Aaron E. Carroll

In a paper published early this year in Nature Human Behavior, scientists asked 500 Americans what they thought about foods that contained genetically modified organisms.

The vast majority, more than 90%, opposed their use. This belief is in conflict with the consensus of scientists. Almost 90% of them believe GMOs are safe — and can be of great benefit.

The second finding of the study was more eye-opening. Those who were most opposed to genetically modified foods believed they were the most knowledgeable about this issue, yet scored the lowest on actual tests of scientific knowledge.

In other words, those with the least understanding of science had the most science-opposed views but thought they knew the most. Lest anyone think this is only an American phenomenon, the study was also conducted in France and Germany, with similar results.

If you don’t like this example — the point made here is unlikely to change people’s minds and will probably enrage some readers — that’s OK because there are more where that came from.

A small percentage of the public believes that vaccines are truly dangerous. People who hold this view — which is incorrect — also believe that they know more than experts about this topic.

Many Americans take supplements, but the reasons are varied and are not linked to any hard evidence. Most of them say they are unaffected by claims from experts contradicting the claims of manufacturers. Only a quarter said they would stop using supplements if experts said they were ineffective. They must think they know better.

Part of this cognitive bias is related to the Dunning-Kruger effect, named for the two psychologists who wrote a seminal paper in 1999 entitled “Unskilled and Unaware of It.”

David Dunning and Justin Kruger discussed the many reasons people who are the most incompetent (their word) seem to believe they know much more than they do. A lack of knowledge leaves some without the contextual information necessary to recognize mistakes, they wrote, and their “incompetence robs them of the ability to realize it.”

This helps explain in part why efforts to educate the public often fail. In 2003, researchers examined how communication strategies on GMOs — intended to help the public see that their beliefs did not align with experts — wound up backfiring. All the efforts, in the end, made consumers less likely to choose GMO foods.

Brendan Nyhan, a Dartmouth professor and contributor to The Upshot, has been a co-author on a number of papers with similar findings. In a 2013 study in Medical Care, he helped show that attempting to provide corrective information to voters about death panels wound up increasing their belief in them among politically knowledgeable supporters of Sarah Palin.

In a 2014 study in Pediatrics, he helped show that a variety of interventions intended to convince parents that vaccines didn’t cause autism led to even fewer concerned parents saying they would vaccinate their children. A 2015 study published in Vaccine showed that giving corrective information about the flu vaccine led patients most concerned about side effects to be less likely to get the vaccine.

A great deal of science communication still relies on the “knowledge deficit model,” an idea that the lack of support for good policies, and good science, merely reflects a lack of scientific information.

But experts have been giving information about things like the overuse of low-value care for years, to little effect. A recent study looked at how doctors behaved when they were also patients. They were just as likely to engage in the use of low-value medical care, and just as unlikely to stick to their chronic disease medication regimens, as the general public.

In 2016, a number of researchers argued in an essay that those in the sciences needed to realize that the public may not process information in the same way they do. Scientists need to be formally trained in communication skills, they said, and they also need to realize that the knowledge deficit model makes for easy policy but not necessarily good results.

It seems important to engage the public more, and earn their trust through continued, more personal interaction, using many different platforms and technologies. Dropping knowledge from on high — which is still the modus operandi for most scientists — doesn’t work.

When areas of science are contentious, it’s clear that “data” aren’t enough. Bombarding people with more information about studies isn’t helping. How the information contained in them is disseminated and discussed may be much more important.

This article originally appeared in <a href="https://www.nytimes.com/2019/07/22/upshot/health-facts-importance-persuasion.html" id="a8dafbc2-2c2d-3a1c-a907-fb39f39a9987">The New York Times</a>.

(The Upshot: The New Health Care)

Health Facts Aren't Enough. Should Persuasion Become a Priority?

SINGAPORE — Singapore’s health care system is sometimes held up as an example of excellence and as a possible model for what could come next in the United States.

When we published the results of an Upshot tournament on which country had the world’s best health system, Singapore was eliminated in the first round, largely because most of the experts had a hard time believing much of what the nation seems to achieve.

It does achieve a lot.Americans have spent the last decade arguing loudly about whether and how to provide insurance to a relatively small percentage of people who don’t have it. Singapore is way past that. It’s perfecting how to deliver care to people, focusing on quality, efficiency and cost.

Americans may be able to learn a thing or two from Singaporeans, as I discovered in a recent visit to study the health system,although there are also reasons that comparisons between the nations aren’t apt.

A population that is healthier

Singapore is an island city-state of around 5.8 million. At 279 square miles, it’s smaller than Indianapolis, the city where I live, and is without rural or remote areas. Everyone lives close to doctors and hospitals.

Another big difference between Singapore and the United States lies in social determinants of health. Citizens there have much less poverty than one might see in other developed countries.

