COVID-19 and racial disparities

On the eve of Martin Luther King Day, let us take a quick look at the impact of COVID-19 by race/ethnicity. CDC reports rates of COVID-19 cases, hospitalizations and deaths. One can see that cases are higher for most minorities (except Asians) relative to Non-Hispanic Whites, but hospitalizations and deaths are much higher for all minorities relative to Non-Hispanic whites. COVID-19 hospitalizations are more than triple the rate for Whites and deaths are more than double.

Source: CDC, as of 17 Jan 2021

One solution to this issue would be to increase vaccination rates across the board. However, according to a survey in November, minorities are less interested in getting COVID-19 vaccinations compared to Whites.

Much progress is needed to get people vaccinated quickly. Israel’s success may provide some learnings on how to get more people vaccinated, fast!

Large decline in cancer mortality

Last year, I published a paper that showed that cancer mortality rates fell by 24% between 2000 and 2016 and further that new cancer drugs accounted for saving 1.3 million lives in the US over that time period. At the end of the paper, we noted that these numbers may be conservative since the use of new innovations–such as immuno-oncology, CAR-T therapy and others–took off after 2016.

It turns out that our premonition was correct, based on the findings from the American Cancer Society’s 2021 Facts & Figures report.

From 1991 to 2018, the cancer death rate has fallen 31%. This includes a 2.4% decline from 2017 to 2018—a new record for the largest 1-year drop in the cancer death rate.

2021 Cancer Facts & Figures

Improvements in health behaviors (e.g., smoking) and new treatments have paved the way for this improvement, particularly for lung cancer.

Reductions in the lung cancer death rateaccount for almost half of the total drop in the cancer death rate from 2014 to 2018. This is thought to be due to declines in smoking, advances in early detection, and improved treatments, especially for non-small cell lung cancer (NSCLC), the most common subtype.

Note that the data don’t include the impact of COVID-19 as they only run through 2018. Further, despite the dramatic improvement in mortality, cancer is still the #2 killer in the US (ignoring COVID), just behind heart disease. However, COVID-19 has passed cancer in terms of US mortality in the spring and with the recent surge has almost certainly surpassed cancer in the fall/winter of this year.

2021 Cancer Facts & Figures

As always, there is lots of interesting information in the ACS report.

COVID-19 now a top 3 leading cause of death in US

That is the conclusion from Woolf et al. (2020) in JAMA this past week. The authors examine rates of COVID-19 mortality in the US between March and October 2020 and compare them with other leading causes of death from March to October 2018. We see that COVID-19 is currently the 3rd leading cause of death in the US, and that does not take into account the recent spike in cases between November and December.

As we all know, COVID-19 is particularly dangerous for the elderly. The mortality rate for those aged 85 and above from COVID-19 is about 100 times higher than those aged 35-44.

This follows on worse news that a new, more transmissible COVID-19 mutation may now be moving throughout the UK. As NPR reports:

The new mutation could be up to 70% more transmissible than earlier variants of the virus, Susan Hopkins of Public Health England told the BBC. This variant was first identified in the middle of October from a sample taken in September, Hopkins said.
“In early December, while we were trying to understand why [cases in] Kent and Medway continued to increase despite the national restrictions, we found a cluster that was growing very fast.”
…While the new variant seems to be more infectious, scientists say there is no evidence that it leads to a more severe illness, according to Vivek Murthy, who President-elect Joe Biden has announced as his nominee for U.S. Surgeon General. “There’s no reason to believe that the vaccines that have been developed will not be effective against this virus as well,” Murthy said on NBC’s Meet the Press.

In short, it’s time to ramp up vaccine production and get people vaccinated!

COVID-19 and Reductions in Cancer Mortality

That is the topic of a commentary I wrote with co-authors Joanna MacEwan and Farzad Ali, titled “Does COVID-19 Threaten the Progress Pharmaceuticals Have Made in Reducing Cancer Mortality Over the Last 20 Years?” An excerpt is below:

Cancer mortality rates have fallen significantly over the last 20 years. Between 2000 and 2010, overall age-adjusted cancer mortality rates decreased by about 1% per year globally.1 In the United States, the trend has been equally pronounced. Overall, the US cancer mortality rate declined by 29% between 1991 and 2017, translating into an estimated 2.9 million fewer cancer deaths.2 Notably, cancer mortality rates fell by 2.2% between 2016 and 2017, the sharpest single-year drop on record.2 
This leaves us with 2 key questions: what factors are the primary causes of the reduction in cancer mortality, and does COVID-19 threaten to stall this progress?3 

We discuss the role of pharmaceuticals in this decline in cancer mortality and what can be done to insure cancer mortality continues to fall even during the COVID-19 pandemic.

