Wednesday, January 29, 2014

Someday putting doctors out of business

In the KGI business class today, a student idly asked “when will computer replacing doctors in diagnosing patients?” Another student said “never”.

It's pretty clear that technology will reduce the labor-intensity of medical care, and also shift some tasks from expensive high-skill people to inexpensive low-skill people. Look at how cars were made by Gottlieb Daimler, Henry Ford, Toyota in the 1970s and then today.

Computers will over the next 20-40 years replace some or all of the role of doctors in diagnosing conditions. It’s too labor intensive and expensive not to become a target. The question is not if, but when, where first and how fast.

The claim will be that it's intended to improve consistency and accuracy, but the real reason will be cost. Claimed improvements in quality — such as for patients who lack access to a specialist for diagnosis — could be used to overcome the opposition of highly educated, compensated and organized physicians.

The initial push thus will come from an organization that both has scale economies and a record of innovating to save pennies. My prediction is that the first major use in North America will fall in one of three categories:
  • US government, probably the Veterans Health Administration
  • An HMO, almost certainly Kaiser Permanente; or
  • A provider serving rural areas, most likely the First Nations and Inuit Health Branch of Health Canada.

Sunday, November 10, 2013

Biotech not like other high tech

In teaching, research and talking to industry professionals, I am often tempted to refer to “high technology” industries, “technology startups” and the like. This would tend to emphasize the commonality between IT and biotech.

And then there are days like Friday, when I’m reminded that biotech — and human health more generally — is completely different.

The occasion was an event on pharmaceutical quality, organized by KGI’s student chapter of the Parenteral Drug Association, a professional organization concerned with drug quality and safety issues.

The students invited three industry speakers.

First up was Susan Weber of Baxter introduced us to the principles of Quality by Design, i.e. start from a quality goal and work back through the entire design, development an production process. The ideas are more than 20 years old, but apparently have recently have begun to influence pharmaceutical manufacturing in the US.

The second speaker was Marsha Hardiman, a consultant for Concordia Valsource. After showing a stunning video by the American Society for Quality on the consequences of quality failures, she summarized the regulatory and process failures of the New England Compounding Center that have led to 64 deaths so far. Nothing illustrates the difference between a bad drug and a bad iPhone app.

The final speaker was James Sesic of Amgen, talking about the challenges of maintaining regulatory compliance for drugs sold in more than 100 countries.

This was the real eye-opener. We all know about the need for drug companies to spend years and hundreds of millions of dollars to get the first NDA or BLA approval. Sometimes we talk about getting the second approval — e.g. in Europe or Japan after the US. But I’ve never heard anyone talk about the rest of the world.

How does a company like Amgen handle approval in dozens of countries? The richest countries have their own large-scale regulatory systems (US, Japan, Canada, Europe), the smallest grant approval after certified approval from one of the major regulators, while a range of countries attempt to form their own regulatory judgements without a lot of resources.

On top of that, regulatory approval is required for any major change in the production process. Normally this discourages companies from making major changes, but if there’s a major improvement in the process — or the company needs to comply with new regulations — it will go through the process.

One example is getting approval to shift manufacturing to a new factory. The company will have to apply for approval in dozens of countries and cannot sell drugs in country X from the new factory until regulatory agency X has approved such production.

If a drug has several deliver modalities — concentration, IV vs. injection, etc. — then when multiplied by the disparate languages, marking requirements and other national regulations, a single blockbuster drug could be sold in 100s of SKUs. Double that with separate SKUs from the old and the new factory.

When it takes 4-6 years for all the countries to approve the change, then an Amgen needs to keep track of all those 200? 500? SKUs (for one drug) to know which SKU is legal to sell in one country.

Contrast this to the rollout of the latest iPhone, a product that (unlike software or PCs) must satisfy strict government and operator requirements to be sold in a given country. Apple launched the iPhone 5c in 10 countries in September, added 60 countries between October 25-November 1, and expects to have more than 100 countries by the end of the year (i.e. in less than 4 months).

The process of global drug regulation seems pretty inefficient, and we pay for this inefficiency through higher costs (or lack of access by smaller countries to non-blockbuster drugs). It would be nice if we could develop a drug regulatory system where the first review is highly rigorous but the remaining process is streamlined so that drug companies spend their money on development (and safety), not SAP and paperwork.