The tax system is progressive. The bottom 20 percent of Singaporeans in income pay less than 10 percent of all taxes and receive more than a quarter of all benefits. The richest 20 percent pay more than half of all taxes and receive only 12 percent of the benefits.

Everyone lives in comparable school systems and the government heavily subsidizes housing. Rates of smoking, alcoholism and drug abuse are relatively low. So are rates of obesity.

All of this predisposes the country to better health and accompanying lower health spending. Achieving comparable goals in the United States would probably require large investments in social programs and there doesn’t appear to be much of an appetite for that.

There’s also a big caveat to Singapore’s success. It has a significant and officially recognized guest worker program of noncitizens. About 1.4 million foreigners work in Singapore, most in low-skilled, low-paying jobs. Such jobs come with some protections and are often better than what might be available in workers’ home countries, but these workers are also vulnerable to abuse.

Guest workers are not eligible for the same benefits (including access to the public health system beyond emergency services) that citizens or permanent residents are, and they aren’t counted in any metrics of success or health. Clearly this saves money and also clouds the ability to use data to evaluate outcomes.

Things to like, for the left and right

The government’s health care philosophy is laid out clearly in five objectives.

In the United States, conservatives may be pleased that one objective stresses personal responsibility and cautions against reliance on either welfare or medical insurance. Another notes the importance of the private market and competition to improve services and increase efficiency.

Liberal-leaning Americans might be impressed that one objective is universal basic care and that another goal is cost containment by the government, especially when the market fails to keep costs low enough.

Singapore appreciates the relative strengths and limits of the public and private sectors in health. Often in the United States, we think that one or the other can do it all. That’s not necessarily the case.

Dr. Jeremy Lim, a partner in Oliver Wyman’s Asia health care consulting practice based in Singapore and the author of one of the seminal books on its health care system, said, “Singaporeans recognize that resources are finite and that not every medicine or device can be funded out of the public purse.”

He added that a high trust in the government “enables acceptance that the government has worked the sums and determined that some medicines and devices are not cost-effective and hence not available to citizens at subsidized prices.”

In the end, the government holds the cards. It decides where and when the private sector can operate. In the United States, the opposite often seems true. The private sector is the default system and the public sector comes into play only when the private sector doesn’t want to.

In Singapore, the government strictly regulates what technology is available in the country and where. It makes decisions as to what drugs and devices are covered in public facilities. It sets the prices and determines what subsidies are available.

“There is careful scrutiny of the ‘latest and greatest’ technologies and a healthy skepticism of manufacturer claims,” Lim said. “It may be at the forefront of medical science in many areas, but the diffusion of the advancements to the entire population may take a while.”

Government control also applies to public health initiatives. Officials began to worry about diabetes, so they acted. School lunches have been improved. Regulations have been passed to make meals on government properties and at government events healthier.

In the United States, the American Academy of Pediatrics and the American Heart Association recently called on policymakers to impose taxes and advertising limits on the soda industry. But that’s merely guidance. There’s no power behind it.

In Singapore, campaigns have encouraged drinking water, and healthier food choice labels have been mandated. The country, with control over its food importation, even got beverage manufacturers to agree to reduce sugar content in drinks to a maximum of 12 percent by 2020.

Should beverage companies fail to comply, officials might not just tax the drinks — they could ban them.

What’s really special is the delivery

Singapore gets a lot of attention because of the way it pays for its health care system. What’s less noticed is its delivery system.

Primary care, which is mostly at low cost, is provided mostly by the private sector. About 80 percent of Singaporeans get such care from about 1,700 general practitioners. The rest use a system of 18 polyclinics run by the government.

As care becomes more complicated — and therefore more expensive — more people turn to the polyclinics. About 45 percent of those who have chronic conditions use polyclinics, for example.

The polyclinics are a marvel of efficiency. They have been designed to process as many patients as quickly as possible. The government encourages citizens to use their online app to schedule appointments, see wait times and pay their bills.

Even so, a major complaint is the wait time. Doctors carry a heavy workload, seeing upward of 60 patients a day. There’s also a lack of continuity. Patients at polyclinics don’t get to choose their physicians. They see whoever is working that day.

Care is cheap, however. A visit for a citizen costs 8 Singapore dollars for the clinic fees, a little under $6 U.S. Seeing a private physician can cost three times as much (still cheap in American terms).

For hospitalizations, the public vs. private share is flipped. Only about 20 percent of people choose a private hospital for care. The other 80 percent choose to use public hospitals, which are — again — heavily subsidized. People can choose levels of service there (from A to C, as described in an earlier Upshot article), and most choose a “B” level.

About half of all care provided in private hospitals is to noncitizens of Singapore. Even for citizens who choose private hospitals, as care gets more expensive, they move to the public system when they can.

