To prescribe or not prescribe? In the realm of the nation’s opioid epidemic, it’s an important question.

Research has shown that inappropriate use of prescription opioids is part of the reason behind a dramatic rise in opioid-related deaths since 2000. By 2015, the amount of opioids prescribed in the U.S. had tripled — enough for every American to be medicated around the clock for three weeks, at 5 milligrams of hydrocodone every four hours. 

Now, new research by a trio of Stanford scholars shows how different insurance strategies affect the volume of opioid use and could help stem inappropriate prescribing behaviors. 

The study, released in a working paper by the National Bureau of Economic Research, was co-authored by Stanford Health Policy’s Laurence C. Baker, M. Kate Bundorf, and Daniel P. Kessler. Baker and Bundorf are professors in the Department of Health Research and Policy at the Medical School; Kessler is a professor in the Law School and Graduate School of Business, and a senior fellow at the Hoover Institution.  All are also senior fellows at the Stanford Institute for Economic Policy Research (SIEPR).

Their study — the first to investigate the effect of the form of Medicare drug coverage on opioid use — found that enrollment in Medicare Advantage, a combined medical and drug insurance plan, significantly reduces the likelihood of beneficiaries filling an opioid prescription, as compared to enrollment in a stand-alone drug plan.

Compared to beneficiaries enrolled in stand-alone plans, those enrolled in the integrated Medicare Advantage plan were 37 percent less likely to get an opioid prescription, according to their analysis of drug claims from 2014.

The researchers also found that enrollment in integrated insurance coverage under Medicare Advantage had a disproportionate effect on the likelihood of filling an opioid prescription from the nation’s highest opioid-prescribing doctors — the top 1 percent of prescribers in Medicare Part D. The lower likelihood of prescriptions from these high prescribers to Medicare Advantage enrollees accounted for more than half of the reduction, according to their findings.

To understand the scope of this health plan-related effect and what’s at stake, consider the backdrop laid out in the study:

Since its implementation in 2006, Medicare Part D has become the nation’s largest purchaser of prescription opioids. More than 42 million Americans are enrolled in Medicare Part D — either under the stand-alone drug plan or the integrated Medicare Advantage plan.

What’s more, opioid prescriptions are concentrated among a relatively small group of “high prescribers.” 

According to research published in the 2016 edition of JAMA Internal Medicine, more than one-third of opioid prescriptions under Medicare Part D were made by about 8,000 doctors, making up the top 1 percent of prescribers. And according to the Office of the Inspector General of the Department of Health and Human Services, “extreme use” and “questionable prescribing” have put almost 90,000 beneficiaries at serious risk for opioid misuse or overdose.

Because the researchers did not examine patient health outcomes, they could not definitively determine that enrollment in Medicare Advantage reduced only inappropriate opioid use. However, because the reduction in opioid use came disproportionately from high prescribers, and previous work has found that Medicare Advantage enrollees had higher prescription drug use overall, the reduction in use that the researchers found was targeted rather than a result of a broader effort to restrict access to treatment.  

The researchers’ results support the conclusions of previous work that integration of prescription drug coverage with the other benefits provided by Medicare Advantage plans

improves the quality of care. Further study will be needed to drill deeper into the reasons behind the impact of Medicare Advantage plans, and whether a similar effect occurs in non-elderly populations, the researchers said.

The national opioid epidemic has grown at such breakneck speed that public health experts have been left scrambling to keep up. Though they understand the reasons behind the abuse — the solutions are complicated and costly.

Yet there appears to be some success at reducing at least one area of opioid abuse.

In new research by Health Research and Policy’s Eric Sun, the risk for chronic opioid use among patients with musculoskeletal pain actually decreased slightly between 2008 and 2014. 

The Stanford Medicine assistant professor of anesthesiology and pain medicine found that measures such as avoiding opioid use soon after diagnosis can further reduce the risk of addiction, especially among patients at highest risk for chronic opioid use.

"We found that early opioid use after diagnosis is predictive of opioid use longer term, suggesting that it may be prudent to minimize opioid use where possible for patients with musculoskeletal pain,” said Sun, whose research was published earlier this week in the Annals of Internal Medicine.