Friday, November 1, 2013

Administration: 93 million will lose existing health insurance

From Forbes, October 31:
Obama Officials In 2010: 93 Million Americans Will Be Unable To Keep Their Health Plans Under Obamacare
[by] Avik Roy

It turns out that in an obscure report buried in a June 2010 edition of the Federal Register, administration officials predicted massive disruption of the private insurance market.

Section 1251 of the Affordable Care Act contains what’s called a “grandfather” provision that, in theory, allows people to keep their existing plans if they like them. But subsequent regulations from the Obama administration interpreted that provision so narrowly as to prevent most plans from gaining this protection.

“The Departments’ mid-range estimate is that 66 percent of small employer plans and 45 percent of large employer plans will relinquish their grandfather status by the end of 2013,” wrote the administration on page 34,552 of the Register.

The Departments’ mid-range estimate is that 66 percent of small employer plans and 45 percent of large employer plans will relinquish their grandfather status by the end of 2013,” wrote the administration on page 34,552 of the Register. …

Another 25 million people, according to the CBO, have “non group and other” forms of insurance; that is to say, they participate in the market for individually-purchased insurance. In this market, the administration projected that “40 to 67 percent” of individually-purchased plans would lose their Obamacare-sanctioned “grandfather status” and become illegal, solely due to the fact that there is a high turnover of participants and insurance arrangements in this market. (Plans purchased after March 23, 2010 do not benefit from the “grandfather” clause.) The real turnover rate would be higher, because plans can lose their grandfather status for a number of other reasons.

As to the number of people facing cancellations, 51 percent of the employer-based market plus 53.5 percent of the non-group market (the middle of the administration’s range) amounts to 93 million Americans.

Wednesday, June 19, 2013

The Myriad decision

The Supreme Court on June 13 issued its decision on Association for Molecular Pathology v. Myriad Genetics. The case was of great interest because we covered it in class this semester. To understand the case, I read more than a dozen articles or commentaries on the decision: in addition to the decision itself, the most useful were a same day report by Jason Rantanen of the PatentlyO blog and this week’s analysis by John Conley of the Genomics Law Report.

The case centered on Myriad’s discovery of two genes (BRCA1 and BRCA2) linked to a greater inherited risk of breast cancer. In an opinion written by Clarence Thomas, the Supreme Court held unanimously that Myriad was not allowed to patent isolated DNA but was allowed to patent composite DNA (cDNA).

As the slip opinion reported:
At issue are claims 1, 2, 5, 6, and 7 of U. S. Patent 5,747,282 (the ’282 patent), claim 1 of U. S. Patent 5,693,473 (the ’473 patent), and claims 1, 6, and 7 of U. S. Patent 5,837,492 (the ’492 patent).
and specifically whether these claims are patentable under 35 U. S. C. §101::
§ 101 - Inventions Patentable:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
The difficulty of the case (as with any patent case the Supreme Court is willing to accept) is trading off two conflict policy goals. As Thomas wrote:
patent protection strikes a delicate balance between creating “incentives that lead to creation, invention, and discovery” and “imped[ing] the flow of information that might permit, indeed spur, invention.”
From the opinion, it’s clear that this case sets a key precedent for DNA patenting, building upon the seminal Diamond v. Chakrabarty (1980) and last year’s Mayo v. Prometheus.

Diamond v. Chakrabarty was the first case to allow DNA patenting, in this case for a new organism created by GE microbiologist who created a new oil-degrading microbe. In retrospect, it doesn’t seem controversial at all, but the case was narrowly decided by a 5-4 margin, with the two conservative judges joined by three swing justices (vs. the four liberal justices).

Last year, in the Prometheus case, a unanimous SCOTUS held that “Laws of nature, natural phenomena, and abstract ideas are not patentable.”.

Despite efforts by Myriad to argue otherwise, in this case the court agreed with plaintiffs that isolating DNA corresponded to “laws of nature”. In this regard, they seemed heavily influenced by the opinion of William Bryson, the dissenting justice in a 2-1 decision favoring Myriad at the U.S. Court of Appeals for the Federal Circuit.

