Taking Appropriate Protective Measures During Pandemics

A new model determines the optimal strength of containment measures for infectious diseases, taking into account the costs of both the disease and containment. Courtesy of Max Planck Institute for Dynamics and Self-Organization (MPI-DS)

Aug. 20, 2026

In the event of a pandemic, every society is confronted with the question: How can the spread of an infectious disease be effectively contained without restricting daily life more than necessary? Measures such as mandatory mask-wearing or contact restrictions can reduce the spread of disease but also entail social, economic, and psychological costs.

Optimization of countermeasures
Researchers from the Theory of Complex Systems group at the Max Planck Institute for Dynamics and Self-Organization (MPI-DS) developed a model to optimize such countermeasures while taking containment costs into account. Using numerical methods, it calculates the optimal intensity of intervention based on the characteristics and severity of a disease. The aim of the study was not to make statements about specific diseases or to propose specific measures, but rather to identify general principles of optimal pandemic control. The flexible optimization model can also be combined with different disease models: “Through this, we hope to contribute to preparations for possible future outbreaks,” explains Laura Müller, the study’s first author.

Abrupt transition at optimal level of containment
The research group investigated general patterns of optimal infection control using a classical mathematical model in which individuals are either susceptible, infected, or temporarily immune following recovery. “Surprisingly, it becomes very clear that optimal measures follow a threshold structure,” explains Viola Priesemann, professor and group leader at the MPI-DS. “For mild diseases, the model shows that it is optimal not to impose any containment measures. However, once the disease reaches a certain severity, a high level of containment is optimal.”

Which combination of measures is chosen is a societal and political question and depends on the characteristics of the disease. “It is very surprising that the transition in the idealized model occurs absolutely abruptly,” Priesemann further emphasizes. Measures that represent a compromise between optimal intervention and no control at all ultimately result in higher overall costs: Either they are more extensive than necessary or they are insufficient to effectively curb the spread of infection.

Seasons and vaccinations influence the course of infection
The researchers also examined the influence of fluctuating infection rates over the course of the year. Their findings showed that optimal containment in winter must increase in tandem with the higher probability of infection. With such optimized containment of infectious diseases like influenza, there would be at most a small wave of infection in the spring instead of the typical waves of infection during the winter months. Mathematically, it can be calculated that this small wave occurs exactly 3 months after the peak of seasonality.

The new optimization framework also allows to determine how measures can optimally be reduced during vaccination campaigns.
In addition, it is possible to calculate the costs incurred when measures are implemented too late: Due to the exponential growth at the beginning of a pandemic, even minor delays in launching measures lead to notably higher infection rates. In the case of severe diseases, this results in considerable additional costs; in the case of mild diseases, however, it does not.

The study is based on a simplified mathematical model and therefore cannot make direct statements about individual diseases or specific pandemic situations. However, it has identified a novel, fundamental principle of optimal infection control: the clear threshold that emerges when containment costs are factored in at a very general level. Such principles provide guidance when society and policymakers need to develop strategies, even when the exact details of disease spread are not yet known. The results can thus help provide a stronger scientific basis for future decisions. Which measures are ultimately implemented, however, remains a societal and political decision.

Surveyed Americans Unsure Which Vaccines the CDC Recommends During Pregnancy

July 22, 2026

Most Americans are unaware or unsure of which vaccines the Centers for Disease Control and Prevention (CDC) recommends during pregnancy, according to a nationally representative survey of U.S. adults by the Annenberg Public Policy Center (APPC) of the University of Pennsylvania.

Almost half of those surveyed know that the CDC recommends getting the seasonal flu vaccine during pregnancy, while less than a third know that the CDC recommends the Tdap (tetanus, diphtheria, and pertussis), RSV (respiratory syncytial virus), and hepatitis B vaccines if one is not up to date on these vaccinations.

For the Covid-19 vaccine, over half of those surveyed do not know whether the CDC recommends it during pregnancy. While the CDC says that the vaccine helps reduce the risk for severe illness from Covid-19, it does not explicitly say the vaccine is recommended. Instead, the CDC says its Covid-19 vaccine recommendations are “now based on individual decision making, which emphasizes considering the benefits and risks of vaccination.”

Women of childbearing age (ages 18-49) are significantly more knowledgeable than other adult groups about most of the CDC’s pregnancy vaccine recommendations, although many in this group are also not sure what is recommended.