His co-authors are Jasmin Moshfegh, who is working on her PhD in health policy, and Steven Z. George, director of musculoskeletal research at Duke University School of Medicine.

Patients with lower back or chronic neck, shoulder and knee pain are at the highest risk for opioid abuse since pain meds are typically prescribed to help ease their spasms. 

Patients who suffer musculoskeletal pain may unwittingly transition to chronic opioid use, which means filling 10 or more prescriptions or having a supply for at least 120 days. The prescription drugs include hydrocodone, hydromorphone, methadone, morphine, oxymorphone, and/or oxycodone. Those don’t include heroin and synthetic opioids such as fentanyl.

Sun and his fellow researchers at the Stanford University School of Medicine used a large health-care database to assess the risk and risk factors for chronic opioid use among more than 400,000 “opioid-naïve” patients — those who have never been prescribed painkillers before — recently diagnosed with pain in the knee, neck, lower back or shoulder. 

The sample was restricted to privately insured patients, thereby excluding other policy-relevant populations, such as those who were prescribed pain medications under Medicare or Medicaid.

They found that risk for chronic opioid use ranged from 0.3 percent for knee pain to 1.5 percent for multiple-site pan and decreased for some anatomical regions during the timeframe studied. They discovered a relative decline of 25 to 50 percent across all pain types from 2008 to 2014.

Opioid Abuse

Opioid abuse has its roots in the late 1990s when pharmaceutical companies assured the medical community that patients would not become addicted to pain relievers. Clinicians began prescribing them at greater rates because they worked so well and seemed safe.

Today, more than 130 people die every day from opioid-related drug overdoses from prescription pain relievers, heroin and synthetic opioids such as fentanyl, according to the U.S. Department of Health and Human Services, From 2002 to 2017, there was more than a fourfold increase in opioid deaths, with some 49,000 people dying in 2017.

The Centers for Disease Control and Prevention estimates that the total economic burden of prescription opioid misuse alone in the United States is $78.5 billion a year, including the costs of health care, lost productivity, addiction treatment and criminal justice involvement.

“While our research found that only about 1 percent of patients with musculoskeletal pain progress to chronic opioid use, this is potentially concerning because it’s an extremely common diagnosis,” Sun said. “By pointing out the scope of the issue and identifying factors that place patients at risk, we hope this research will guide further efforts aimed at reducing opioid use among patients with musculoskeletal pain.” 

A cutting-edge treatment for blood cancer in children with promising short-term remission rates has nevertheless come under intense scrutiny due to its unprecedented cost.

Acute lymphoblastic leukemia (ALL) is the most commonly diagnosed pediatric cancer. Though treatment advances have driven five-year survival rates above 90 percent, relapse is common and ALL remains a leading cause of death from childhood cancer. Those surviving relapse typically require hematopoietic stem cell transplant (HSCT) to remain in remission.

The Food and Drug Administration last year approved tisagenlecleucel as the first anti-CD19 CAR T-cell therapy for relapsed or refractory pediatric ALL. While tisagenlecleucel-induced remission rates are encouraging compared with those of established therapies — more than 80 percent compared with less than 50 percent — a one-time infusion costs $475,000.

That makes it one of  the most expensive oncologic therapies out there, even though no studies exist to show how the therapy maintains remission in the long term.

So Stanford Health Policy’s John K. Lin, Jeremy Goldhaber-Fiebert and Douglas K. Owens, in collaboration with Kara Davis, an assistant professor of pediatrics at Stanford’s Lucile Packard Children’s Hospital, set out to determine whether the treatment is cost-effective. 

Their study was recently published in the Journal of Clinical Oncology.

“CAR-T cells are an exciting new therapy to help us cure more patients with ALL,” said Davis. “They use a patient’s own immune system to precisely target their leukemia and remission rates are impressive for patients with relapsed disease.”

The researchers note, however, it remains unknown whether tisagenlecleucel is sufficient to cure relapsed or refractory disease without a transplant.