However, all three appeals justices (and all seven SCOTUS justices) agreed with Myriad’s argument that the cDNA is manmade and thus entitled to patent protection. As Thomas wrote:
the lab technician unquestionably creates something new when cDNA is made. cDNA retains the naturally occurring eons of DNA, but it is distinct from the DNA from which it was derived. As a result, cDNA is not a “product of nature” and is patent eligible under §101…
What was particularly instructive was the final (III) section of Thomas’ opinion, which begins: “It is important to note what is not implicated by this decision.” If DNA patents are invalid, two related patent areas are not.

Myriad did not assert any method claims: if it had, it appears the court would have upheld them. Similarly, the case did not consider any applications (of the unique DNA knowledge Myriad developed). As appeals justice Bryson noted, Myriad both “was in an excellent position to claim applications of that knowledge. Many of its unchallenged claims are limited to such applications.”

What is unresolved is the business impact of this decision. On the one hand, Myriad’s lead attorney told Bloomberg “We have a very strong patent estate around the BRCA test,” protected by 24 patents. However, as Bloomberg (and others) noted, the plaintiffs and many other private and university entities will be offering BRCA-based tests. Presumably these competing approaches will test the validity (or effectiveness) of Myriad’s patent barriers.

A second key issue is how the decision will impact university research, Rantanen (writing at PatentlyO) suggests two possible outcomes:
In terms of the effects on my friends here at the University, I can see at least two consequences. First, it may allow researchers more freedom to engage in whole-genome sequencing because they won't need to deal with a multitude of isolated DNA patents for individual sequences. On the other hand, because early-stage research on newly discovered DNA sequences cannot be patented, it may encourage companies - and perhaps universities - to pursue greater secrecy over those early stage discoveries. Social research norms may cut against this - particularly in universities - but there may be some increased pressure, particularly at the margins, towards secrecy of potentially valuable inventions.
Finally, there is some question as to how long the cDNA distinction will last. Forbes quoted biochemist and IP attorney Brenda Jarrell as disputing the court’s finding that “cDNA is not a product of nature.” When this argument is made in a future case, it’s possible that the justices will reconsider where they have drawn the line between nature and invention.

That points to the final (and best) part of the entire Slip opinion, the concurrence by Antonin Scalia:
I join the judgment of the Court, and all of its opinion except Part I–A and some portions of the rest of the opinion going into fine details of molecular biology. I am unable to affirm those details on my own knowledge or even my own belief.
Although Scalia relied on the same distinction between natural and manmade used by the other eight justices, he did not feel qualified (nor necessary) to deliver a two page tutorial on molecular biology. Some would count this as an all-too-rare example of humility by any member of the high court.

Friday, May 24, 2013

Spurring innovation for a system that doesn't want it

This morning I watched the #bigdatamed conference at Stanford via webcast. (I was out of town earlier in the week, presenting my own talk, and couldn’t attend the conference in person.)

The session I watched was entitled “Big Data Opportunities for Healthcare Startups,” part of a larger conference on “Big Data in Biomedicine: Driving Innovation for a Healthier World.” The panel consisted of Ori Geva (Medial Research), Vicki Seyfert-Margolis (MyOwnMed), Risa Stack (GE Healthmagination), Warren Hogarth (Sequoia Capital) and moderator Chris Longhurst (Lucile Packard Children’s Hospital).

It’s no revelation to say that healthcare startups face huge regulatory barriers that drive up both costs and capital requirements. It’s hard to imagine — other than perhaps a defense company — a more regulated industry the USA. (And, as I showed in a 2008 study, aerospace communications allowed market entry by some underfunded entrepreneurs who later gave us Qualcomm).

It’s also no revelation that most stakeholders — patients, doctors, taxpayers, the government, insurance companies — would like better outcomes at a lower cost, but have different perspectives on how to do so.

Notably, former Kleiner Perkins exec Risa Stack cited last week’s published lament by healthcare IT entrepreneur Jonathan Roth about the difficult of bringing innovation into the healthcare system. She cited his three barriers:
1. Healthcare consumers don’t shop. … With hardly an exception, patients can’t shop for healthcare. Why compare prices if you have no choice? Caregivers can’t shop for healthcare either. Few doctors know the true cost of tests and treatments they refer patients to. The setup leaves caregivers with little incentive to differentiate themselves by embracing innovative care delivery practices.

