricing would lead to meaningful competition among hospitals. Competition among hospitals would lead to lower prices. Hence, the AHA is obliged to fight it on behalf of its members. But the American Academy of Family Physicians is not.

On the theme of overpaying, Medicare paid $7.8bn for insulin. If it paid the same prices as another federal government department (the VA) it would only have paid $3.4bn. That is just one of many examples of Medicare overpaying because of a lack of political will.

Verily partners up with Santen for vague eye are projects.

Continuous vital sign monitoring in hospitals is becoming a thing at glacial pace. (See Isansys in the last post). This study compares the performance of devices from EarlySenseSensiumVitalConnect, and Masimo. In summary, all are accurate for heart rate, some better than others for respiratory rate.

The lack of continuous monitoring still hits the broader headlines occasionally. In this case,  Cleveland area hospitals are highlighted for their tardy adoption following the tragic loss of a patient in 2016. 

Also continuous monitoring, continuous monitoring of urine for the early detection of kidney problems from Serenno Medical.

Virta Health – treating diabetes the old school way with diet – has been running a pilot with the VA for a year.

Scientists in Korea have developed a biosensor test for the early detection of Alzheimer’s.

An opinion piece from a Professor at Stanford University School of Medicine, outlining how a simple market segmentation would improve healthcare delivery. In practice, I think market forces are already doing this to some extent. The reasonably wealthy/reasonably healthy are starting to adopt concierge medicine, or direct primary care. But, I can’t figure out how direct primary care is going to work for the chronically ill. The direct primary care model often touts unlimited primary care access as a benefit. But, unlimited access to a primary care doc is just like having unlimited paid time off. It only works if you don’t take the “unlimited” literally. And that won’t apply to many chronically ill people with multiple comorbidities of course. And if that mild criticism wasn’t enough for you, Molly Osberg well and truly shreds direct primary care – the comments are definitely worth a read too.

Oh, and Iora Health, one of the pioneers for direct primary care, just closed $126m in funding.

Apple finds itself squaring up to Epic and assorted hospitals in the fight to allow patients to easily access and share their healthcare data. Epic and friends voice concerns about patient privacy. But, dig deeper. Allowing Apple access to health records is a long-term threat to Epic’s dominant position in the EMR business. And, as noted above (on price transparency), hospitals aren’t keen on competition either. Preventing people from easily moving their data from one provider to another is just one more way to lock in a patient’s revenue stream. Which is really unfortunate. Because, if consumers had better access to quality data, better access to pricing, and could easily move their health records from one provider to another, hospitals that delivered greater value to consumers would ultimately thrive. And those that did not would either raise their game or wither and die. Everybody wins – except sucky hospitals.

Also in the EMR connectivity game, Innovaccer raises $70m to create longitudinal patient records to support the transition to value-based care. I imagine it will find “patient privacy concerns” to be a challenge too. 

Caretaker Medical of finger-cuff blood pressure fame, adds ECG via a partnership with VivaLNK. Ditto VitalConnect, partnering with CorVitals to add arrhythmia detection.

Researchers in New South Wales come up with a superior algorithm for early detection of sepsis.

Machine learning is steadily making inroads into the diagnostic space. PhysIQ adds a patent to estimate cardiopulmonary function from wearables.

Challenging assumptions or a little Christmas humor…Parachute use to prevent death and major trauma when jumping from aircraft: randomized controlled trial.

Fortnightly Healthtech Update #15

An industry analysts pretty jaded view of the state of innovation and meaningful change in healthcare from the JP Morgan Healthcare Conference.

Isansys claims a small deployment down under as the world’s first hospital-wide wireless monitoring install. If you know a better claim to first, please let me know….

The internet of insecure things…..GE has a problem with the security of its patient monitors.

Hips and knees are falling out of favor with participants in Medicare’s bundled payment program. Two reasons are given. First, providers have wrung all the excess cost out of the bundle – primarily by cutting skilled nursing for rehab. Second, the shift to outpatient surgical centers might increase the risk. As noted in the last fortnightly, this saving has been achieved without a drop in quality. Good news is, providers are moving onto more complex bundles. Let’s hope they don’t lose their shirts.

