Paying for (Artificial) Intelligence

 The US Government Will Pay Doctors to Use These AI Algorithms

The US Centers for Medicare & Medicaid Services (CMS) recently said it would pay for use of two AI systems: one that can diagnose a complication of diabetes that causes blindness, and another that alerts a specialist when a brain scan suggests a patient has suffered a stroke. The decisions are notable for more than just Medicare and Medicaid patients—they could help drive much wider use of AI in health care.

The incentives for technology adoption are related to professionalism, organizational, and economic. In this case are the latter one.Though, FDA must have cleared both algorithms, and this is the case. 

We could also think on using AI for reducing radiation exposure in CT scans for example..

EU is lagging behind developed countries in this (and many other issues)...

Prioritising population health or the economy (2)

  The Pandemic Information Gap. The Brutal Economics of COVID-19

Joshua Gans has updated his former book on covid economics. And says:

Pandemics are an information problem. Solve the information problem and you can defeat the virus. There is a big difference between knowing someone you interact with is infectious and having to make a guess as to whether that person is infectious. In the former case, you can act and limit the interactions. In the latter case, you have to take a risk. And, in evaluating that risk, what we care about is not just whether you become infected but also whether you might pass that infection on to others.

 The difference between perfect knowledge and no knowledge is what causes an infectious disease to have an impact on social and economic interactions. With perfect knowledge, some people get sick, they are isolated, and life is (for most of us) essentially unchanged. With no knowledge at all and no interventions to prevent infections, then for COVID-19, at its peak, about 21 million people in the United States alone would likely be infectious at one time. With no restrictions on activity, the probability that you interact with one of the infectious people on a given day is 21 million divided by 327 million (the US population), or 6.4 percent.4 However, suppose you interact with only 10 people per week. In that situation, the probability that you are able to avoid any of those infected people is about 50-50. When going to public spaces, you may interact with over a hundred people per week. In that case, your probability of avoiding an infected person becomes close to zero. In other words, perfect knowledge allows you to avoid all infected people. No knowledge makes it near certain that you will encounter at least one infected person.

 Without knowledge of how many people are infected and whether particular people are carriers of thecoronavirus, we are forced to take drastic actions.

True. Pandemics are an information problem but information will never be perfect and complete. Uncertainty sorrounds us. And pandemics are more than an information problem. Because you may know who is infected, and not "act and limit interactions". Therefore, emotions, incentives and expectations count. We do have also a behavioral problem. And if it is behavioral, it has ethical implications. And finally, that's life, decisions with or without information and behavioral and ethical implications of such decisions.

Anyway, a useful introductory text.

Measuring covid-19 related deaths

 Excess mortality. Measuring the direct and indirect impact of COVID-19

Estimations of excess mortality can give an overall understanding of the impact of  COVID-19, by not only comparing deaths that are directly attributable to the virus, but also by 
taking into account indirect mortality. There may be, for example, deaths due to health systems not being able to cope with other conditions or to the longer-term impact of the pandemic on population health. There has been preliminary evidence of disruptions to continuity of care, hindering people suffering from heart attacks, strokes and other conditions from seeking necessary treatment in emergency rooms, or of delays in accessing regular preventive or ongoing care in primary care practitioner (GP), surgeries or outpatient departments for chronic conditions.

And a clear message on this OECD report:

 Comparing the number of reported deaths from all causes against the average over the previous five years, Spain recorded a 61% increase in overall mortality over this 10-week period. At the beginning of April (Week 14), Spain also recorded the highest excess mortality for a single week, with a 154% increase in mortality, equating to more than 12 500 additional deaths compared to the 5-year average for the same week.

The politics of care in US

The Politics of Care. From COVID-19 to Black Lives Matter. Edited by Boston Review

After the US elections, now it's time to rebuild. A new book provides some insights.

WHERE DO WE GO from here? Our answer must connect to a broader politics that addresses the deep structural roots of the problems we face in the United States. We must build for a better future, not just climb out of the rubble of this pandemic, brush ourselves off, and return to business as usual. We need a new politics of care, one organized around a commitment to universal provision for human needs; countervailing power for workers, people of color, and the vulnerable; and a rejection of carceral approaches to social problems. The question now is how to connect that vision to programmatic responses that address the needs of the moment and beyond. We need to aim at “non-reformist reforms”—reforms that embody a vision of the different world we want, and that work from a theory of power-building that recognizes that real change requires changing who has a say in our political process.