2. The biggest buyer stifles innovation. The government is the biggest buyer in the healthcare market. In fact, it represents more than half of all healthcare buying in the U.S. Unfortunately, the big spender has not prioritized supporting innovation, but rather minimizing the scenario of maximum regret: the audit, the lawsuit, the death. This compliance burden breeds risk-averse behaviors among healthcare providers, not creativity. Furthermore, the government rewards caregivers for a narrowly defined set of activities, limiting their appetite for innovation and thus for entrepreneurial services and technology.

3. Service, quality, and competitive pricing aren’t rewarded. By and large, doctors continue to be paid based on how many services and tests they provide. Here and there, they are also paid for reporting some data. They are not commonly incentivized to compete on price and quality, and there’s no obvious way to display that information for shoppers.
While Stack thought high-deductible plans might make patient more cost-sensitive, she offered no hope on the other two.

In her talk, former FDA official (turned healthcare social media entrepreneur), Vicky Seyfert-Margolis, noted the pressures coming to the system for greater real world effectiveness. For example, right now oncology drugs are tested for late stage cancers that have failed chemo (with benefits measured in terms of 3 month survival) rather than earlier in the process. Since her husband helped get Obama re-elected, I was not surprised at her optimism about the effect of the Affordable Care Act and Accountable Care Organizations. However, Stack noted that ACOs would tend to increase patient switching costs and decrease choice (which by the way would exacerbate Roth’s concerns).

Still, there was one glimmer of hope. Given HIPAA, the one party that could actually do something to enable patient data availability for better outcomes is the patient him/herself. There are important technical and economic barriers to overcome, but also potential payoffs.

As VC Warren Hogarth noted, this would be an improvement over the current system, where patients (for example) give their tissue to a hospital for (say) a cancer test and then the hospital controls who and how that sample is used. The technical problem, noted Ori Geva, is how will the data be made available in a portable way so that the patient can share it on other platforms.

Several panelists noted the opportunity (and challenge) was what’s in it for the patient. On the one hand, patients [likely an unusually motivated subpopulation] opt in to sharing their data on Patients Like Me. Many patients also want to be treated as co-managers of their own data or (as Seyfert-Margolis noted), the household “chief medical officer” (i.e. mom) does so on behalf of the family.

One challenge is that (as noted) patients don't shop. Another is the (realistic) fear that medical groups, insurance companies or HMOs will use the data to reduce choices for patients in the name of cost containment. (I have no problem with Kaiser choosing cancer treatment A over cancer treatment B based on my genotype, but I start to get pissed when they tell me I can’t have a diagnostic because I’m “low risk”).

On the one hand, the normal way of getting something through the system is to appeal to payers. If entrepreneurs can find a way to directly appeal to patients — as have 23 & Me, Patients Like Me and the various quantified self efforts — perhaps this will create other opportunities for innovation that don’t require winning the end-to-end cooperation of the entire system.

Friday, April 26, 2013

Saving animals, dooming people

Earlier this month, neurobiologists in Milan had their lab broken into in a deliberate attempt to slow their research.

As the Wall Street Journal reported:
The lab was targeted because its work, like just about every other medical advance and effort of our time, involves mice. Hundreds of small, cute, furry mice, which in this case had been genetically modified for protein mutations meant to model, as [Dr. Michela] Matteoli puts it, "what goes wrong in the synapse."

The animal-rights crowd decided it had better plans for the mice. So on Saturday five members of Italy's "Stop Green Hill" group (initially formed to protest a nearby dog-breeding facility) broke into the Milanese lab and "occupied" the area housing 800 animals, mostly mice but also some rabbits. They chained themselves by their necks to the facility's doors, Ms. Matteoli says, ensuring that any attempt at forced entry by the police could "harm them really seriously. They could kill them, break their necks."
The researchers lost mice, and others were scrambled such that it was impossible to associate a specific animal with a given experiment. The damage is estimated to cost a year of work and hundreds of thousands of euros. The research is intended to develop drugs “that might arrest neuron destruction in Alzheimer's patients, or avoid the synaptic dysfunctions apparent in autism.”