For a well-sourced read on – well, just about everything wrong with US healthcare – thank the American College of Physicians. One of the observations is that an eye-watering 31% of US healthcare spending goes on admin. A number that is so high in my view because of the massive fragmentation of both providers and payers. That complexity just needs so many administrators to coordinate everything – and often still dropping the ball in my personal experience. How can we cut that admin overhead that we all pay…? One way is single payer (aka Medicare For All). Another might be direct primary care with healthcare cost sharing ministries.

More on what the US could learn about universal healthcare from other countries here.

And One Medical, a direct primary care startup, has filed for an IPO. Direct primary care for the price of a latte a week

And yet more on new primary care models here.

The personal health wearables business has another casualty with UnderArmour dropping out.

Plenty of players (eg. Isansys above), but precious little adoption: VivaLNK elbows its way into the wireless vital sign monitoring market.

As does BioIntelliSense, with FDA clearance for a 30 day patch to monitor vitals in the home. The home use and 30 day life position the BioIntelliSense BioSticker squarely in the “readmission prevention” market, so it should appeal to many hospitals. Not clear to me if the device is semi-reusable, but that would help to keep the unit costs down if that’s true. Most interestingly, the company is pursuing a subscription model, with sensors provided free to providers.

Evidence that Medicare is overpaying docs for post-surgical follow-up that never happens. On the face of it, this could be an argument against bundled payments. But it’s not. Bundled payments pay for outcomes. If a doc decides they can achieve a perfect outcome without burdening the patient with unnecessary follow-up visits, good luck to them. Let them keep the extra, because the total price of that bundle should be cut by a fraction each year to encourage innovation in clinical practice.

McKinsey has a piece on what hospital care will look like in 2030. It’s mostly a rosy picture for patients – mhealth apps that avoid hospital visits altogether, online appointment booking, no waiting rooms (just-in-time visits), walls that change color to reflect or enhance a mood etc. It’s all good – but overlooks the obvious question. Why would our healthcare industry pay for any of that when they don’t get reimbursed for it in turn? And hospital visits aren’t going to dry up anytime soon while the American Hospital Association spends $26m a year on lobbying. Unless we push really hard on the ACO model and make sure value-based care fulfills its promise. Better outcomes at lower cost.

Tangential I know, but US life expectancy actually increased for the first time in four years. But only by a month. Lower death rates from cancer and opioid overdoses are the reason.

Alphabet’s Verily playing catchup with Apple, adds FDA cleared irregular heartbeat capability. Although Verily is more focused on clinical trials than Apple is on the face of it.

Allegheny Health Network reports both good financial and clinical outcomes by adopting a faster test for sepsis that gives results in just 90 minutes. Aiming to better that, researchers in Switzerland are working on a sepsis test that can be completed in 15 minutes.

Fortnightly Healthtech Update #14

A glasses-attaching wearable to monitor eating habits – because I’m pretty sure we all lie to our calorie counting apps at least some of the time…. 

CMS loosens up the purse strings to try digital health for moms-to-be and kids with complex needs.

Some positive results for the Medicare bundled payment experiment, lower costs for hip and knee replacements with no loss of quality. The study doesn’t see any change with other bundles though (there are 48 possible bundles in total). That doesn’t surprise me – and it’s not necessarily a bad thing. By far the most popular bundle with hospitals was hips and knees. Basically, it was a very low risk way to get into the bundled payment game – a relatively simple, repeatable procedure. So, the reason why we’re not seeing savings with other bundles yet could just be a lack of volume for providers to learn from.

Not an option for those that sleep in the buff, smart pajamas for monitoring vitals and sleep patterns.

Boston/Paris based Cardiologs picks up $15m in funding for its afib diagnostics algorithm. Abstract for a clinical study here.

Ah, the tensions of the journey to value-based care….BCBSMN tries to steer patients to out-patient clinics for lower cost care. Hospitals that stand to lose out, sue.

Why aren’t patients electronically accessing their medical records? Because – IMO – there’s rarely anything to be gained by doing so….I can go to my docs portal, I can login, I can see my data, and then…so what, there’s nothing useful to do with it.