Not only is it a very expensive treatment, it also has expensive and serious side effects, such as cytokine release syndrome, which can cause symptoms ranging from fevers and muscle aches to lethal drops in blood pressure and difficulty breathing, requiring ICU-level care.

“Given [its] high cost and broad applicability in other malignancies, a pressing question for policymakers, payers, and clinicians is whether the therapy’s cost represents reasonable value,” the authors wrote.

In order to evaluate the economic value of tisagenlecleucel, the authors created a computer simulation that modeled children with relapsed or refractory ALL. Since the durability of the therapy’s effects are unknown, they evaluated the therapy at different levels of long-term effectiveness.

Through their analyses, the authors found that at the simulation’s most optimistic projections, patients receiving tisagenlecleucel would, on average, live over a decade longer than those receiving alternative therapies. At these projections, the therapy would be cost-effective. However, at more modest long-term outcomes, its clinical and economic benefits declined.

Image

“The durability of CAR-T cell therapy is the main question. For many children with relapsed or refractory ALL, their disease is fatal — if the therapy results in sustained remissions, it would represent an important advance,” said lead author Lin, who is a VA Health Services Research and Development Fellow at Stanford Health Policy.

In the end, the researchers concluded that at tisagenlecleucel’s current price, its economic value is “uncertain,” as the true cost-effectiveness depends on its long-term performance, which has yet to be determined. They noted that substantial price reductions would improve its cost-effectiveness even if its long-term performance is relatively modest.

“A commonly asked question for many expensive, novel therapies is why can’t prices be lowered at least until we know how well they work,” said Goldhaber-Fiebert. 

The “Catch-22” here is that long-term effectiveness data are needed to justify a drug’s price, but current uncertainty about its effectiveness, along with its high price, means developing the data to justify its price occurs much more slowly.

“CAR-T is an individually tailored treatment that has thus far been produced for a relatively small number of patients,” Goldhaber-Fiebert said. “Its current production costs may well be very high, and certainly the companies that have spent heavily on its research and development are interested in achieving returns on their investments.”

When the effectiveness of a therapy is uncertain, some pharmaceutical companies and health policy experts have proposed something called outcomes-based payment, which is essentially a “money-back guarantee” if the therapy doesn’t work as intended.

Novartis, which developed tisagenlecleucel, has a money-back guarantee; if the patient does not achieve initial remission within the first month, they are not responsible for the payment of the therapy.

“However, we find that because most children have an initial remission, this does not materially improve cost-effectiveness,” Lin said.

He suggested that if the money-back guarantee were extended to see whether the patient relapses within a year, such a guarantee would improve the treatment’s cost-effectiveness.

The other co-authors of the study are James I. Barnes, Alex Q.L. Robinson, Benjamin J. Lerman, Brian C. Boursiquot and Yuan Jin Tan.

African-American doctors could help reduce cardiovascular mortality among black men by 19 percent — if there was more racial diversity among physicians, according to a new study led by Stanford Health Policy’s Marcella Alsan.

After conducting a randomized clinical trial among 1,300 black men in Oakland, the researchers found that the men sought more preventive services after they were randomly seen by black doctors for a free health-care screening compared to non-black doctors.

“We found that, once African-American men were at the clinic, even though all services were free, those assigned to a black doctor took up more services,” such as flu shots and diabetes and cholesterol screenings, said Alsan, an economist and infectious disease physician who focuses on  health and socioeconomic disparities here at home and around the world.

“It was surprising to see the results,” said Alsan, an associate professor of medicine at Stanford Medicine, a faculty fellow at the Stanford Institute for Economic Policy Research, and an investigator at the VA Palo Alto Health Care System. “Prior to doing the study, we really were not sure if there would be any effect, much less the magnitude. The signal in our data ended up being quite strong.”

Those signals include the men were 29 percent more likely to talk with black doctors about other health problems and seeking more invasive screenings that likely required more trust in the person providing the service. They found subjects assigned to black doctors increased their uptake of diabetes and cholesterol screenings by 47 percent and 72 percent, respectively.