Columnist Anne Jolis concluded:
[C]onsidering the zealousness of the people who have a problem with their work, Ms. Matteoli and her colleagues may hit on the cure for Alzheimer's before they convince any "Stop Green Hill" types that studying the caged mice and bunnies is worth it. Even Ms. Matteoli, polite to a fault, admits that she has been "impressed" with some of the "really completely crazy" online comments about news articles on the lab attack.

"One guy wrote that you should only study mice if you're developing medicine for mice. I mean, this is the level of—it's very hard to understand." Here's hoping she doesn't try too hard—and can get back to her regular work soon.
I realize I’m a species bigot, but I never got why people valued primitive animals over human life. (Dolphins and chimps I understand — at least up to a point.) The advances will come too late to help me, but perhaps they could help my in-laws (in another decade) or people of my generation (in 30-40 years).

The biopharma industry will continue to have these sorts of problems unless it can find a way to persuade the public of the necessity of their research and research methods. The anti-capitalist, anti-science challenge seems greatest in Europe, but the U.S. is far from exempt.

Monday, April 15, 2013

Myriad: does it matter?

Today marks the Supreme Court hearings in the Myriad patent case. It is no exaggeration to say this is the most closely watched life science patent case in more than a decade.

In one sense, the case has the potential to be as significant as Diamond v. Chakrabarty. That is the 1980 Supreme Court decision that (to quote a 2005 review) “held that a live, man-made microorganism is patentable subject matter” under the patent code.

There is more than a little posturing on both sides. The patients’ rights, open IP crowd is promulgating op-eds in the Washington Post and Los Angeles Times asserting that no company “should own our DNA”. Not surprisingly, the news pages of the New York Times are promoting sympathetic stories on the anti-IP crowd.

In response, the CEO of Myriad Genetics submitted an op-ed to Forbes and letters hoping to rebut these op-eds. The argument is summed up by the title of the Forbes piece: “We're Not Patenting Your Genes, But Our Research.”

We used the HBS case on Myriad in my Innovation Management class at KGI this semester. The students really enjoyed it. It provided very interesting discussion questions at the intersection of genomic medicine, patent law and public policy. It didn’t hurt that this is a high-profile unsettled question of the law.

For the legal issues, a great place to start is the PatentlyO blog. For example, a December 3 article listed the conventional wisdom among legal scholars. A February 11 article lists the questions to be addressed by the two parties:
  1. Are human genes patentable?
  2. Did the court of appeals err in upholding a method claim by Myriad that is irreconcilable with this Court's ruling in Mayo Collaborative Servs. v. Prometheus Labs., Inc., 132 S. Ct. 1289 (2012)?
  3. Did the court of appeals err in adopting a new and inflexible rule, contrary to normal standing rules and this Court's decision in MedImmune, Inc. v. Genentech, Inc., 549 U.S. 118 (2007), that petitioners who have been indisputably deterred by Myriad's "active enforcement" of its patent rights nonetheless lack standing to challenge those patents absent evidence that they have been personally threatened with an infringement action?
The arguments on each side were highlighted in an August 17 summary of the appeal at the Genomics Lab Report, as well as a blog posting the same day on Derek Lowe’s “In the Pipeline” blog.

However, probably the most interesting article was in my Sunday morning paper here in San Diego, by Bradley Fikes, the longtime biotech reporter who’s one of the few North County Times reporters after it was acquired last year by the San Diego Union. His lead on the Union story captures it all:
No matter which way the Supreme Court rules on the Myriad Genetics BRCA breast cancer gene test patenting case, the importance of such patents is diminishing over time. Biotechnology is moving beyond patents derived from naturally occurring gene sequences. The most important biotech patents nowadays are becoming synthetic gene and RNA sequences, proteins and other indisputable contrivances of human ingenuity.
The article highlights the important pioneering efforts of Craig Venter (now in San Diego) who competed with the Human Gene Project.

In the long run, will Myriad matter? The experts consulted by Fikes think not, for two reasons. First, the relevant patents are expiring and thus at some point won’t be enforced anyway. Second, the IP strategies of biotech companies are getting more sophisticated, emphasizing synthetic organisms rather than isolated DNA strains from natural organisms. Meanwhile, as the New York Times reports, Myriad’s latest business model is using trade secrets for its database.

So in the end, this may be tempest in a teapot. In a decade or two, the industry will have moved on, and the patent lawyers will need something else to debate.