VC’s clearly think bundled payments are here to stay, sinking an extra $27m into Aver. They might be right, unless we all go down the ACO path instead. Or just pull the plug on value-based care.

More potential lawsuit woes for Apple, being sued for patent infringement by Masimo. The basis seems to be that instead of licensing the technology, Apple hired key people away. Quite a nifty strategy if you can pull it off actually.

Is digital health failing before it’s even really got started…? Maybe. The good news is a Stanford Medicine survey finds that the vast majority of docs see value in self-reported health data. The bad news…a third are looking for help on how to use AI. So that’s a big fail right there.  Docs really shouldn’t need to know how the algorithm works, unless they want to get into research.

Israel-based Clew raises a B funding round for it’s AI-based platform for predicting patient deterioration. 

Can AI reverse the ‘unsustainable’ trajectory of spine care? In a word, no. Because spine surgery rarely brings much benefit.

Interesting use of Fitbit’s to track the spread of flu.

Payers and providers are squaring up to each other on how to address surprise bills. I’m honestly not confident this will help. The problem isn’t surprise medical bills. The problem is just medical bills. I’m mindful of this case from a couple of years back, where a woman begs people not to call an ambulance because she couldn’t afford it.

Rock Health reports that digital health funding dropped a little to $7.4bn in 2019, Still the second highest funding year on record though.

Masimo to acquire some assets from NantHealth, basically a device integration play. 

The Camden Coalition has previously been seen as a great success in addressing the high cost of high utilizers. Unfortunately, that no longer seems to be true.

Enrollment in Medicare’s MSSP program is stalling as CMS tries to get ACO’s to take on actual financial risk earlier. Meanwhile, it’s reported that one of Medicare’s other ACO programs (Next Generation) saved Medicare $184m in 2018. That’s good, but bear in mind the total budget for CMS is $1.2tn….

Fortnightly Healthtech Update #13

As the diabetic population continues to grow, medtech vendors are looking for alternatives to the unpleasant finger stick to measure glucose levels. Continuous Glucose Monitoring (CGM) wearables are a commercial reality, while researchers have explored using tears as a reliable glucose measure. Now, researchers in India have a plan to use saliva.  Talking of diabetes, incidence is lowest among caucasians in the US.

Patient monitoring of a different kind: Roosevelt General Hospital in Portales, New Mexico tells patients to monitor bank accounts after malware infection

A possible speed bump for Apple, a New York doc claims patent infringement for the afib algorithm.

Payers are gearing up to address the social determinants of health, with programs from the University of Pittsburgh Medical Center, BCBSRI, and Cigna. Meanwhile, UPMC are implementing remote blood pressure monitoring using Vivify Health for new moms after they are discharged. The aim is to improve care quality by catching hypertension early, and reduce costs  by managing follow-up appointments more appropriately. 

Last time I noted how the “healthcare system” fails mothers. Apparently if’s failing kids too, with fewer hospitals able to treat pediatric inpatients. But at least the US isn’t unique in that, with the UK facing a lack of pediatric ICU beds.

In Pennsylvania, the Rural Health Model spins up to see if Medicare can incentivize better access to care for rural communities.

Creative accounting of a good kind, payer and provider split the investment in a social worker to try and keep frequent flyers out of the ED.

A doctor’s disarmingly honest account of how a hospital stay often turns out to be very bad for your health. It’s a long read, so you might prefer the podcast instead. 

From the National University of Singapore, an RFID enabled biosensor less than 1mm wide.

Mobile telestroke program cuts door to needle time for stroke patients by 30 minutes. It’s mostly underpinned by simple process changes, enabling critical treatment steps to run in parallel instead of sequentially. 

Apple was recently granted 35 patents, the most healthcare related being a mattress for measuring vitals signs, and exercise intensity via a PPG sensor.

Many of the ancestry/genealogy companies have two revenue streams. Charging individuals for testing their DNA is the obvious one. But, less well known, a second revenue stream comes from selling the individuals anonymized data to big pharma. 23andMe is doubling down there, inviting individuals to add their medical history to the 23andMe database. As the article notes, that could be a hugely powerful dataset – but should it be owned by a private, for-profit company?