The proposals from the progressive wing of the Democratic Party, and even those from democratic socialists, are missing what we might call a New Deal for Public Health. Here, social movements are indispensable. In particular, the AIDS movement of the past forty years offers a template for the kinds of mobilization we’ll need to achieve our goals: not only bringing the virus under control, but also building a future where something like this never happens again.

What might that include? The Medicare for All component has been mapped out, but less obvious and just as crucial is a new, robust, long-lasting infrastructure of care. For example, people need to be able to stay home when they are sick. Yet employment in this country systematically undermines our ability to care for ourselves and others. Women have it the worst, especially immigrant women and women of color, for they are the ones with the highest burdens of paid and unpaid care. The truth is stark: more than 32 million workers lack access to paid sick days. While 93 percent of the highest-wage workers have access to paid sick days, only 30 percent of the lowest-wage workers do. The risk of job loss and precarious scheduling all add to employees’ difficulties. Without anything like universal sick pay or income insurance, self-employed and gig economy workers are cast adrift at a time like this.

Diagnostic Testing for the COVID-19 Pandemic (again)

 Rapid Expert Consultation on Critical Issues in Diagnostic Testing for the COVID-19 Pandemic

Yesterday I was thinking about the implications of having spare structural capacity in hospitals for pandemics and disasters. With a larger capacity, the need for lockdown to preserve health system operations would be less important. However, how much capacity is needed is uncertain. Therefore, the cost and benefit of spare capacity of the health system is very difficult to estimate. And I would add, it is really much more difficult to manage such investment, because spare capacity in physical terms is not enough, you would need also spare capacity in human resources!. Maybe there is a technological innovation that I can foresee. Forget it. 

Right now the hotest issue are the tests and the vaccine. And NASEM has released a rapid consultation of interest about tests.

This rapid expert consultation draws attention to four critical areas in developing diagnostic testing and strategies to reduce the number of COVID-19 infections and deaths: (1) advantages and limitations of reverse transcription polymerase chain reaction (RT-PCR) testing for viral RNA; (2) the status of POC testing; (3) testing strategies, namely, considerations in the deployment of types and sequences of tests; and (4) next-generation testing that offers the prospect of highthroughput, rapid, and less expensive testing.

If you want to know the state of the art, this is the document to read. I can't see anywhere any cost-benefit approach of different options...We are still dealing with precision and accuracy, i.e. effectiveness.

PS. Quite surprising that today everybody is talking about a vaccine and its 90% effectiveness without any scientific paper being published. Can you accept that?. Information merchants looking for attention.

 

Editing human genome

 MODERN PROMETHEUS. Editing the Human Genome with Crispr-Cas9

Drug approval and geographic differences

 Approval of Cancer Drugs With Uncertain Therapeutic Value: A Comparison of Regulatory Decisions in Europe and the United States

We know that the regulation of medical devices is quite different between US and Europe, and with COVID tests we have experienced such divide. In drugs, one could expect a closer approach to approval. However, this is not the case. 

Regulatory agencies may have limited evidence on the clinical benefits and harms of new drugs when deciding whether new therapeutic agents are allowed to enter the market and under which conditions, including whether approval is granted under special regulatory pathways and obligations to address knowledge gaps through postmarketing studies are imposed.

In a matched comparison of marketing applications for cancer drugs of uncertain therapeutic value reviewed by both the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA), we found frequent discordance between the two agencies on regulatory outcomes and the use of special regulatory pathways. Both agencies often granted regular approval, even when the other agency judged there to be substantial uncertainty about drug benefits and risks that needed to be resolved through additional studies in the postmarketing period.

Postmarketing studies imposed by regulators under special approval pathways to address remaining questions of efficacy and safety may not be suited to deliver timely, confirmatory evidence due to shortcomings in study design and delays, raising questions over the suitability of the FDA’s Accelerated Approval and the EMA’s Conditional Marketing Authorization as tools for allowing early market access for cancer drugs while maintaining rigorous regulatory standards.