So Singapore isn’t really a more “private” system. It’s just privately funded. In effect, it’s the opposite of what we have in the United States. We have a largely publicly financed private delivery system. Singapore has a largely privately financed public delivery system.

There’s also more granular control of the delivery system. In 1997, there were about 60,000 ambulance calls, but about half of those were not for actual emergencies. What did Singapore do? Itdeclared that while ambulance services for emergencies would remain free, those who called for nonemergencies would be charged the equivalent of $185.

Of course, this might cause the public to be afraid to call for real emergencies. But the policy was introduced with intensive public education and messaging. And Singaporeans have identifier numbers that are consistent across health centers and types of care.

“The electronic health records are all connected and data are shared between them,” said Dr. Marcus Ong, the emergency medical services director. “When patients are attended to for an emergency, records can be quickly accessed, and many nonemergencies can be then cleared with accurate information.

“By 2010, there were more than 120,000 calls for emergency services, and very few were for nonemergencies.”

The good times may not last

Singapore made big early health leaps, relatively inexpensively, in infant mortality and increased life expectancy. It did so in part through “better vaccinations, better sanitation, good public schools, public campaigns against tobacco” and good prenatal care, said Dr. Wong Tien Hua, the immediate past president of the Singapore Medical Association.

But in recent years, as in the United States, costs have started to rise much more quickly with greater use of modern technological medicine. The population is also aging rapidly. It’s unlikely that the country’s spending on health care will approach that of the United States (18 percent of gross domestic product), but the days of spending significantly less than the global average of 10 percent are probably numbered.

Medical officials are also worried that the problems of the rest of the world are catching up to them.They’re worried that diabetes is on the rise. They’re worried that fee-for-service payments are unsustainable. They’re worried hospitals are learning how to game the system to make more money.

But they’re also aware of the possible endgame. One told me, “Nobody wants to go down the United States route.”

Perhaps most important, the health care system in Singapore seems more geared toward raising up all its citizens than on achieving excellence in a few high-profile areas.

Without major commitments to spending, we in the United States aren’t likely to see major changes to social determinants of health or housing. We also aren’t going to shrink the size of our system or get everyone to move to big cities.

It turns out that Singapore’s system really is quite remarkable. It also turns out that it’s most likely not reproducible. That may be our loss.

This article originally appeared in <a href="https://www.nytimes.com/2019/04/22/upshot/singapore-health-system-lessons.html" id="5c1f5873-7042-3033-ae1c-7f321d20d2b4">The New York Times</a>.

(The Upshot) ; (The New Health Care)

What Can the U.S. Health System Learn From Singapore?

I’m a pediatrician, and when I see adolescents in my clinic, I ask them if they are drinking alcohol (among other risky activities). Then I counsel them if they answer in the affirmative. I want young people to be safe.

But doctors lack the evidence base — we don’t have enough studies — to know how much of a difference this makes.

Here’s why we may want an answer. Excessive drinking is responsible for 88,000 deaths per year in the United States, about one in 10 deaths among working-age adults. The cost in 2010 was almost $250 billion.

And drinking is a serious problem among adolescents. More than 9 percent of those 12 to 17 years drink alcohol, and almost 5 percent engaged in binge drinking in the last month.

We know how to screen people for alcohol misuse. It’s not hard. The Alcohol Use Disorders Identification Test-Consumption asks three simple questions: how often people drink, how much, and how often they drink heavily.

The Single Alcohol Screening Question asks simply, “How many times in the past year have you had 5 (men)/4 (women) or more drinks in a day?”

The latter is positive (when the answer is more than once) in about 80 percent of people who have unhealthy use of alcohol (as defined by the National Institute on Alcohol Abuse and Alcoholism) and negative in about 80 percent of people who do not. That’s pretty good for a quick one-question screen.

There are also specific screening tools that have been validated for pregnant women, adolescents and older people.

Asking a question in a visit costs very little, other than time. It also has few harms. So why not recommend it for everyone?

Screening is useful when a condition is prevalent, when we can screen in a cost-effective manner, when early diagnosis makes a difference, and when we know that acting on that screen will make a difference.

Susan Curry, immediate past chairwoman of the U.S. Preventive Services Task Force, told me: “There have simply not been enough studies that assess the effects of screening and behavioral counseling in adolescents to address this issue. In addition, many of the existing studies have been conducted in school settings, and it is unclear if the results would translate to primary care.”

Theeffectsof behavioral counseling on adults are proven enough to show a moderate benefit. The same holds true for pregnant women, although the evidence is a bit less strong. But when it comes to adolescents, the evidence is almost nonexistent.

In a large systematic review, there were two studies (out of 68 total) that addressed how counseling interventions might affect adolescent alcohol use. One was a randomized controlled trial of 119 urban adolescents screened in a medical clinic, published in 2015, that used peer networks and motivational interviewing. Overall, it failed to achieve statistical significance with respect to reducing alcohol use.