Uncertainty is a dominant finding in this survey. For each of four vaccines the CDC recommends during pregnancy, at least 45% of respondents are not sure whether the CDC recommends it.

The findings come from an APPC survey conducted on the research company SSRS’s Opinion Panel Omnibus platform among 1,031 U.S. adults from July 1-5, 2026. For further details, download the topline or see the end of this release.

“Our findings suggest that many Americans do not know which vaccines are recommended during pregnancy,” said Ken Winneg, APPC’s managing director of survey research. “The encouraging news is that relatively few people think that the CDC recommends vaccines that are not advised during pregnancy. The challenge is overcoming the widespread uncertainty over what the CDC does recommend.”

CDC recommendations during pregnancy
According to CDC guidance, four vaccines routinely recommended during pregnancy include:

Seasonal influenza vaccine during flu season.
Tdap vaccine during every pregnancy, preferably between 27 and 36 weeks’ gestation.
RSV vaccine during weeks 32-36, administered during RSV season, which is September through January in most of the continental United States. Additional doses are not recommended during subsequent pregnancies.
Hepatitis B vaccination for individuals who have not already been vaccinated.

The CDC says: “Covid-19 vaccination offers the greatest benefit if you are at higher risk for severe illness, including if you are pregnant. Pregnancy increases your risk of becoming very sick from Covid-19.” It adds that if you get sick with Covid-19 during pregnancy, you are at increased risk of complications that can affect your health and the health of your baby. The CDC recommends that Covid-19 vaccination decisions be made through individual decision making after considering the benefits and risks.

Other medical professionals more clearly back the Covid-19 vaccine during pregnancy. The American College of Obstetricians and Gynecologists (ACOG) strongly recommends that pregnant individuals be vaccinated against Covid-19 and “continues to recommend that all pregnant and lactating individuals receive an updated COVID-19 vaccine or ‘booster.’”

Awareness of recommended vaccines is low
Survey respondents were asked whether the CDC recommends each of eight vaccines during pregnancy. In addition to the five vaccines noted above, respondents were asked about three vaccines the CDC does not recommend during pregnancy: MMR or measles, mumps and rubella; chickenpox or varicella; and human papillomavirus or HPV.

Among U.S. adults overall, about a quarter to less than half of adults correctly identify any of the CDC-recommended vaccines, with the fewest (23%) knowing that the CDC recommends the hepatitis B vaccine during pregnancy and the most (48%) knowing the seasonal flu vaccine.

With the exception of RSV, women age 18 to 49 are significantly more likely than other groups (all adult men and women over 49 years old) to correctly identify all of the CDC-recommended vaccines. Among women 18 to 49, the vaccine least known to be recommended during pregnancy is the hepatitis B vaccine (30%) and the most known is the seasonal flu vaccine (60%).

Uncertainty dominates
For most of the vaccines, the most common response to whether the CDC recommends it during pregnancy is “not sure.” For each of the four recommended vaccines and the Covid-19 vaccine, from 45% to 66% of the overall adult population was not sure whether these were recommended. Among women of childbearing age, from 32% to 58% were not sure for each of the vaccines.

Few incorrectly think non-recommended vaccines should be taken
Current CDC guidance advises against administering the MMR and chickenpox vaccines during pregnancy because they are live-virus vaccines. The MMR and chickenpox vaccines should be given only before or after pregnancy. HPV vaccination is not recommended during pregnancy and should be delayed until after pregnancy if needed.

The survey found that relatively small groups of people incorrectly think that CDC recommends these vaccines:

21% say the CDC recommends the MMR vaccine during pregnancy.
17% say the CDC recommends human papillomavirus (HPV) vaccination.
17% say the CDC recommends chickenpox (varicella) vaccination.
Women of childbearing age are no more likely than others to incorrectly identify these vaccines as recommended. For the HPV vaccine, for example, 20% of women 18-49 years old incorrectly say it is recommended during pregnancy, compared with 16% of others, which is not a statistically significant difference.

For all three of these vaccines, the bigger issue is that a majority of respondents are not sure whether the CDC recommends that they should or should not be taken during pregnancy.

“Many Americans routinely get their immunizations from their primary healthcare providers or at local pharmacies, said Patrick E. Jamieson, director of APPC’s Annenberg Health and Risk Communication Institute, which oversees the health surveys. “It is critical that those professionals pay special attention to vaccine recommendations for recipients who are pregnant.”