On that note, some concern about Facebook encouraging users to share their health data so that they can get recommendations for preventative care. I think there will always be concerns here. Many tech companies change their terms of service frequently, which is always a concern. On the other hand, if it helps people take care of their health better than the traditional “healthcare system” does, it’s worth thinking about.

Related to that, a Rock Health survey finds that only 10% of consumers would be willing to share their data with a tech company (free download). But, before you read too much into that, here’s the exact question that was asked in the survey: “Please indicate which of the following individuals or organizations you would be willing to share your health information with…”. 

So I’ve done primary research for a living. With every data set that I ever collected, I’d always look at the data afterwards and think “Doh, I wished I’d asked <fill in the blank> instead”. In this case, I’m thinking a more insightful question might be “Would you share your health information with a tech company if it saved you $500 a year in your out of pocket healthcare costs?” I think you’d get a very different answer from 10% would be willing to share. Different again if you suggested $100, and different again for $1,000. A more informative approach perhaps. Everything has its value, everything has its price.

Also in the very same Rock Health research survey…Apparently, consumers are significantly less likely to share their data with their own doctor than they were two years ago. Why…? I would guess because consumers can’t see any demonstrable value in doing so. Because sharing data with any entity – a doctor or a tech company – only makes sense if it leads to lower costs and/or better quality care. If it doesn’t, you’re just wasting your time.

A wearable to help identify swallowing disorders.

Progressive, but not clear to me how this would work in practice without consumption on-site: Massachusetts Looks to Open New Drug-Monitoring Centers.

Coincidentally in Massachusetts, Lowell Hospital is working with Frontive to leverage Amazon’s Alexa to try and improve post-discharge orthopedic care.

Apparently house calls by doctors have been decreasing for years. Partly – if this article is to be believed – because the reimbursement was so low. But, for homebound people, there’s real value there. This is something that Accountable Care Organizations should be willing and able to revive. If that home visit – by a doctor, an NP, or a paramedic – can keep someone healthy at home, there has to be a payoff there for an ACO. 

Healthtech Fortnightly Update #10

Clinicians may have the opportunity to use wearables to help PTSD sufferers.

A good primer on bundled payments for anyone that wants to understand more about that value-based care experiment. Oh, and enrollment in Medicare’s BPCI voluntary bundled payment program dropped 16% when providers had to take on risk. A number of reasons why are offered in the article, and they’re all valid to some extent. But, this program was introduced in April 2013. So, in reality, that’s over 6 years for providers to figure out their costs, what causes readmissions etc. Clearly, that hasn’t been a priority for some.

Sensor-based medication adherence firm Proteus Digital Health shows some success with hepatitis C, more clinical details here

Skilled Nursing Facilities (SNFs) fret that value-based care is going to drive them out of business unless they adopt technology. Well, they’re right to worry, because there’s likely to be a bloodbath. There are over 15,000 SNFs, many of them smaller, independent facilities. And many of them are about to get squeezed out of the market. First, Medicare is trying to bypass SNFs altogether, and discharge people directly to their homes, with home health agencies to provide rehab. That’s because it’s cheaper, all other things being equal (although comparing  readmission rates would be interesting…). In addition, in the past, patients needing further rehab have been free to choose any SNF when discharged from hospital. As Medicare tries to make hospitals accountable for reducing readmissions, hospitals need more control over the entire episode of care, end-to-end. So, they are going to develop preferred partnerships with SNFs so they can manage and improve quality. Ergo, there will be far fewer patient days in SNFs, and only the SNFs that can guarantee quality outcomes and integrate seamlessly with hospitals will survive. A good opportunity for companies likes EarlySense and Curavi Health to help out. Even so, some SNFs will go out of business, others will be gobbled up by large SNF corporations or healthcare systems. Expect that 15,000 SNFs to become 12, maybe even 10, thousand over the next few years.

The most impressive thing from the Apple Watch heart study isn’t it’s ability to detect afib. It’s the fact that they collected data from 419,000 people. In the era of AI and machine learning, that’s a game changer. Because, all other things being equal, in the era of AI and machine learning, the person with the most data wins.

The VA obviously likes what it’s seen in it’s diabetic foot pilot with Podimetrics that it’s expanding use to all clinics.