Health care spending as a share of GDP rises to 17.7%

In a paper by Martin et al. (2020) in Health Affairs, CMS’s Office of the Actuary (OACT) has its annual health care spending estimates for 2019. They found that health care spending in the US was $3.8 trillion in 2019, and increase of 4.7%. The share of the economy dedicated to health care rose slightly, from 17.6% in 2018 to 17.7% in 2019.

What were the key drivers of these rising health expenditures?

In 2019 faster growth in spending for hospital care, physician and clinical services, and retail purchases of prescription drugs—which together accounted for 61 percent of total national health spending—was offset mainly by expenditures for the net cost of health insurance, which were lower because of the suspension of the health insurance tax in 2019.

The study found that spending on hospitals rose by 6.2%, physician services 4.7%, and retail pharmacy costs by 5.7%.

Are rising prices–especially drug prices–to be blamed?

With the exception of hospital prices, the answer is ‘no’. Hospital prices rose by 2.0% and physician prices rose by a modest 0.8%. Retail pharmacy prices, however, actually fell for the second year in a row.

Retail prescription drug prices decreased 0.4 percent in 2019 after a larger decline of 1.0 percent in 2018 as price growth slowed for brand-name drugs and declined for generic drugs.

Why does the US spend so much more on health care than other countries?

If we break down the numbers, it turns out that medical costs–rather than pharmacy cost–are the key drivers of differences in health care spending per capita. This is the findings from Kaiser Family Foundation (KFF) study comparing US spending against an average from 9 other countries rich countries
(i.e., Austria, Belgium, Canada, France, Germany, the Netherlands, Sweden, Switzerland and the United Kingdom).

While the US spends $513 (+58%) more on drug and medical goods, spending on inpatient and outpatient services are $3906 (+144%) higher in the US compared to the average of these 9 countries.

Hat tip: Adam Fein.

Comparing health care for older adults across 23 countries

How well do different countries take care of their elderly and near elderly populations? A study by Macinko et al. (2020) provides some useful statistics.


  • Population of interest. Individuals aged 50 and above sampled across 23 high- and middle-income countries.
  • Data. The data come from 2015 or 2016 years for the following data sets: Brazilian Longitudinal Study of Aging; Mexican Health and Aging Study; Korean Longitudinal Study of Ageing; Survey of Health, Ageing and Retirement in Europe, which includes Austria, Belgium, Croatia, Czech Republic, Denmark, Estonia, France, Germany, Greece, Israel, Luxembourg, Italy, Poland, Portugal, Slovenia, Spain, Sweden, and Switzerland; Chinese Health and Retirement Longitudinal Study; Health and Retirement Study in the US. While these are large, robust data sources, about 18% of the sample was dropped due to missing variables.
  • Measures. The measures included: (i) whether individuals had a doctors visits in the past year (as a measure of access to care); (ii) more than 15 doctors visit the past year (to indicate over-utilization), (iii) two or more hospitalizations in the past year, and (iv) whether health expenditures reached 25% of household income. While this measure are easily accessible across a number of different countries, they are very crude.


  • Access: “About 11 percent of respondents reported having no doctor visit in the previous year. But this figure varied considerably, with two countries (Czech Republic and Luxembourg) reporting rates of less than 5 percent and two countries (Greece and Mexico) reporting approximately 23 percent of adults without a doctor visit in the past year.”
  • Excess utilization. This was largely the reverse of the access story. Nine percent of the sample reported ≥15 doctor visits in the past year, ranging from a low of <5% in Brazil and Mexico to a high of nearly 16% in Italy.
  • Two Hospitalizations. The lowest rate (<2% of individuals) was observed in South Korea, whereas the highest rate (>9% of individuals) was observed in Mexico and the US .
  • Catastrophic cost. European countries were the best performers here. In Denmark, France, Germany, and Sweden <1% of people had health care costs reach >25% of their income, whereas for >10% of the elderly in Brazil, China, Mexico, and South Korea had health care costs >25% of their income.