The researchers calculated that black doctors could reduce cardiovascular mortality by 16 deaths per 100,000 per year, accounting for 19 percent of the black-white gap in cardiovascular-related deaths. They believe that the results would be even larger if extrapolated to other leading causes of death that are amenable to prevention, such as cancer and HIV/AIDS. 

“I was definitely surprised,” said Owen Garrick, president and COO of Bridge Clinical Research, an Oakland-based organization that helps clinical researchers find patients from targeted ethnic groups. “If you ask most people, they feel that there is some impact of black men seeing black doctors — but it has never been quantified using an experimental design.”

Alsan and Garrick, along with U.C. Berkeley graduate student Grant Graziani, published their findings in this working paper for the National Bureau of Economic Research.

Garrick, himself an African-American physician, said black doctors tend to present themselves in a manner that puts a black patient at ease, making him more willing to open up and agree to certain care. “The black doctor might explain the medical services in a way that the black patient more clearly understands.”

Garrick called the findings “astounding,” but he warned that increasing the number of black doctors and getting black men to routinely see them are no small tasks.

There is a yawning gap between white physicians and those of color. While African-Americans comprise about 13 percent of the population, only 4 percent of physicians and less than 6 percent of medical school graduates are black, according to the study.

This is compounded by African-American men having the lowest life-expectancy in the country, due to lack of health insurance, lower socioeconomic status and structural racism. 

And there remains a distrust of the U.S. healthcare system at least partially attributed to the infamous Tuskegee study that began in 1932, when the U.S. Public Health Service began following about 600 African-American men in Tuskegee, Alabama. Some two-thirds of the men had syphilis, and USPHS declined to inform those afflicted by the disease. Even after penicillin became the standard of care for syphilis treatment in the mid-1940s, the USPHS continued to withhold treatment. The study was finally halted when a whistleblower went to the press in 1972.

Alsan— with her colleague Marianne Wanamaker at the University of Tennessee — published a study in The Quarterly Journal of Economics in February that found the 1972 Tuskegee study revelation was correlated with a reduction in health-seeking behavior and increases in medical mistrust and mortality among African-American men.

Image

The men who participated in the recent study were recruited from barbershops and flea markets in Oakland, a city known for its diversity, yet plagued by a 20 percent rate of poverty.

Field officers —including minority and low-income pre-med students from around the Bay Area — approached men to enroll in the study. After obtaining written consent, the men were given a short survey about socio-demographics, health care and mistrust. For completing the survey, the men received a voucher with up to $25 for their haircut or, in the flea market, a cash incentive.

The men were also given a coupon to receive a free health-care screening for blood pressure, BMI, cholesterol and diabetes at the clinic where the Stanford team operated on Saturdays in the fall and winter of 2017-2018. The patients who did not have transport to the clinic were given free rides courtesy of Uber. Attendance at the clinic was encouraged with another $50 incentive.

Subjects and the 14 participating doctors were told that they were taking part in a Stanford study designed to improve preventive health-care for African-American men.

On top of the significant increases in patients who agreed to diabetes or cholesterol screenings if suggested by a black doctor, the researchers found that the men were 56 percent more likely to get a flu vaccine if randomized to one of the African-American doctors.

The results suggested the more invasive the test, the greater the advantage of being assigned a black doctor. And the findings were even stronger among subjects who had a high mistrust of the medical system as well as those who had limited prior experience with routine medical care.

“In curative care, the patient feels ill and then may seek out medical care to fix the problem,” Alsan said. “But in preventive care, the patient may feel just fine — but must trust the doctor when he is told that certain measures must be taken to safeguard health.” 

The policy implications would suggest that medical schools need to open the pipeline to students from diverse backgrounds who are training for health-care professionals. 

Garrick recommends exposing more young people of color to the field of medicine and helping them to become more competitive applicants through tutoring and interview prep.

“And you need advocates,” he said. “Since much of the medical school selection process is subjective, you need to get people on the selection committees who will relate and see the potential of black applicants as much as people relate to other applicants.”

