It seemed like it was time to really prove the risk for kids. We already knew that poor sleep can be a predictor of suicidal ideation and behavior. Suicide is a leading cause of death among adolescents, who have high rates of sleep disturbance. But it’s still rare to find studies on the risk of suicidal behaviors in the transition from childhood to early adolescence.

We simply established that there was an association between sleep disturbance, starting at around age 10, and later suicidal thoughts and behavior. For the kids we studied, those thoughts and behaviors were not in their initial interview. But they could have occurred any time in the two years between first assessment and follow-up.

We could give all sorts of hypotheses. But we don’t really know at this point what is causing it. I think it’s time right now to start unpacking it to find out. So, the next step is to look at lots of risk factors for long-term prediction.

There has not been a comprehensive study of CAR-T cell therapy that calculates overall non-relapse mortality. If we can learn the cause of adverse events, what does that mean for patients and how can we use that information in the clinic?

We found the non-relapse mortality rate for patients with lymphoma and multiple myeloma one year after treatment was about 6.8 percent, which was higher than we expected. We identified 7,604 patients across 46 studies published through March 2024.

More than half of the patients died from infections. We looked for specific pathogens, and most of the known infection-related deaths reported were listed as caused by COVID-19. However, about two-thirds of the pathogens were not identified or not reported.

The second most common cause was death from a secondary malignancy, so another cancer had developed.

We would like to see improved guidelines for clinicians who have patients on CAR-T cell treatment and raise awareness about the risk of infections and secondary malignancies. There also can be better patient education; if you have a fever, call your doctor immediately. We asked researchers to report these deaths in more detail in the future. Now we see that studies are starting to report this.

The United States has health disparities along many different dimensions. About 20 years ago, a paper called “Eight Americas” approached this problem by studying groups of people in the U.S. based on where they lived, their racial identity, and their income. We wanted to revive that approach and look at how life expectancy disparities have evolved since 2000.

We had about a 13-year gap in life expectancy across the U.S. groupings in 2000, and by 2010, that gap had grown, spanning 71.2 years to 85.2 years. The decade from 2010 to 2019 was not good for life expectancy in the United States. It really stagnated. In the end, the gap spanned almost 16 years, from 70.2 to 86 years. American Indian and Alaska Native populations had the shortest life expectancies and Asian American populations had the longest. All U.S. groups saw declines in life expectancy during [the COVID-19 years we measured, 2020 and 2021], but the size of the declines really varied.

I do not think there is one solution to this problem, but I would like to see much more attention given to why the United States is so far behind other high-income countries. Research points toward fundamental causes such as income, education, employment opportunities, and housing conditions. There is a little bit of risk in focusing on specific causes of death because it can obscure these society-wide features that are probably responsible.

Public health is inherently political. Policies don’t just appear—they are enacted by legislators, implemented by state agencies, and influenced by political ideologies. Yet, political metrics like voting records and state trifectas are rarely used in public health research. I was motivated to investigate the political determinants of health through a fresh lens.

We analyzed over a decade of data and found consistent associations between political conservatism and poorer health outcomes. The strongest correlations came from metrics rarely used in public health—state trifectas and political ideologies of elected officials.

For example, premature mortality rates were significantly higher in Republican trifecta states, which had 55.4 more deaths per 100,000 person-years compared to Democratic trifectas in 2016. To capture the immediate effects of governance, we focused on outcomes for which the time period from exposure to the health event is very fast (such as getting vaccinated and getting, or losing, health insurance coverage) as opposed to health events that could take a long time to show up, such as getting and being diagnosed with cancer. Our findings underscore the tangible impact of political decisions on population health.

This research highlights the need for evidence-based governance. Legislators, advocacy groups, and public health agencies must prioritize health equity when making policy decisions. These associations are not deterministic—they reflect practices that can change. Looking forward, we plan to investigate additional health outcomes and refine causal analysis methods. By making our data publicly available, we hope to encourage further research and informed policymaking.

Boston University has a brain bank for participants exposed to repetitive head impacts. The vast majority of donors have been American football players, but we’ve also collected brains from ice hockey players, rugby players, and others. While we saw a clear dose-response relationship (more exposure increases health consequences) in football players, it was only recently that we had enough data to analyze ice hockey. This study is crucial to understanding how risks apply across different sports.

What surprised me was how strong the relationship was. The odds of developing CTE increased by about 34 percent for each additional year of play. We also saw significant links between years of play and tau protein accumulation in the brain, which correlated with dementia and impaired daily function. While enforcers—players who fight more often—had higher risks, the overall takeaway is that years of play are the primary driver of CTE risk.