I’ve been of the opinion for many years that the best path to bring healthcare costs under control is to make it a fully transparent and competitive market for consumers. A bit like buying a car, or a fridge, or a central heating upgrade. The push for value based care – ACOs, bundled payments etc. – haven’t really got into that yet. Now the federal government is pushing harder for transparency, and predictably the hospitals are starting to squeal louder. I touched on site neutral payments in my last piece, and predictably hospitals aren’t keen on that either. Because both transparency and site neutrality will mean lower costs for you and I, and lower margins for the hospitals.

Still on the topic of site neutrality, Fresenius reports huge growth in home dialysis. Which is just as well for a couple of reasons: First, Fresenius has to go to home dialysis if it wants a business because Medicare is pushing on that for cost reasons. Second, that’s the main reason it acquired NxStage. Relatively speaking, the US has low rates of home dialysis. Because, it seems, decades ago Medicare policy drove people into dialysis centers. Apart from the cost, I have to believe that in a country like the US with large, sparsely populated rural areas, home dialysis can provide a much better quality of life for many people.

Forest Devices wins a pitch competition for a device to help EMS crews rapidly diagnose a stroke.

Some people are very upset that Google has access to patient data through a partnership through Ascension Health. Google has since clarified things a bit. Me, I’m not that concerned about it. If healthcare is going to become affordable and accessible, we need a revolution, not tinkering. And revolution never comes from within, always from the outside. So I’m all for new approaches that might dramatically improve quality and lower costs.

Humana reports $3.5bn savings from value based care in its Medicare Advantage program.

Docs still don’t like EHRs. Never have, almost certainly never will if it takes away from their patient time. But, hard to deny that having a permanent record that could be shared is better than paper and all its limitations. Nevertheless, it unfortunately seems that EHRs are a good idea badly executed.

But wait, another plus for EHRs….the UKs NHS reports good result with the early detection of in-hospital sepsis. I think this is the full study here, so it looks like an algorithm running in Cerner is the business end of that. We can only speculate how much better the results might be if fed by continuous data from patient monitoring.

More good news on sepsis, University of Colorado has developed a predictive algorithm that runs against the EHR to predict sepsis in children.

Fortnightly Healthtech Update #9

The social determinants of health (SDOH) have rightly become a strong focus in the US in the last couple of years. Insightful then that medical students are poorly trained in nutrition. SDOH is warming up sufficiently that there is an analytics startup focused on just that – called appropriately enough Socially Determined.

The traditional split between providers and payers is fading away, with UnitedHealth buying Vivify for remote patient monitoring. As it’s the Optum division doing the buying, I don’t think this has any role in the payer side of UnitedHealth. But, monitoring patients in the home is a cost containment/reduction play, as providers try to reduce hospital re-admissions. So, most likely RPM is a service that will be offered to providers, and potentially more powerful when integrated with its population health offering.

There’s some rough and tumble in the primary care end of the business right now. There are a number of different business models, each focused on one of the big problems for patients: Timely access to a primary care doc. On one hand, there is the growth of concierge medicine and direct primary care (DPC). It’s hard to put a solid number on that apparently, but it’s there. At the other, budget end of the spectrum, are retail clinics – roughly 6,000 or so of them if Consumer Reports is to be believed. The third approach is telehealth services, such as Teledoc and MDLIVE . And that competitive dynamic has led Walgreen’s to pull back somewhat on in-store clinics. As the article notes, direct-to-consumer (D2C) telehealth has emerged as a serious competitor. And retail clinics have seen explosive growth too, growing 8% in numbers in 2018. Expect a bit of a cull and some M&A in each of concierge medicine, telehealth, and retail clinics before each finds its competitive niche.

This is pretty cool, MedRhythms is getting into trials with hospitals to improve stroke rehab. I first saw this company a couple of years ago when they were part of the Philips Healthworks incubator. Like most moments of pure genius, it’s beguiling in its simplicity, using music to help rewire the brain.

Also a simple idea with potential, a pacifier that analyzes a newborn’s saliva.

I mused last time on the potential liability that will follow the adoption of AI in healthcare. It seems Memorial Sloan Kettering are thinking about it too, but don’t have any more clarity. But then again, AI can help avoid malpractice lawsuits. None of this is going to put the brake on AI funding thought, with Viz.ai scoring $50m from Google Ventures, among others.