 Hockney

The long and bumpy road to CRISPR (2)

 Editing Humanity. The CRISPR Revolution and the New Era of Genome Editing

In 2017 I wrote a post about the book by Jennifer Doudna, A Crack in Creation, now Kevin Davies, the editor of the CRISPR journal has published a new book on CRISPR. It is an effort to put all the information and details about CRISPR in one book. Therefore, if you want to now the whole story (or close to) this is the book to read. If you are interested in a general approach, then the Doudna book is better.

It is quite relevant the chapter that explains the role of Francis Mojica in CRISPR (chapter 3), and the chapter 18, on crossing the germline and what happened about the scandal of genome editing by JK.

“When science moves faster than moral understanding,” Harvard philosopher Michael Sandel wrote in 2004, “men and women struggle to articulate their own unease.” The genomic revolution has induced “a kind of moral vertigo.”49 That unease has been triggered numerous times before and after the genetic engineering revolution—the structure of the double helix, the solution of the genetic code, the recombinant DNA revolution, prenatal genetic diagnosis, embryonic stem cells, and the cloning of Dolly. “Test tube baby” was an epithet in many circles but five million IVF babies are an effective riposte to critics of assisted reproductive technology.

With CRISPR, history is repeating itself,

That's it, great book.

 

NGS: from research to clinical use

 Expanding Use of Clinical Genome Sequencing and the Need for More Data on Implementation

During the past 5 years, next-generation sequencing (NGS) has transitioned from research to clinical use.At least 14 countries have created initiatives to sequence large populations (eg, All of Us, Genomics England), and it is projected that more than 60 million people worldwide will have their genome  sequenced by 2025.

If this is so,

 Understanding how NGS is being used and paid for is critical for determining its clinical and economic benefits and addressing current and future challenges to appropriate implementation.

 Without consistent information on clinical utility and how NGS tests are implemented in clinical care, it is not possible to develop an understanding of benefits and harms associated with NGS. It is not always the case that evidence of clinical utility leads to improved outcomes, and evidence about implementation is required to complete the assessment of the effects on population health. Implementation science is intended to support the integration of findings from scientific evidence to uptake in routine clinical care in the ongoing cycle of a learning health care system.

This is clearly a call for action. However, in my country the call is for both developing NGS  (nothing has started) and asessing implementation. Is there anybody in the room?

Hockney

A nation that failed too many of its citizens

THE HEALTH OF A NATION. A President Looks Back on His Toughest Fight. The story behind the Obama Administration’s most enduring—and most contested—legacy: reforming American health care.

This is the case of US and health reform. In the last issue of New Yorker, Barack Obama explains why health reform was so difficult when he proposed Obamacare, in few words:

"one person’s waste and inefficiency was another person’s profit or convenience"

A lot of details inside the article. Useful hints to understand the intricated world of Washington politics. 

"both the politics and the substance of health care were mind numbingly complicated. I was going to have to explain to the American people, including those with high-quality health insurance, why and how reform could work"

 One thing felt certain: a pretty big chunk of the American people, including some of the very folks I was trying to help, didn’t trust a word I said. One night, I watched a news report on a charitable  organization called Remote Area Medical, which provided medical services in temporary pop-up clinics around the country, operating out of trailers parked at fairgrounds and arenas. Almost all the patients in the report were white Southerners from places like Tennessee and Georgia—men and women who had jobs but no employer based insurance or had insurance with deductibles they couldn’t afford. Many had driven hundreds of miles to join crowds of people lined up before dawn to see one of the volunteer doctors, who might pull an infected tooth, diagnose debilitating abdominal pain, or examine a breast lump. The demand was so great that patients who arrived after sunup sometimes got turned away. I found the story both  heartbreaking and maddening, an indictment of a wealthy nation that failed too many of its citizens.

Now that somebody wants to block its application, there is a need to emphasize the great job that was done.

President Obama in the Oval Office as a vote on the Affordable Care Act approached. Proposals for some sort of universal health care in the U.S. stretched back a century but had always been defeated.Photograph by Pete Souza

Uneven human germline genome editing regulation

Human Germline and Heritable Genome Editing: The Global Policy Landscape

Françoise Baylis et al. have made an excellent review of current genome editing regulations around the world. And the conclusion is that there is no agreement on this extremely delicate issue. Somebody (UN) should do something to fix such a mess.