I want people to understand that it’s not just about symptomatic concussions, it’s the accumulation of thousands of impacts over a career that increases risk. Policies that reduce impacts can make a big difference. I hope this study helps players and their families make informed decisions about their health while encouraging safer ways to play the sport.

We’re seeing an explosion of new data sources, like street-view imagery, being used to make decisions. But there’s often a disconnect—people using these tools don’t always have the public health knowledge to interpret the data correctly. We wanted to highlight the importance of combining data science and domain expertise to ensure interventions are accurate and impactful.

We discovered that the relationship between built environment features and health outcomes isn’t straightforward. It’s not just about having sidewalks; it’s about how often people are using them. Improving physical activity levels in a community could have a far greater impact on health outcomes than simply adding more infrastructure.

It also revealed the importance of understanding the local context. For instance, Google Street View data sometimes misclassifies sidewalks, particularly near highways or bridges, leading to inaccurate conclusions. Relying solely on this data, without accounting for these nuances, could result in less effective interventions.

I hope this research sparks conversations about integrating local context into data-driven decisions. By collaborating with urban planners, public health departments, and tech companies, we can ensure the data reflects reality and leads to more effective health interventions.

The purpose of this study was a survival analysis of people with end-stage kidney disease, when their kidneys are failing and they need dialysis.

I spent about two-and-a-half weeks on the Pine Ridge Reservation, working in the emergency room and also as a transplant coordinator. I see how challenging it is for them to get high-quality health care to keep their diabetes managed. Some of these people live five or six hours away from the transplant facility.

Compared to non-Hispanic white diabetes patients, American Indians and Alaska Natives survived longer on dialysis, with an 18.4 percent decrease in risk of death. But they are also diagnosed with end-stage disease much sooner. They progress rapidly through the stages of chronic kidney disease, coming to the emergency room because their kidneys are no longer working.

Once they have diabetes, there’s not a lot of good follow-up. I’d like to get chronic kidney disease patients into doctor visits before they even get to the end-stage, whether it’s in a local clinic or a nephrology hospital unit. I would also love to see more discussions about advanced care and end-of-life planning.

Adolescents who were exposed to cannabis in utero demonstrate a higher prevalence of mental health problems, including anxiety, depression, and attention deficit hyperactivity disorder. These findings underscore the importance of understanding how prenatal exposures can influence mental health outcomes later in life.

The study suggests that prenatal exposure to cannabis could disrupt normal brain development. These disruptions can lead to alterations in brain architecture—in regions associated with emotional regulation, attention, and impulse control—that appear as behavioral and psychological challenges during adolescence. For instance, significant changes were observed in the prefrontal cortex, which is critical for managing decision-making.

There is an urgent need for increased awareness about the risks of cannabis use during pregnancy. Public health campaigns and health care provider education could play crucial roles in encouraging expectant mothers to abstain from cannabis.

Activists, and others, have been asking for reparations for a very long time. It’s not a new conversation, but it’s one that public health hasn’t been very involved in, even though we know there is a strong correlation between higher socioeconomic status and better health outcomes. We wanted to think about the relationships between structural racism and health and what an intervention might look like to close the Black-White wealth gap.

We used a simulation to model reparations payments over ten years. We found that when we made payments totaling about $900,000 to each Black household, these reparations would have reduced premature mortality by about 30 percent for Black adults over the age of 18. We also found a similar relationship for overall mortality, not just those who died before the age of 65. There was an all-cause mortality reduction of about 25 percent in Black adults.

I’d like to see H.R. 40, a bill that would establish a commission to study and develop reparation proposals, actually come to the floor of Congress. It’s been introduced in Congress each year since 1989, but it’s never gone for a vote outside of committee. I feel encouraged, though, that these conversations are happening on a state level in some places, like in California, and in smaller cities and municipalities. Ultimately, I’d like us to spend more time thinking about what reparations might look like.

Dementia prevention is not just about memory but also about preserving the independence and dignity of our patients. Dementia affects not only their personal and family lives but also their caregivers. With the availability of disease-modifying drugs for Alzheimer’s and the opportunity for lifestyle modification, the sooner we can diagnose these patients, the sooner we can put them on a path to better brain health.

We found 62.3 percent of people over age 65 had undiagnosed mild cognitive impairment, 12.3 percent had dementia, and only 25.5 percent had no cognitive impairment. Black individuals had more than double the odds of White individuals of having a diagnosis. The prevalence of undiagnosed cognitive impairment in the community could be higher because patients who are plugged in to primary care are different from people who do not have access to primary care.

We need more targeted interventions for communities at high risk, so people are not diagnosed at late stages when the treatment options are limited. Health care providers can emphasize brain health in routine visits and offer culturally sensitive resources to reduce stigma and encourage people to seek screening. Some health care systems are piloting digital screening tools embedded in their electronic health systems, but more needs to be done.