Google is lined up to acquire Fitbit for $7.35 a share. That’s a premium on recent trading, but a long way from the circa $50 a share post IPO. Fitbit has been trying to pivot from fitness to healthcare for a couple of years. Makes sense as fitness trackers are a dime a dozen now. Becoming a true medical device though needs a big step up in accuracy, dependability, and reliability. It’s one thing to get pretty good data from most of the people, most of the time. It’s something else entirely to get great data from all of the people, all of the time. Contrast Fitbit’s woes with Apple growing its wearable revenue by 54%. Scroll a little further down in that Apple article though, and you’ll find Withings also trying to expand from consumer devices to professional healthcare. Respectfully, I have to disagree with the CEO of Withings who says “Data is the key to improving nearly every aspect of our healthcare system.” I’ve had this discussion with many people. Data is not the answer. The key to improving healthcare is to pay providers to do what we want them to do – namely deliver better outcomes at lower cost. If we get the financial incentives right, everything else will follow – including the use of insights from data. But unless we change the financial incentives to reward the outcomes we want, all the data in the world isn’t going to change anything.

Many different viewpoints on what the Google/Fitbit deal means btw, including accelerating the pivot to healthcareGoogle can’t do hardware, and it’s all about the data.

A new paper-based approach to detecting sepsis – and I don’t mean paper-based records, I mean a sensor made from paper that won an award.

Medicare refines the reimbursement policy on  remote patient monitoring, actually making it more likely to be adopted. This is goodness, major cost and suffering savings here if care models can fall into line and support RPM for chronic care. And that is, after all, dependent on providers being clear on how to make money from it. Something which has been missing from the reimbursement guidelines thus far. Re earlier point on data only being truly useful when the financial incentives are there.

New to me, KnowFalls has a video monitoring solution as an alternative to sitters to help prevent patient falls. As it acknowledges, it’s not the first company to go down this path, but it claims to be more cost effective. Personally, I’m not sure these solutions will ever be widely adopted. For one thing, patient privacy rights in Europe can make video monitoring difficult. Also, many of the elderly patients being monitored will unfortunately be suffering from dementia. The idea of some bodiless voice talking to patients with cognitive difficulties just feels like it could be a very disturbing experience. 

The CEO of Medically Home makes a financial case for treating chronically ill patients at home. What leapt out at me here is that in the US we spend $800bn annually on brick-and-motor hospitals (I haven’t double checked that, and it’s not sourced). Problem is, while we continue to build rooms for hospital beds, hospital CFOs will want to fill them. Isn’t the accountable care organization (ACO) the best way out of this cycle..? Define outcomes standards for managing, say, heart failure. Then it’s up to the ACO to find the most cost effective way to meet those standards for each patient. With of course downside risk if they go over budget.That is something Medicare is pushing for with site neutral payments, but legal obstacles are rightly or wrongly slowing that down.

Here’s Rock Health’s latest report on digital health consumer adoption, the key highlight for me is how patients are willing to share their health data, but few providers are geared up to work with that.

Interesting move for AliveCor as it tries to fend off Apple, starting a partnership with Huami.

Amazon continues to fill out its employee health offering, with the acquisition of Health Navigator.

To wrap up, interesting insight into the impending crisis that is the declining health of millenials. Which is odd, because the preconception I had is that millennials are super focused on being  healthy. Note, this is US millennials, not sure if it applies to other countries too. Either way, it’s not good reading….

Fortnightly Healthtech Update #8

Potentially very cost effective – using a smart speaker to monitor respiration of sleeping babies. I wonder if this is something that could be adapted for chronically sick Medicare beneficiaries?

Philips joins physIQCurrent Health, and Isansys – among others – in using machine learning for early detection of patient deterioration. And also works with the DoD to detect infection early in active duty personnel.

More AI, improving radiologists ability to spot breast cancer. One of the goals is to reduce the number of biopsies required. Great in theory, probably not in practice. Imagine the scenario: A patient whose cancer goes undetected until late, sues the radiologist. In court,the lawyer asks the radiologist why they didn’t order a biopsy. Answer, “Because the AI told me not to.” Some medical tests get ordered solely to ward off litigation.