Do read the whole article if you are interested more of these statistics.

While one should be careful of interpreting these results as clearly indicating whether a health care system is good or bad–for instance, quality of care is not well-measured–these are helpful basic statistics for cross country comparisons.


Spending on specialty drugs: 2010-2017

Specialty drugs are often injectibles, biologic drugs, or other drugs that require specialized administration, handling, or ongoing clinical assessment. Specialty drugs are often more expensive than standard pharmaceuticals. As the use of specialty pharmaceuticals has become more common, a paper by Hill, Miller and Ding (2020) aims to quantify how much spending on specialty pharmaceuticals has grown. Using data form MEPS between 2010 and 2017, the authors find:

We found that net spending on retail specialty drugs more than doubled from 2010–11 to 2016–17. Specialty drugs accounted for 37.7 percent of retail and mail-order prescription spending net of rebates in 2016–17

The figure below shows that there were large increases in spending on specialty drugs. While the average annual increase in spending on specialty pharmacy increased by 17.4% per year (gross), after rebates, this figure was only 14.9% (net). Rebates were especially large for the Medicaid population.

Increased spending should not be viewed as negative in and of itself. While price of course influences expenditures, so does quantity. In fact, Hill and co-authors find that the share of people using at least one specialty pharmacy drug increased from 2.2% of the population to 5.0% of the population. Specialty drug fills as a percentage of overall fills increased from 1.0% of fills in 2010 to 2.3% of fills in 2017. Thus, the quantity of medicines received rose. If specialty pharmacy represents significant innovation and high value to patients, the rise in spending may be considered a good thing.

To give a parallel example, spending on smartphones over recent decades has grown tremendously, but because consumers have received great value for this, no one is complaining about the rise of smartphone expenditures. In the case of specialty pharmaceuticals, value assessment is needed to insure that the health and societal benefits received are worth these additional costs.


Healthcare spending by age: An international comparison

We all know that health care spending per person in the US is highest among all countries. But are these results consistent across age groups? That is the question posed by Papanicolas et al. (2020). Using 2015 OECD data from the US and 7 other high-income countries (i.e., Australia, Canada, Germany, Japan, the Netherlands, Switzerland, and the United Kingdom), the authors find the following:

In the US, per capita health care spending was $9524, or 1.9-fold higher than the mean for the 7 comparator countries….The absolute difference between US spending and that of the other countries for ages 0 to 4 years was $3899, and that difference decreased at approximately age 5 years, after which it slowly increased. The difference increased faster after age 65 years, peaking at $18 645 for ages 80 to 84 years.

Papanicolas et al. (2020)

In short, health care spending in the US is higher across basically all ages.

Quantifying the digital divide

Telemedicine and video-conferencing are helping to bridge the gap in access to care during the COVID-19 pandemic. These approaches, however, only work if patients have access to the technology needed to engage in telemedicine or videoconference visits.

A paper by Roberts and Mehrota (2020) uses data from the 2018 American Community Survey (ACS) and finds that:

Overall, 41.4% (95% CI, 40.4%-42.4%) of Medicare beneficiaries lacked access to a desktop or laptop computer with a high-speed internet connection at home, and 40.9% (95% CI, 40.0%-41.8%) lacked a smartphone with a wireless data plan…The proportion of beneficiaries without either form of digital access was 26.3% (95% CI, 25.5%-27.1%), and this proportion varied across demographic and socioeconomic groups. 

In summary, about 2 of 9 Medicare beneficiaries have no digital access, 3 of 9 have partial access (either laptop/desktop or smartphone) and 4 out of 9 have full access (both laptop/desktop and smartphone). Hopefully these numbers have improved over the last two years. Nevertheless, the digital divide causes more than economic challenges, but in the time of COVID-19 also causes challenges in accessing health care for the elderly as well.

More opioid negative externalities: Pediatric opioid poisoning

Some people use opioids as a recreational drug. If people choose to ruin their own life, that is their problem. Others take opioids when prescribed from physicians and unwittingly become addicted. In either case, there is one scourge that may not fully be appreciated: parents who take opioids risk having their kids go into their medicine cabinets and–intentionally or unintentionally–use them.