Survey methodology
The data reported here come from a survey conducted for the Annenberg Public Policy Center by SSRS, an independent research company. The findings come from a nationally representative probability sample drawn from SSRS’s Opinion Panel, conducted July 1-5, 2026, among 1,031 U.S. adults. It has a margin of sampling error of ± 3.4 percentage points at the 95% confidence level. All figures are rounded to the nearest whole number and may not add to 100%. Combined subcategories may not add to totals in the topline and text due to rounding.

Source: Annenberg Public Policy Center (APPC) of the University of Pennsylvania

Why Taking a Sick Day Depends on More Than Being Sick

Courtesy of Adobe stock/makistock; provided by University of Technology Sydney

July 1, 2026

People may be too sick to work, but their job is too insecure to stay home. New research led by the University of Technology Sydney (UTS) shows the decision is heavily influenced by pay, job security and gender.

The study, published in Applied Economics, examines how workers’ health and economic circumstances dictate how many sick days they actually take.

“Employers and policymakers often focus on reducing absence, but workers who attend while unwell may recover more slowly, spread infection to colleagues, and be less productive,” said lead researcher Dr. Nancy Kong, a senior research fellow at the UTS Centre for Health Economics Research & Evaluation.

Drawing on data from the Household, Income and Labor Dynamics in Australia (HILDA) survey from more than 15,000 Australians between 2005 and 2016,  Kong and her co-authors, Dr. David Rowell from the University of Queensland and professor Peter Zweifel from the University of Zurich, examined patterns of sick leave across the workforce.

“We focused on this period to avoid the COVID years, when major changes in public health rules, workplace practices and leave policies occurred at the same time, and could have blurred the relationship between job conditions and sick leave,” said Kong.

The study revealed a clear divide: workers in casual and fixed term jobs take only around one day of sick leave a year on average, compared with about four days for permanent employees.

Even accounting for variables such as occupation, job satisfaction, household circumstances, living arrangements, marital status, education and place of residence, non-permanent workers still take around three fewer sick days each year.

“This does not necessarily mean casual and fixed-term workers are healthier,” said Kong. “A more likely explanation is that taking time off is riskier when work is insecure. Non-permanent workers may have less access to paid sick leave. They may also worry that saying no to work, even when ill, could affect future hours or their chances of keeping their job. For a permanent employee, staying home with influenza might be inconvenient; but for a casual worker it may trigger financial stress.”

The study also found that economic insecurity plays a role, with workers living in areas with higher unemployment tending to take less sick leave.

For instance, when the local unemployment rate rises by five percentage points, sick leave drops. While this amounts to a fraction of a day per individual, across a standard team this adds up to significant forgone recovery time.

This pattern is consistent with a simple concern: when jobs feel harder to replace, workers may be less willing to take time off.

“They may worry that being absent could make them seem less reliable or increase the risk of losing work,” said Kong.

The effect of wages proved more nuanced. Higher wages alone did not consistently dictate leave behavior. However, among workers in poorer health, higher wages were strongly associated with taking more sick leave, suggesting higher income earners possess a financial buffer.

“Higher-paid workers generally have workplace support and leave entitlements that mean they are supported to take time off when they are unwell without fear of a potential financial penalty,” said Kong.

The most consistent finding was a distinct gender gap. Across every analysis, men take fewer sick days than women, averaging about half a day less per year (a 23% difference), even when matching with similar health and job circumstances.

“This may reflect differences in health needs, caring responsibilities or how likely people are to seek medical care,” Kong said. “But it also points to workplace cultures and gender expectations about ‘toughness’, reliability and working through illness.”

Ultimately, the study highlights that sick leave rates are not simply a reflection of physical health. For employers, low sick leave rates should not be automatically viewed as a sign of success; they may also indicate a culture of fear.

“Workplace cultures should not reward people for attending when unwell or treat legitimate sick leave as a lack of commitment,” said Kong. “Reducing stigma is also particularly important in addressing the gender gap.”

Kong said for policymakers, the study points to the importance of secure work, accessible paid leave and workplace practices that support people to recover when unwell. This is particularly relevant during periods of increased cost-of-living pressure, workforce shortages and seasonal illness. “A fair and effective sick leave system should support productivity while ensuring workers do not have to choose between protecting their health and protecting their income.”