And yet more….Medicare thinks AI could cut fraud. I’m convinced it could – but in my experience, it also takes political will, and that’s not always there in abundance….

Concierge medicine taken to the extreme: The $10,000 dollar all-day physical. Which possibly does more harm that good, with the optional full-body CT scan, since the American College of Radiology advises imaging should only be used when there is a clear medical benefit.

A bit late to the party perhaps, but Fitbit is looking to add apps to help users detect afib. Talking of afib, AliveCor goes to the well again for $6m.

There’s a reason for that: Physician-led ACOs performed better than hospital-led ACOs. Hospitals often stand to make more money from readmitting a patient than they stand to lose in readmission penalties. Physicians have no such conflict of interest.

A very niche solution for patient monitoring, targeting airline passengers who are taken ill during flights. Just how niched..? I’m not sure, Google wouldn’t readily give it up.  But I do know this – I’d readily give up the possibility of in-flight vitals monitoring for an extra inch of legroom.

Working as planned then – Medicare reimbursement and incentives starting to see more patients discharged to home care, fewer to skilled nursing.

Brigham and Women’s experimenting with the Fruit Street Health telemedicine platform to try and ward off dementia.

We often hear about the controversially high price of new, innovative meds. But here’s exploitative pricing for something that’s been around for almost 100 years, the price of insulin increases 600% over 17 years.

Remote patient monitoring continues to show success, with St. Luke’s serving 36,000 patients in the first year.

Maybe I missed it before, but Google’s/Alphabet’s Verily is starting to get more active in population health, working with Atrius in California.

Worryingly, I find myself increasingly skeptical about value-based care. Allegedly, a growing percentage of healthcare payments are tied to value-based care. And yet, very few Medicare ACO’s are taking financial risk, they are in upside only models. In other words, if they beat spending targets, they get a bonus. If they overspend, well, we’ll just let it slide…So, we have value-based care, with payments linked to quality targets, but no penalty if you miss the quality targets. So, where’s the value in that..? Maybe Intermountain Healthcare can restore my faith.

And finally, a couple of new biosensors, from Integrated Device Technology (beware press release), and a team at Duke University.

Fortnightly Healthtech Update #7

Current Health (formerly Snap40) partners to add axillary temp and spirometry to the parameters it measures.

Detecting antibiotic levels in real-time.

Bioformis ups its game with FDA clearance for BioVitals to help manage chronic conditions.

MC10 partners with University of Rochester to collect real-world evidence for how people live with Parkinson’s. Reminds me a bit of this collaboration between IBM and Pfizer.

Interesting perspective that I hadn’t considered before: Dated US regulatory policies are endangering the US lead in healthcare technology.

Direct primary care (DPC) might help achieve half of the quadruple aim, clinician satisfaction and patient satisfaction. But lower healthcare costs overall, and higher quality overall, I’m not so sure. It’s also typically positioned as an up-sell to people who already have health insurance. But just thinking out loud, maybe there’s another angle. If you can’t afford a traditional insurance plan, would a low cost primary-care-only option with a strong focus on preventative care be better than nothing…?

Sounds like a lawsuit waiting to happen, Rensselaer County launches online emergency room for Medicaid patients. I’m sure the intent is good, to prevent unnecessary ER visits. But, pitching an app called “ER Anywhere” to a patient population might just be asking for trouble.

I expect to see more of this from large employers tired of rising healthcare premiums: Walmart to use Embold Health’s analytics to herd it’s employees to high quality providers.

Can a simplified cardiac risk score also predict strokes?

In case anyone was still in doubt, new research published in JAMA suggests 25% of all healthcare spending could be waste (free registration required). And that doesn’t include administration. Candidly, a primary care doc once told me the rule of thumb in primary care is 3 admins for every medical practitioner. Their practice at the time had a 5 to 1 ratio! I’m inclined to think that admin overhead is a direct result of the complexity of the payer/reimbursement model. And they say that government managed healthcare would be inefficient…

Best practice for in-hospital rapid response teams (RRT) include having a dedicated RRT, and the ability for anyone to trigger the RRT without fear of reprisal.