A recent paper by Toce et al. (2020) in JAMA Pediatrics uses the National Poison Data System (NPDS) and finds that 338,476 children or young adults experienced opioid poisoning between 2005 and 2017. The vast majority of these poisonings occurred in very young children (3 or younger) or teenagers (13 to 19 years old).

The authors also find that:

PDMP [prescription drug monitoring program] and pain clinic legislation appeared to be associated with significant reductions in opioid poisoning among children 4 years and younger and adolescents between 15 and 19 years

While opioids are needed by some subset of patients, keeping opioids out of the hands of children should be a public health initiative we can all get behind.

Fentanyl tops drug overdose death rates

RAND has an interesting report titled “The Future of Fentanyl and Other Synthetic Opioids.” What is truly startling is the rise in deaths due to synthetic opioids like fentanyl in the last 5 years: from 3,000 in 2013 to approximately 30,000 in 2018. In fact, synthetic opioids are now involved in twice as many deaths as heroin.

Further, synthetic opioid-related overdoses often occur when individuals use fentanyl along with other illegal substances.

Fentanyl truly has become a top public health priority.

Health care battle: US vs. England

A new study by Choi et al. (2020) compares the health of individuals aged 55-64 in the US compared to England. They use data from Health and Retirement Study (HRS)–in the US–and the English Longitudinal Study of Ageing (ELSA) for 2008-2016. Health is measured across 16 outcomes: 5 self-assessed outcomes, 3 directly measured outcomes, and the prevalence of 8 key comorbidties. They find:

Cross-country differences in health were in favor of England for all health outcomes except ADL limitations, depression, and measured blood pressure, which were not significantly different between countries.

Of perhaps more interest is that they compare health outcomes among individuals in the top compared to bottom income deciles across countries. In this analysis, they find:

Among individuals in the lowest income group in each country, those in the US group vs the England group had significantly worse outcomes on many health measures (10 of 16 outcomes in the bottom income decile); the significant differences in adjusted prevalence of health problems in the US vs England for the bottom income decile ranged from 7.6% (95% CI, 6.0%-9.3%) vs 3.8% (95% CI, 2.6%-4.9%) for stroke to 75.7% (95% CI, 72.7%-78.8%) vs 59.5% (95% CI, 56.3%-62.7%) for functional limitation. Among individuals in the highest income group, those in the US group vs England group had worse outcomes on fewer health measures (4 of 16 outcomes in the top income decile); the significant differences in adjusted prevalence of health problems in the US vs England for the top income decile ranged from 36.9% (95% CI, 33.4%-40.4%) vs 30.0% (95% CI, 27.2%-32.7%) for hypertension to 35.4% (95% CI, 32.0%-38.7%) vs 22.5% (95% CI, 19.9%-25.1%) for arthritis.

The authors do note that these differences persist even after statistically controlling for demographic factors, educational level, smoking, and body mass index. Note that the outcomes are relatively crude measures and the data cover only individuals aged 55 to 64. Further, the study describe what these differences are, but not they why of why these occur. Nevertheless, these results do provide a helpful snapshot comparing the health of Americans and the English across income strata.

What explains the decline in COVID deaths despite the rising number of cases?

Take a look at these two graphs. The first gives the number of new COVID-19 cases from John Hopkins University. It is a scary graph.

The second graph is the number of COVID deaths from IHME. Not nearly as bad.

CDC reported death numbers are similar. In fact, the number of deaths in recent weeks is no different from normal according to the CDC.

One reason for rising cases is that we are doing more testing. Thus, we are catching more cases. People who are hospitalized with critical illness are more likely to be tested and thus as we expand the number of tests, we are now catching more of the less severe cases with the additional testing.

However, testing does not explain everything. We see that as cases are rising, so is the number of hospitalizations. Thus, this is not just a testing story. Real–not just detected–cases are rising.

Most likely, the existence of more effective treatments (e.g., remdesivir and dexamethasone) and physicians gaining experience treating the disease is responsible. Among hospitalized individuals, death rates have been falling dramatically in recent weeks.

There is much still to be learned about the causes of these divergent trends and much more research is needed.