The other was a 2017 randomized controlled trial in Switzerland schools (not clinics) of a web-and-text-based peer-network messaging system to reduce single-occasion binge drinking. It decreased the chance of that happening by 5.9 percent in the intervention group. But it had no effect on how much youths drank overall or the estimated peak blood alcohol concentration during binge drinking.

That’s all we’ve got. And before 2015, there was nothing.

“We are calling for additional research in all areas related to screening adolescents for alcohol use and providing counseling interventions to those who need them,” Curry said. “It is also important to explore whether such interventions could reduce other risky behaviors. In the meantime, we encourage primary care clinicians to use their clinical judgment when deciding whether to screen their adolescent patients for alcohol use.”

There’s disagreement on the definition of unhealthy alcohol use for adults.

For men, risky drinking is more than four drinks on any day or more than 14 drinks in any week, according to the national institute. And for women, it’s more than three drinks on any day or more than seven drinks in any week. (The institute defines a drink as 12 ounces of beer, 5 ounces of wine or 1.5 ounces of spirits.)

The American Society of Addiction Medicine, on the other hand, defines “at-risk use” as “use that increases the risk for health consequences.”

Young people tend to consume more alcoholic beverages when they’re drinking than adults do. Most of their drinking is binge drinking. Excessive drinking among adolescents — what we’re talking about screening for here — accounts for more than 4,300 deaths each year.

The economic costs of excessive drinking in just this population were $24 billion in 2010. Of those who drive, about 8 percent report having driven after drinking alcohol; 20 percent report having been in a car with someone who had been drinking. About 20 percent of teenage drivers who have died in crashes had been drinking.

Physicians who assure confidentiality and earn the trust of their patients can screen for alcohol use in adolescents pretty well. It’s not that we can’t uncover the problem. I’m often humbled by what my young patients will tell me in private. The issue is that we don’t have the knowledge for what to do with a “yes.” We don’t know that what we’re doing actually helps.

Although it endorsed screening of adults and pregnant women in a recent statement, the U.S. Preventive Services Task Force did not recommend universal screening for alcohol use among adolescents in primary care settings. That speaks more to a failure of our knowledge base — we need more studies — than to the failure of the recommending body to recognize the importance of the problem.

This article originally appeared in <a href="https://www.nytimes.com/2019/04/08/upshot/teenagers-risky-drinking-doctors-counseling.html" id="107d2140-b293-3418-aa75-98e3ebafbc5e">The New York Times</a>.

Can Doctors Talk Teenagers Out of Risky Drinking?

More than 30 million people in the United States have diabetes. The vast majority of them have Type 2 diabetes. Some of those are testing their blood sugar at home, but the best research is telling us that they do not need to — that in fact it is a waste of money.

It is not a small problem. The waste is running into the billions of dollars and it is costing all of us money through the health care system.

For people with Type 1 diabetes, blood glucose monitoring and insulin administration is the standard of care. Patients need to check their blood sugar a number of times a day, then give themselves insulin to replace what would have been made in the pancreas. Treatment for Type 2 diabetes, however, does not involve these critical calculations of insulin. It is usually maintained with pretty regular administration of the same drugs on a set schedule.

Self-monitoring for blood glucose, therefore, may be unnecessary. This has been tested in well-designed studies.

The Monitor trial, published two years ago in JAMA, was a pragmatic trial that took place in 15 primary care practices in North Carolina. Patients with Type 2 noninsulin-treated diabetes were randomly assigned to one of three groups.

People in the first group were told to check their blood glucose once a day. People in the second were told to check their blood glucose once a day and then were given tailored advice depending on the results from the meter. The third group was told not to check blood sugar at all.

After one year (a pretty impressive length for a study like this), there were no differences in the hemoglobin A1C levels (the best way to monitor long-term blood glucose control) between the three groups. There were also no differences in the health-related quality of life measures for the patients. There were no differences in the number of times they experienced hypoglycemia, how much care they needed and how many progressed to the need for insulin.

In other words, there were no measurable differences in how patients fared, whether they checked blood sugar or not.

This evidence, while the best to date, confirmed what previous work had shown. A 2012 Cochrane review assessed all the randomized controlled trials through 2011 that had investigated how testing for blood glucose at home improved outcomes. It included 12 trials involving more than 3,200 patients. By 12 months, the overall benefit to testing, with respect to lab values, was statistically insignificant. There were never any benefits with respect to patient satisfaction.

Still, not everyone is on board. Critics of this most recent trial said it did not prove that blood glucose monitoring could not help: It is possible that with better training, or more attention to detail, there might be ways to make this work.

The point of pragmatic comparative effectiveness trials like this, though, is to test how practices work in the real world. In these high-quality primary care practices, even with customized help in interpreting the measurements (which is more than most patients get), testing blood sugar did not make a difference.