The summary in one figure:

 Some details:

Regarding human germline genome editing (not for reproduction), potentially relevant documents yielded no relevant information for 56 of 96 countries. Among these 56 countries are 22 of the 29 countries that have ratified the Oviedo Convention, which prohibits heritable human genome editing but leaves the status of germline genome editing to be determined by each country. Among the remaining 40 countries with relevant information, 23 (58%) prohibit this research—19 prohibit it outright and four prohibit it with potential exceptions. Eleven countries explicitly permit human germline genome editing; and six countries are indeterminate

 

 

On healthcare and its contribution to decline in mortality

 The (Still) Limited Contribution of Medical Measures to Declines in Mortality

The Questionable Contribution of Medical Measures to the Decline of Mortality in the United States in the Twentieth Century

 “Medical measures appear to have contributed little to the overall decline in mortality in the United States since about 1900.” Readers might assume that this statement is from a recent research article or policy report featuring the social determinants of health. But no, it is from the 1977 seminal Milbank Quarterly article by John and Sonja McKinlay titled “The Questionable Contribution of Medical Measures to the Decline of Mortality in the United States in the Twentieth Century.”

 In the section of Milbank Quarterly Classics, David Kindig explains his favourite topic while reviewing the McKinlays 1977 article:

 “if they [medical measures] were not primarily responsible for it [the decline in mortality], then how is it to be explained?” They did not answer this question themselves, but referred back to McKeown, who had concluded that “the main influences were: (a) rising standards of living, of which the most significant feature was a better diet; (b) improvements in hygiene; and (c) a favorable trend in the relationship between some micro‐organisms and the human host.”2 However, in their conclusion they magnified this point, stating that “profound policy implications follow from either a confirmation or a rejection of the thesis. 

 For many years, I taught a session of my population health course featuring the two contemporary papers that frame what we know today—the first by McGinnis and colleagues, based on CDC surveys, argues that medical care is responsible for about 10% of preventable mortality, and the second an econometric analysis by David Cutler argues that medical care was responsible for 50% improvement in certain causes of mortality over the period of 1960 to 2000. When students are shocked by this range, I remind them that, in a world that still predominantly assumes the pre‐McKinlay reality of medical care being close to fully responsible for preventing or curing disease and death, it is still a profound statement to many that much more than medical care goes into the production of health.

A must read. There are still many unanswered questions.

Hopper

Health reform, a lost chance (once again) (2)

 Consenso por un sistema sanitario del siglo XXI

A new report and new proposals for health reform. The authors start saying that Informe Abril (released 29 years ago. Yes! I was there) was not finally considered as a source for any reform, as we all know. After that, there are 40 reports available, and nothing happened. After three decades, the probability of a health reform, in the wake of current pandemic and political weakness, is still lower in my opinion. Authors highlight the current difficulties of the health system and generic approaches to be considered.

However, health reform requires a consensus on values, and such consensus will not arise from experts, it is a social and political consensus. Expert consensus is a way to reflect concerns about a topic and ways to tackle it, but on policy terms its relevance and impact is minor.

Anyway, it's up to you your final assessment. 

Covid and the Value of Statistical Life

 COVID-19 and Uncertainties in the Value Per Statistical Life

Do the Benefits of COVID‐19 Policies Exceed the Costs? Exploring Uncertainties in the Age–VSL Relationship

For an individual, VSL can be derived by dividing WTP by the risk reduction. A population-average VSL of $10 million indicates that the typical individual is willing to pay $1,000 to decrease the chance of dying in a given year by 1 in 10,000. Individual WTP also can be summed across individuals expected to accrue the risk reduction. If 10,000 people will experience a 1 in 10,000 risk reduction and are each willing to pay $1,000 for the risk change, the total value is $10 million (10,000 times $1,000), and one less person would be expected to die that year as calculated by 10,000 times 1/10,000.

But

Individual WTP, then, is the fundamental measure—the $1,000 in this case. The conversion to a $10 million VSL is simply for convenience. 