Source: University of Technology Sydney

Estimated Effectiveness of 2024-2025 COVID-19 Vaccines in Adults

Licensed via Adobe Stock

June 15, 2026

Vaccine effectiveness (VE) estimates are needed to monitor the effect of updated COVID-19 vaccinations, say Wiegand, et al. (2026) who sought to assess the effectiveness of 2024-2025 COVID-19 vaccines against medically attended COVID-19 among adults 18 years and older in the U.S. This case-control study with a test-negative design included patient encounters with a COVID-19–like illness discharge diagnosis code and a molecular or antigen SARS-CoV-2 test within 10 days before to 3 days after the encounter date, from Sept. 5, 2024, to Sept. 2, 2025. Encounters were captured in VISION (Virtual SARS-CoV-2, Influenza, and Other Respiratory Viruses Network), a multisite, electronic medical record–based network of health care systems, including 381 emergency department/urgent care (ED/UC) departments and 246 hospitals in six states.

In 333,262 eligible ED/UC encounters (median [IQR] age of patients, 54 [35-72] years; 60% female) and 97,663 eligible hospitalizations among immunocompetent adults 18 years and older (median [IQR] age of patients, 72 [59-81] years; 53% female), estimated VE was 26% (95% CI, 23%-29%) against COVID-19–associated ED/UC encounters, 35% (95% CI, 30%-40%) against COVID-19–associated hospitalization, and 41% (95% CI, 28%-51%) against COVID-19–associated critical illness 7 to 299 days after vaccination. Among immunocompetent adults 65 years and older (122 663 ED/UC encounters and 63 958 hospitalizations), estimated VE was 26% (95% CI, 22%-30%) against COVID-19–associated ED/UC encounters, 35% (95% CI, 29%-40%) against COVID-19–associated hospitalization, and 41% (95% CI, 28%-52%) against COVID-19–associated critical illness 7 to 299 days after vaccination. Among 32,629 hospitalizations in immunocompromised adults 18 years and older, estimated VE against COVID-19–associated hospitalization was 24% (95% CI, 13%-34%). VE estimates waned with more time since vaccination.

In this test-negative case-control study, 2024-2025 COVID-19 vaccination was associated with reduced likelihood of medically attended COVID-19–associated outcomes among immunocompetent and immunocompromised adults, highlighting the importance of adults receiving recommended COVID-19 vaccinations.

Reference: Wiegand RE, et al. Estimated Effectiveness of 2024-2025 COVID-19 Vaccines in Adults JAMA Intern Med. Published Online: June 15, 2026. doi: 10.1001/jamainternmed.2026.1936

Testing AI Against Public Health’s Existing Tools

Courtesy of Sylvia Zhang, Penn Engineering

June 8, 2026

A new University of Pennsylvania-led randomized controlled trial has found that AI-powered chatbots can make vaccine-hesitant parents more likely to say they will immunize their children against human papillomavirus (HPV), but no more than standard written public health materials.

The findings raise questions about when, how and to what extent AI enhances public health communications. “Comparing a chatbot to nothing isn’t really a fair test. The interesting question is whether it does better than what public health agencies already have out there. In our study, it didn’t,” says Sharath Chandra Guntuku, research associate professor in computer and information science (CIS) and the study’s senior author.

Described in a new paper in JAMA Network Open, the trial — which included nearly 1,300 participants in the United States, the United Kingdom and Canada — found that skeptical parents who interacted with the chatbots were more likely than those who received no intervention to say they intended to immunize their children.

But spending a few minutes reading standard written materials provided online about the benefits of the HPV vaccine from governmental health agencies like the Centers for Disease Control and Prevention (CDC) produced essentially the same effect.

“It would have been easy to compare an AI chatbot with no intervention or a very weak control condition and find a positive result,” says Neil Sehgal, a doctoral student in CIS and the study’s first author. “But we wanted to know whether the chatbot added value beyond what public health agencies already provide.”

As chatbots have become more powerful and widely used, researchers around the world have begun to explore the extent to which AI can change people’s minds, a question made all the more urgent by the rising tide of vaccine hesitancy.