Choosing Wisely, an educational campaign aiming to reduce unnecessary medical tests and procedures, advises against routine home glucose monitoring for patients with Type 2 diabetes who are not on insulin. It says that there is no benefit and that there are potential harms. This argument is supported by the American Academy of Family Physicians, the Society of General Internal Medicine and the Endocrine Society.

Of course, there are exceptions. When patients are acutely ill, or changing regimens, or finding that their blood sugar is not well controlled, testing may be appropriate. Such decisions should be made in consultation with a physician.

But for most people with Type 2 diabetes not on insulin, testing is inappropriate most of the time. That message is not getting through. At the end of last year, another study was published in JAMA Internal Medicine that quantified the prevalence of glucose testing in adults. Researchers examined a database that contained data on more than 370,000 commercial health insurance and Medicare Advantage beneficiaries who had Type 2 diabetes.

Of the more than 23 percent of patients who were using testing strips, more than half were probably doing so despite widespread recommendations that they shouldn’t. They were using a median of two testing strips a day at a cost of more than $325 per year per patient.

On another front, The New York Times has reported that “a surprisingly large number of people” who use insulin are using less because they cannot afford it, putting themselves in danger.

With a health care system as complicated as ours, it is hard to take money from one pot and shift it easily to another. Efficiency in each system is crucial. The fact that a necessary facet of diabetes care is increasingly out of reach — while unnecessary and potentially harmful care is easily overused — illustrates how much work still needs to be done.

This article originally appeared in <a href="https://www.nytimes.com/2019/03/11/upshot/a-diabetes-home-test-can-be-a-waste-of-time-of-money.html" id="e3a64703-6be4-3c70-a79b-4ee5858fc58c">The New York Times</a>.

A Diabetes Home Test Can Be a Waste of Time and Money

The racist photo in the medical school yearbook page of Gov. Ralph Northam of Virginia has probably caused many physicians to re-examine their past.

We hope we are better today, but the research is not as encouraging as you might think: There is still a long way to go in how the medical field treats minority patients, especially African-Americans.

A systematic review published in Academic Emergency Medicine gathered all the research on physicians that measured implicit bias with the Implicit Association Test and included some assessment of clinical decision making. Most of the nine studies used vignettes to test what physicians would do in certain situations.

The majority of studies found an implicit preference for white patients, especially among white physicians. Two found a relationship between this bias and clinical decision making. One found that this bias was associated with a greater chance that whites would be treated for myocardial infarction than African-Americans.

The Implicit Association Test has its flaws. Although its authors maintain that it measures external influences, it is not clear how well it predicts individual behavior. Another, bigger systematic review of implicit bias in health care professionals was published in BMC Ethics, also in 2017. The researchers gathered 42 studies, only 15 of which used the Implicit Association Test, and concluded that physicians are just like everyone else. Their biases are consistent with those of the general population.

The researchers also cautioned that these biases are likely to affect diagnosis and care.

A study published three years earlier in the Journal of the American Board of Family Medicine surveyed 543 internal medicine and family physicians who had been presented with vignettes of patients with severe osteoarthritis. The survey asked the doctors about the medical cooperativeness of the patients, and whether they would recommend a total knee replacement.

Even though the descriptions of the cases were identical except for the race of the patients (African-Americans and whites), participants reported that they believed the white patients were being more medically cooperative than the African-American ones. These beliefs did not translate into different treatment recommendations in this study, but they were clearly there.

In 2003, the Institute of Medicine released a landmark report on disparities in health care. The evidence for their existence was enormous. The research available at that time showed that even after controlling for socioeconomic factors, disparities remained.

There is significant literature documenting that African-American patients are treated differently than white patients when it comes to cardiovascular procedures. There were differences in whether they received optimal care with respect to a cancer diagnosis and treatment. African-Americans were less likely to receive appropriate care when they were infected with HIV. They were also more likely to die from these illnesses even after adjusting for age, sex, insurance, education and the severity of the disease.

Disparities existed for patients with diabetes, kidney disease, mental health problems, and for those who were pregnant or were children.

The report cited some systems-level factors that contributed to this problem. Good care may be unavailable in some poor neighborhoods, and easily obtained in others. Differences in insurance access and coverage can also vary by race.

But the report’s authors spent much more time on issues at the level of care, in which some physicians treated patients differently based on their race.

Physicians sometimes had a harder time making accurate diagnoses because they seemed to be worse at reading the signals from minority patients, perhaps because of cultural or language barriers. Then there were beliefs that physicians already held about the behavior of minorities. You could call these stereotypes, like believing that minority patients would not comply with recommended changes.

Of course, there is the issue of mistrust on the patient side. African-American patients have good reason to mistrust the health care system; the infamous Tuskegee Study is just one example.