The only thing that surprises me about this is that patients still have to pay some of the cost: Devoted health to use Apple Watch with Medicare Advantage patients. I would be shocked if most payers/providers aren’t paying for wearables like this in a few years – and financially encouraging patients to use them. If they’re not, it probably means the push for value-based care has failed.

The ECRI Institute releases its top 10 technology hazards for 2020. Not for the first time, the problem of over-alarming makes the list.

In Pennsylvania a patient tragically expired in the ER after being left unattended. A potential market opportunity for wearables perhaps. But unfortunately, hard to make a business case for it that is going to make a hospital CFO jump for joy.

Funding for digital health startups has cooled a tad, but still an estimated $1.3bn in Q3 according to Rock Health.

Fortnightly healthtech update #6

Firefly raises an additional $10m for concierge primary care, bags new CEO.

An incredibly low-tech way to spot patient deterioration early – listen to their loved ones.

A vision for the Internet of Medical Things (IoMT) in the UK

Apple announces three new health studies – but Jessica Baron at Forbes adds a big dollop of caution.

Corvia Medical and PhysIQ team up for a study in heart failure management.

Henry Ford in Detroit has some success cutting re-admissions with remote patient monitoring. Now it’s looking for partners to help it expand and grow.

Picking up on the thread from last time that value-based care might not be just around the corner: Stakeholders push back on bundled payments for oncology. Also, current value-based care models need greater emphasis on specialty care. Hard to argue with that. It’s true, patents with chronic conditions account for the majority of US healthcare spending. But, if we focused on preventative medicine instead of treating sickness – through a single payer system for example – arguably we wouldn’t need so much specialty care now would we.

Rapid advance in medical wearables runs headfirst into reality: Docs call for standards for blood pressure measurement.

Using a ballistocardiogram to monitor heart failure patients at home. It’s OK, I didn’t know what it was either…

Hospital execs in New Hampshire fret that high deductible health plans might be a barrier to care. Which is really hospital CEO code for “high deductible plans hurt our revenue”. And they’re right – but that’s part of the point. I can tell you from personal experience, a high deductible changes your thinking. If you’re personally paying 100% of the bill for everything It makes you spend only on the “must haves” in healthcare. And that’s OK. If we’re going to rein in healthcare spending – which we must, because it’s not sustainable – we need to give up the “nice to haves”. And providers will feel some of that pain, short term. But there is a perverse incentive with high deductible plans. What happens when you hit the deductible limit? Well, you just pile in and take care of all those “nice to haves” while the insurer is paying for them…

That said, high deductible insurance plans would be more effective if consumers had tools to compare both price and quality when shopping for healthcare. Often, they don’t. But, here’s a study showing that price transparency tools seem to work. Which is interesting, because Walmart just opened it’s first physician-staffed clinic with transparent pricing upfront.

One more reason for healthcare execs to fret: Amazon launches Amazon Careas a pilot for its’ employees. Coming soon to….probably pretty much everyone once they’ve worked out the wrinkles I imagine. It looks like a nice example of focusing on the customer need. Not the more traditional healthcare approach of “What can we do that we’ll get reimbursed for…?”.

And just to pile on the pressure, Sam’s Club (owned by Walmart) is introducing $1 telehealth visits with a subscription. This is interesting too. Telehealth options are usually offered by an employer. This is the first I’m aware of that is D2C. So what’s the target market…? Is it people without health insurance? Or, is it people who have insurance, but want a more cost effective/convenient way to access primary care?

An interesting history lesson: Americans already paid for universal health care. Just not for fellow Americans.

A good summary of how far Medicare has got down the road with value-based care. Not very. Just to reinforce that, although physicians participate in value-based care, most of their revenue come from good ol’ fashioned fee-for-service.

Eko raises $20m for it’s machine learning driven cardiac monitoring.

Also in funding news, InsightRX raises $10m for precision dosing – using data to optimize the therapeutic dose and minimize side effects.

Sad but true: Slowing the growth of healthcare costs – not actually cutting costs – is seen as  a triumph. But at least BCBSMA has achieved that with an 8 year value based program.

A long, but worthy read: How to avoid getting fleeced in the emergency room.