 In a recent study with Ryan Sullivan and Jason F. Shogren, I compare the effects of three approaches often used to adjust for age: an invariant population-average VSL; a constant value per statistical life-year (VSLY); and a VSL that follows an inverse-U pattern, peaking in middle age. We find that when applied to the U.S. age distribution of COVID-19 deaths, these approaches result in average VSL estimates of $10.6 million, $4.5 million, and $8.5 million. The differences in these values is substantial enough to alter the conclusions of frequently cited analyses of social distancing.

 Table II. VSL by Age Group (in 2019 millions of dollars)

Age Group	Invariant VSL	Constant VSLY	Inverse‐U Relationship
Under 1 year	$10.63	$13.88	$5.38
1–4 years	$10.63	$13.74	$5.38
5–14 years	$10.63	$13.37	$5.38
15–24 years	$10.63	$12.64	$5.38
25–34 years	$10.63	$11.76	$8.50
35–44 years	$10.63	$10.63	$10.63
45–54 years	$10.63	$9.19	$10.72
55–64 years	$10.63	$7.54	$8.15
65–74 years	$10.63	$5.68	$8.15
75–84 years	$10.63	$3.72	$8.15
85 years and over	$10.63	$2.03	$8.15

 Table III. COVID‐19 Age‐Weighted Value (in 2019 millions of dollars)

	Invariant VSL	Constant VSLY	Inverse‐U Relationship
Total value, all COVID‐19 deaths	$937.6 billion	$394.8 billion	$773.4 billion
Average VSL, weighted by COVID‐19 deaths by age	$10.63 million	$4.47 million	$8.31 million

Table IV. Effect of Alternative Approaches on Analytic Results

		————————————————‐Benefits————————————‐————
		Lives	Original	Invariant	Constant	Inverse‐U
	Costs	Saved	Approach	VSL	VSLY	Relationship
Thunström et al. (2020)	$7.2 trillion	1.24 million	$12.4 trillion	$13.16 trillion	$5.54 trillion	$10.30 trillion
Greenstone and Nigam (2020)	N/A	1.76 million	$7.94 trillion	$18.72 trillion	$7.88 trillion	$14.64 trillion
Acemoglu et al. (2020)	$2.15 trillion	8.7 million	N/A	$92.44 trillion	$38.93 trillion	$72.31 trillion

Does this make sense? It seems quite high. If we value identified life more than a statistical life, can you imagine the final figure?

Whether the social distancing policy considered by Thunström et al. (2020) yields net benefits varies depending on the valuation approach. The authors use an invariant VSL but apply a somewhat lower value than we use in our analysis ($10 million rather than $10.63 million). However, both our invariant VSL and inverse‐U approaches lead to positive net benefits. Under our invariant approach, the benefits increase by almost $800 billion due to differences between the VSL estimates. Benefits decrease when using the inverse‐U approach, but not by a large enough amount to drop below estimated costs. Under the constant VSLY approach, benefits decrease by a substantial amount and the policy no longer appears cost‐beneficial.

While Greenstone and Nigam (2020) do not include a cost estimate in their calculations, the effects of our three approaches on their featured benefit estimates are significant. The benefit estimates more than double when applying the invariant VSL approach rather than their age‐adjusted approach. Interestingly, their estimates are very similar ($7.94 vs. $7.88 trillion) to the results using our constant VSLY method, while applying our inverse‐U estimates almost doubles the value in comparison to their inverse‐U approach. This result reflects the relative steepness of their curve at older ages as well as our assumption that values level off at older ages under the inverse‐U approach. As noted earlier, the additional sensitivity analyses reported by the authors also show siginificant variation in the results.

Acemoglu et al. (2020) have by far the largest estimates of lives saved across the three social distancing studies, which naturally increases the benefit values. Under all three approaches, we find that benefits exceed costs by an order of magnitude. However, Acemoglu et al. (2020) find that approaches other than the scenario reflected in Table IV are more cost‐effective, particularly if they target higher risk, older age groups.

Anyway, let's search a little bit more on that. Let's take the press. I don't think that this helps to take policy decisions in the current pandemic. It's just recreational research.

PS. The price of freedom is 103€ per day in my country (cost of non being free by mistake). Explained here.

Hockney