In its report, the Institute of Medicine recommended strengthening health plans so that minorities were not disproportionately denied access. It urged that more underrepresented minorities be trained as health care professionals, and that more resources be directed toward enforcing civil rights laws.

In practice, it endorsed more evidence-based care across the board. It noted the importance of interpreters, community health workers, patient education programs and cross-cultural education for those who care for patients.

All of this has met with limited success.

In 2017, the Agency for Healthcare Research and Quality issued its 15th yearly report on health care quality and disparities, as called for by the medical institute in 2002. It found that while some disparities had gotten better, many remained. The most recent data available showed that 40 percent of the quality measures were still worse for blacks than whites. Other groups fared worse as well. Measures were worse for 20 percent of Asian-Americans, 30 percent of Native Americans, and one-third of Pacific Islanders and Hispanics.

Of the 21 access measures tracked from 2000 to 2016, nine were improving. Nine were unchanged. Three were worsening.

It would be easy to look at a racist photo from the 1980s and conclude that it was a different time and that things have changed. Many things have not. We know that racism, explicit and implicit, was pervasive in medical care back then. Many studies show that it is still pervasive today. The recommendations from the medical institute in 2003 still hold. Any fair assessment of the evidence suggests much work remains to be done.

This article originally appeared in <a href="https://www.nytimes.com/2019/02/22/upshot/doctors-and-racial-bias-still-a-long-way-to-go.html" id="29bad9ca-9298-3240-a5f0-fd684cbf142e">The New York Times</a>.

Doctors and Racial Bias: Still a Long Way to Go

Are we underestimating the harms of legalizing marijuana?

Those who hold this view have been in the news recently, saying that research shows we are moving too far too fast without understanding the damage.

The United States is in the midst of a sea change in policies on pot and we should all be a bit nervous about unintended consequences.

Vigilance is required. But it should be reasoned and thoughtful. To tackle recent claims, we should use the best methods and evidence as a starting point.

Does Marijuana Increase Crime?

Crime has gone up in Colorado and Washington since those states legalized marijuana. It is reasonable to wonder about the connection, but it is also reasonable to be skeptical about causation.

The best method to investigate this may be synthetic controls. Researchers can use a weighted combination of similar groups (states that are like Colorado and Washington in a number of ways) to create a model of how those states might have been expected to perform with respect to crime without any changes in marijuana laws. Benjamin Hansen, a professor of economics at the University of Oregon, used this methodology to create a comparison group that most closely resembled the homicide trends and levels from 2000-12.

“I picked those years because they were after the tremendous crime drop in the early ‘90s and most predictive of crime today,” he said. “I ended in 2012 because that’s when Colorado and Washington voted to legalize marijuana.”

This model showed that we might have predicted more of an increasein Colorado or Washington just based on trends seen in comparable states, even without legalization. When he compared the two states with the synthetic control, Colorado and Washington actually had lower rates after legalization than you would expect given trends.

This is not evidence that legalization lowers crime rates. But it does suggest that we should not conclude that it increases them. A number of other studies agree.

What About Car Crashes?

A potential misperception involves automobile crashes. Drunken drivers are measurably impaired when their blood alcohol level is above a certain level. We can prove this in randomized controlled trials.

The tests we use for measuring the presence of THC, though, do not measure the level of impairment. They measure whether someone has used marijuana recently. If we legalize the drug, and more people use it, more people will register its recent use even when they are not impaired. So it should be expected that more people involved in car crashes will test positive even if no one is driving while high.

Using a synthetic control approach, Hansen and colleagues showed that marijuana-related fatality rates did not increase more after legalization than what you would expect from trends and other states.

The Concerns About Schizophrenia

Dr. Ziva Cooper is one of the authors of the National Academies of Sciences, Engineering, and Medicine’s comprehensive report on cannabis.

She says some have misinterpreted the report to state that the report’s committee concluded that cannabis causes schizophrenia. It did not.

“This was stated as an association, not causation,” she said. “We do not yet have the supporting evidence to state the direction of this association.”

Cooper, research director of the UCLA Cannabis Research Initiative, went further: “We as a committee also concluded that a history of cannabis use is associated with better cognitive outcomes in people diagnosed with psychotic disorders. The blatant omission of this conclusion exemplifies the one-sided nature of some articles. Nonetheless, the strong association between cannabis use and schizophrenia means that people with predisposing risk factors for schizophrenia should most certainly abstain from using cannabis.”

Marijuana Has Downsides

No one should be under the impression that marijuana is harmless. The potential downsides are well known and I have covered them. Nor should anyone be irrationally exuberant about its upsides. It is not a wonder drug and the proven benefits are also minimal (as I discussed here).

We should be honest about what we do and do not know. We need more research. It is true that much of the literature around marijuana focuses on the negative, but that is “largely due to funding priorities over the last several decades,” Cooper said.