Some links to other media outlets that have written about this research: 

The New York Times

Harvard Business Review

The Daily Mail

ColorLines

As global health assistance for developing countries dwindles, a Stanford student working on her PhD in health policy has developed a novel formula to help donors make more informed decisions about where their dollars should go.

Donors have typically relied predominately on gross national income (GNI) per capita to determine aid allocations. But using GNI is problematic because it effectively penalizes economic growth. It also fails to capture contextual nuances important to channeling aid effectively and efficiently.

So Tara Templin, a first-year Stanford PhD student specializing in health economics, and her Harvard colleague Annie Haakenstad, have developed a framework that estimates funding based on needed resources, expected spending and potential spending into 2030. They believe the more flexible model makes it adaptable for use by governments, donors and policymakers.

“We've observed development assistance for health growth attenuate over the last seven years,” said Templin, who was a research fellow at the Institute for Health Metrics and Evaluation before coming to Stanford. “There are difficult trade-offs, and this entails honing in on the specific challenges and countries most in need.”

Their research published in the journal Health Policy and Planning outlines how their “financing gaps framework” can be adapted to short- or long-run time frames, between or within countries.

“Depending on donor preferences, the framework can be deployed to incentivize local investments in health, ensuring the long-term sustainability of health systems in low- and middle-income countries, while also furnishing international support for progress toward global health goals,” write the authors, who also are Stephen Lim of the University of Washington, Jesse B. Bump of Harvard and Joseph Dieleman, also at the University of Washington.

The authors developed a case study of child health to test out their framework. It shows that priorities vary substantially when using their results as compared to focusing mainly on GNI per capita or child mortality.

The case study uses data from the Global Burden of Disease 2013 Study, Financing Global Health 2015, the WHO Global Health Observatory and National Health Accounts. Funding flows are anchored to progress toward the U.N. Sustainable Development Goals’ target for reductions in the death rates of children under 5. More than six million children die each year before their fifth birthday, so the United Nations set a goal to reduce under-5 mortality to at least 25 per 1,000 live births.

To build their child health case study, the authors relied on a 2015 study that estimated the average cost per child-life saved is $4,205 in low-income countries, $6,496 in lower-middle income countries and $10,016 in upper-middle countries.

The framework considers three concepts. First, expected government spending is constructed from national health accounts, which are standardized financial reports from countries around the world. Second, ability to pay is estimated by looking at countries with similar levels of economic development and looking at associations with country investment in the health sector. Lastly, needed investment considers a health target, the country’s current health burden, and average costs to save children’s lives in each country.

“Our focus is on the gap between the resources needed to reach critical health targets and domestic health spending,” the authors wrote. “We highlight two facets of domestic health resources—expected spending and potential spending—as critical. While donor preferences may vary, basing aid allocation on expected or existing spending levels incentivizes countries to spend less on health. We therefore propose the use of potential spending, which is a measure of a country’s ability to pay, as the domestic resource benchmark.”

Image

Instead of the gap between expected spending and need, their framework focuses on the gap between potential spending and the health resources needed to meet global health targets. In the framework, policymakers can choose which gap they want to target, since this decision can involve many factors.

“By focusing on that gap, donors can catalyze sustained domestic spending while also addressing the resource needs critical to reaching international health goals,” they wrote.

They then looked at 10 countries with the most need for additional child health resources. The gap between expected spending and potential spending was highest in Afghanistan, at 79 percent, and lowest in Cameroon, where expected spending exceeded potential spending.

“Fifty years ago, GNI was the best proxy for countries’ ability to finance their own development and health,” the authors wrote.

But today, more empirical data and technology are available, allowing donors to incorporate a broader set of health financing measures into their decision-making process.

“The flexible but targeted nature of our framework is critical in the current era of global health financing,” said Haakenstad, the lead author. “Our framework helps to ensure the poor and disadvantaged, the majority of which now reside in middle-income countries, are reached by development assistance and other public financing. This funding is critical to reducing death and disability and reaching global targets in health.”

The authors’ research was supported by the Welcome Trust (099114/Z/12/Z).