Image: Polly Becker

In 2013, Summer Hawkins had 17.7 million answers to ponder. The assistant professor in the School of Social Work (SSW), in her second year at Boston College, was investigating the effects of cigarette taxes on the rate of maternal smoking during pregnancy. She had obtained the digital records of every birth certificate issued in 28 states from 2000 to 2010, a growth period for anti-smoking activism. From Alaska to West Virginia, all 28 had employed the U.S. Standard Certificate of Live Birth, which included a parent worksheet asking postpartum women about their cigarette consumption. “This wasn’t a sample,” she says. “It was the population,” the kind of big data that allows social epidemiologists like Hawkins to take in the myriad, sometimes hidden impacts of a government health policy. Now, she needed someone with the mathematical nimbleness to “tease apart” the responses and align them with additional state data.

Christopher “Kit” Baum, a Boston College economics professor of some 40 years standing, whose specialty is econometrics, had for some time been applying statistical analysis to questions such as the “uncertainty determinants of corporate liquidity” (for which he and a colleague considered 10 years of a Standard & Poor’s database of U.S. companies, 201,552 entries in all). Hawkins proposed Baum take up her project.

Baum applied a model called difference-in-differences to decipher the effect that a range of cigarette tax rates had on various cohorts of mothers. (After 11 years of general increases, the taxes in 2010 spanned from Missouri’s $0.17 a pack to Rhode Island’s $3.46.) Tax levels, it was found, had almost no effect on college-educated pregnant women, who smoked at a rate of 1.5 percent. But expectant mothers lacking a high-school diploma smoked 14–22 fewer cigarettes a month for every $1 cigarette-tax increase, constituting an average decrease of 5.3 percent. “We have demonstrated,” Hawkins and Baum wrote in the August 2014 American Journal of Public Health, “that cigarette taxes may be an effective . . . intervention to reduce smoking among the highest risk mothers.”

A subsequent, 2015, study conducted by the pair looked at the effect of cigarette taxes on teenagers. It drew on the responses of 717,543 adolescents to the Youth Risk Behavior Surveys taken in 43 states by the Centers for Disease Control, 1999–2013. “Overall,” Hawkins and Baum noted, adolescents were “72 percent less likely to smoke in 2013 than in 1999.” But among the facts they gleaned: Tax increases, which were seemingly ineffective with older teens, had a significant impact on teens ages 14 and 15, who showed a 2.2 and 1.6 percent decline in smoking, respectively, for every $1 in added cost.

In each of their eight collaborative studies since 2014, Hawkins, rangy and ebullient, posits the questions—e.g., How does banning smoking in public places affect children’s health? She then collects colossal sets of data from, say, hospitals, health insurers, and state agencies over key time spans. Baum, pensive and imposing, develops econometric models, often aided by his doctoral students, and then runs the data through a high-octane server in St. Clement’s Hall. In 2013, he received a courtesy appointment as an SSW professor for his ongoing research contributions.

Last year, Preventive Medicine published Hawkins and Baum’s answer to that question about smoking bans and children. Through an analysis of 828,000 patient reports generated by 114 hospital emergency departments in Massachusetts, New Hampshire, and Vermont, they determined, initially, that restaurant smoking bans have no apparent effect on the health of children ages 0–17. They went on to slice the data more finely by age, however, and found sharp reductions associated with bans among 10-to-17-year-olds—who likely spend more time in public spaces—in treatments for asthma (-12 percent), respiratory infections (-9 percent), and ear infections (-8 percent).

Baum and Hawkins continue to expand their scope. They’ve acquired all death certificates issued in the country between 2005 and 2015 to investigate which state-level policies affect the maternal mortality rate; the United States is one of the few countries in the world where the rate has recently climbed. And in January the two received a three-year, $716,000 grant from the American Cancer Society, to study the 2010 Affordable Care Act’s impact on six preventive services, including breast and cervical cancer screenings, in Maine, Massachusetts, and New Hampshire. The data will extend into the hundreds of millions.

Read more by Zachary Jason