In the report she worked on, only 40 of the 450 pages were about the therapeutic effects of cannabis and cannabinoids, she said, while the other sections were related largely to the negative health outcomes.

She added, “With increased awareness of the clinical potential of cannabinoids, research priorities have shifted to include studying this area” in the last few years.

It is perfectly natural to be concerned that as cannabis products become legal in more states, they will affect more people.

Many of the experts who have done the work highlighted here are still nervous about how we might proceed. No one thinks that children or adolescents should use marijuana. There is little regulation right now and there is potential for the drug to be mixed with other substances to increase its addictive properties. Advertising will probably make claims that will be out of line with reality.

We should be clear about what we are talking about here, though. Adults will make decisions on how to use it just as they do with similar products like alcohol and tobacco. Both are more dangerous than marijuana and it is not even close.

Anecdotes can make compelling cases, but they do not necessarily lead to thoughtful outcomes.

This article originally appeared in <a href="https://www.nytimes.com/2019/01/14/upshot/the-reasonable-way-to-view-marijuanas-risks.html" id="73e38a40-20c4-3a23-a5cc-de6a3cde09a2">The New York Times</a>.

The Reasonable Way to View Marijuana's Risks

A number of recent news articles have brought renewed attention to financial conflicts of interest in medical science. Physicians and medical administrators had financial links to companies that went undeclared to medical journals even when they were writing on topics in which they clearly had monetary interests.

Most agree such lapses damage the medical and scientific community. But our focus on financial conflicts of interest should not lead us to ignore other conflicts that may be equally or even more important. Such biases need not be explicit, like fraud.

“I believe a more worrisome source of research bias derives from the researchers seeking to fund and publish their work, and advance their academic careers,” said Dr. Jeffrey Flier, a former dean of Harvard Medical School who has written on this topic a number of times.

How might grant funding and career advancement — even the potential for fame — be biasing researchers? How might the desire to protect reputations affect the willingness to accept new information that reverses prior findings?

I’m a full professor at Indiana University School of Medicine. Perhaps the main reason I have been promoted to that rank is that I have been productive in obtaining large federal grants. Successfully completing each project, then getting that research published in high-profile journals, is what allows me to continue to get more funding.

A National Institutes of Health regulation sets a “significant financial interest” as any amount over $5,000. It is not hard to imagine that being given thousands of dollars could influence your thinking about research or medicine. But let’s put things in perspective. Many scientists have been awarded millions of dollars in grant funding. This is incredibly valuable not only to them but also to their employers. Journals and grant funders like to see eye-catching work. It would be silly not to think that this might also subtly influence thinking and actions. In my own work, I do my best to remain conscious of these subtle forces and how they may operate, but it is a continuing battle.

Getting positive results, or successfully completing projects, can sometimes feel like the only way to achieve success in research careers. Just as those drivers can lead people to publish those results, it can also nudge them not to publish null ones.

As a pediatrician, I have been acutely aware of concerns that relationships between formula companies and the American Academy of Pediatrics might be influencing policies on feeding infants. But biases can occur even without direct financial contributions.

If an organization has spent decades recommending low-fat diets, it can be hard for that group to acknowledge the potential benefits of a low-carb diet (and vice versa). If a group has been pushing for very low-sodium diets for years, it can be hard for it to acknowledge that this might have been a waste of time, or even worse, bad advice.

There are things we can do to help mitigate the effects of biases. We can ask researchers to declare their methods publicly before conducting research so that they can’t later change outcomes or analyses in ways that might influence the results. Think of this as a type of disclosure.

A 2015 study published in PLOS ONE followed how many null results were found in trials funded by the National Heart, Lung and Blood Institute before and after researchers were required to register their protocols at a public website. This rule was introduced in 2000 in part because of a general sense that researchers were subtly altering their work — after it was begun — to achieve positive results. In the 30 years before2000, 57 percent of trials published showed a “significant benefit.” Afterward, only 8 percent did.

Moves toward open science, and for a change in the academic environment that currently incentivizes secrecy and the hoarding of data, are perhaps our best chance to improve research reproducibility. Recent studies have found that an alarmingly high share of experiments that have been rerun have not produced results in line with the original research.

We could also require disclosure of other potential conflicts just as we do with ties to companies. In early 2018, the journal Nature began requiring authors to disclose all competing interests, both financial and nonfinancial. The nonfinancial interests could include memberships in organizations; unpaid activities with companies; work with educational companies; or testimony as expert witnesses.

When results are clear and methods are robust, we probably don’t need to worry too much about the subtle biases affecting researchers. When the results are minimally significant, however, and interpretations among experts differ, the biases of those who discuss them probably do matter.

Unfortunately, many results fall into this group. A new drug is minimally better than another, so anyone’s associations with the companies thatproducethem matter when people aremakingdecisions about their use or in writing guidelines. The overall effect of individual nutrient changes is small, but it might have built careers, so it is easy for groups to be too dismissive of new findings that might ask them to change their tunes.