Two years ago, one study found that an AI-powered chatbot could reduce beliefs in conspiracy theories. Last year, a randomized controlled trial in China found that giving parents access to a chatbot improved the odds that they either vaccinated their daughters against HPV or scheduled an appointment to do so.

But the designs of those studies and others like them often make it difficult to assess the effectiveness of the chatbots. In the Chinese study, for instance, two weeks of access to the chatbot was compared with no intervention, making it unclear if other persuasive content, like standard public health materials, would have had a similar effect.

“What is new here is the comparison against a strong, realistic control,” says Alison M. Buttenheim, professor of nursing and health policy in family and community health and a co-author of the study.

In contrast, the Penn team set strict guidelines to make the chatbot as comparable as possible with existing public health materials, and to assess its effects over time: participants assigned to the chatbot or written materials groups were exposed to each for the same minimum time of 3 minutes, and all participants were assessed for their intention to vaccinate their children 15 and 45 days after the interventions.

Under these stricter conditions, the innovative features of the chatbots — including their ability to converse with parents in real time — essentially provided no additional benefit, even though parents spent more time engaging with the chatbots than the written materials.

“While chatbot conversations can move intentions immediately,” says Lyle Ungar, Professor in CIS and a co-author of the study, “their advantage disappeared when compared with well-designed public health materials and did not persist over time.”

Indeed, at 45 days, participants who had been assigned to read the public health materials expressed a higher intent to vaccinate their children than either those who interacted with the chatbots or received no intervention.

“AI chatbots are promising, but they should not be assumed to outperform existing tools simply because they are newer or more interactive,” adds Guntuku. “A short read of a CDC webpage held up at least as well as a chatbot conversation, and the effect actually lasted longer.”

The researchers caution that the minimum exposure time to each intervention may overstate the real-world strength of such written public health materials. “Would all HPV-vaccine-hesitant parents choose to spend a full 3 minutes on the CDC webpage?” asks Sehgal. “Maybe, maybe not; chatbots are certainly more interactive.”

In the end, none of the interventions — neither the written materials nor the chatbots — increased the share of parents who said their children had actually received the HPV vaccine within the 45-day window of the study, underscoring the challenge of addressing vaccine hesitancy.

It’s possible, the researchers say, that the study was simply too short and the interventions too focused on communication, rather than structural barriers to medical care like time, money and access to a pediatrician, for participants to vaccinate their children in greater numbers.

“Vaccination is more than a communication problem,” says Buttenheim. “Even if the interventions changed some parents’ minds, it takes a lot to convert intention into action.”

Next, the researchers hope to test the use of AI in more complex scenarios, where chatbots function less as one-off conversational partners than medical concierges, helping schedule appointments, send reminders and liaise with clinicians. “A chatbot might be more useful if it’s paired with other functions,” notes Ungar.

The Penn team is also extending this work to global health settings, including studies of AI-supported vaccine communication in Nigeria. The goal is to understand how chatbot interventions can be adapted to local contexts rather than simply exported from studies conducted in the United States, the United Kingdom and Canada. “For AI tools to be useful in public health, they have to be evaluated in the communities where they might actually be deployed,” says Sehgal. “That means working with partners to understand concerns, language, trust and access, not just translating a chatbot prompt.”

Ultimately, the researchers hope their work encourages a more evidence-based approach to using AI for public health. “We need to evaluate AI tools against realistic alternatives,” says Guntuku. “It’s time to shift the conversation from, ‘Can AI persuade people?’ to more granular questions, like ‘When does AI add meaningful value, for whom and under what conditions?’”

This study was conducted at the University of Pennsylvania’s School of Engineering and Applied Science (Penn Engineering), School of Nursing (Penn Nursing), Leonard Davis Institute of Health Economics (LDI), Perelman School of Medicine (PSOM), Center for Health Incentives and Behavioral Economics and the Annenberg School for Communication (ASC), and was supported by the Penn Medical Communication Research Institute, Penn Global Research and Engagement Fund and the National Institutes of Health (NIH-NIMHD:R01MD018340, NIH-NIMH:R01MH132401 and NIH-NCI:R37CA259210).

Additional co-authors include Sunny Rai of Penn Engineering and LDI; Manuel Tonneau of Oxford, the World Bank and NYU; Anish Agarwal of PSOM; Joseph Cappella of LDI and ASC; and Melanie Kornides of LDI and Penn Nursing.

Source: University of Pennsylvania