If someone has built a body of research investigating which drugs are better for treating certain infections, that person may discount research that argues we shouldn’t use antibiotics at all.

These conflicts aren’t all the same. Academic researchers are arguably running in many directions with the hope of generally heading toward the truth. Companies are, for the most part, interested in making money. That is not a moral judgment; it is economics. Because of this, it might make sense to put in some hard rules regarding companies. For example, people with financial connections to companies that make medications for high blood pressure should not write guidelines on the treatment of that condition,or be on boards or in positions where they are in charge of policies that could be influenced by those ties.

Even with those rules in place, however, we may need some additional guardrails for scientists and physicians to make sure all research is as unbiased as possible. Such moves would protect us not only from financial conflicts, but also from other types as well.

This article originally appeared in <a href="https://www.nytimes.com/2018/12/31/upshot/congratulations-on-the-promotion-but-did-science-get-a-demotion.html" id="96d67d55-e249-3ec3-ad6d-7f546ad7c848">The New York Times</a>.

Congratulations on the Promotion. But Did Science Get a Demotion?

Despite a number of studies questioning the usefulness of very low-salt diets in the last few years, most major medical organizations continue to recommend them. We would assume that they do so from a strong base of evidence.

But with respect to heart failure, there is a shockingly small amount of evidence.

Heart failure occurs when the heart cannot pump enough blood to the rest of the body. About 5.7 million people in the United States suffer from it. Chronic high blood pressure can force the heart to work too hard over time, weakening it and leading to heart failure. Treatment usually involves trying to strengthen the contractions of the heart muscle with drugs, or reducing the volume of blood with other drugs (like diuretics) or by restricting salt intake.

Recently, researchers reported in JAMA Internal Medicine that they searched for randomized controlled trials that evaluated the use of reduced sodium intake to treat heart failure. In all the literature, they found nine studies that involved 479 patients. One of them was published only in abstract form. None involved more than 100 patients. Overall, none were considered at low risk for bias.

There were no data that showed that salt restriction reduced mortality or cardiac disease; affected whether someone was admitted to the hospital; or influenced how long they had to stay if admitted. Of four outpatient studies, two showed no improvement in heart function and two did.

This is a minuscule amount of data on which to base strong recommendations. An accompanying editorial written by Clyde Yancy, a professor of cardiology at Northwestern School of Medicine, noted that only 0.3 percent of the studies that looked at sodium restriction and heart failure were of sufficient quality to be included in this systematic review.

We need better research. Some of that may be on the way. The Geriatric Out of Hospital Randomized Meal Trial in Heart Failure randomly assigned 66 patients to home-delivered low-salt meals after hospital discharge to study how well they work. Its findings are pending.

Another study, still continuing, has randomly assigned 1,000 people to usual care or a very low sodium diet — 1,500 milligrams a day, what the American Heart Association recommends for most adults — to look at outcomes related to heart failure.

Until then, some will argue that there is little harm from these recommendations, so why not continue them? One reason to stop them is that there is a risk of emphasizing salt avoidance at the expense of other — potentially more useful — diet measures, when we really do not know what is best.

For example, the Agency for Healthcare Research and Quality examined the effects of both sodium and potassium intake on health. It reported this year that lower sodium intake “most likely reduces” blood pressure. But it also reported that increasing potassium intake most likely has the same effect, because consuming more potassium causes you to excrete more sodium in your urine. (Of course, excess potassium has its own risks — any major dietary changes should be discussed with a physician.)

Other systematic reviews find that an increased intake of fiber is associated with a lower risk of heart disease.

It is important to note that we also lack a strong evidence base from randomized controlled trials where potassium and fiber are concerned. But recommendations for those diet changes do not seem nearly as vocal or as strident as the push for salt reductions.

Aaron E. Carroll

Articles written by the author

Health Facts Aren't Enough. Should Persuasion Become a Priority?

What Can the U.S. Health System Learn From Singapore?

Can Doctors Talk Teenagers Out of Risky Drinking?

A Diabetes Home Test Can Be a Waste of Time and Money

Doctors and Racial Bias: Still a Long Way to Go

The Reasonable Way to View Marijuana's Risks

Congratulations on the Promotion. But Did Science Get a Demotion?

Scant Evidence Behind the Advice About Salt

Day Care Directors Are Playing Doctor, and Parents Suffer

A missed opportunity for the malpractice system to improve health care

Why politics should be kept out of miscarriages

The potentially lifesaving difference in how a gun is stored

Does counseling help adolescents who drink alcohol? No one knows

Does counseling help adolescents who drink alcohol? No one knows

My friend's cancer taught me about a hole in our health system

A diabetes home test can be a waste of time and money

The unsung role of the pharmacist in patient health