番茄社区

In a study sure to provoke controversy, researchers analyzing data from a large, standard federal survey indicated a small but detectable surge in mental health problems among African Americans in the weeks following publicized killings of unarmed blacks by police.

With data from the regular Behavioral Risk Factors Surveillance System (BRFSS) during 2013-2015, the researchers established that each police killing of an unarmed African American was associated with 0.14 additional "poor mental health days" in the previous month among blacks living in the same state (95% CI 0.07-0.22).

Such effects were most prominent in the first 1-2 months after such killings, reported Jacob Bor, SD, of Boston University, and colleagues .

Although 0.14 days may appear to be a small, even negligible effect, Bor and colleagues noted that it's nearly as great as the mental health burden associated with diabetes.

BRFSS data for the study covered 103,710 African American respondents, of whom 39,993 were exposed to at least one police killing of an unarmed black person in their state in the 3 months prior to participating in the BRFSS survey. That survey is conducted nationwide on a permanent, rolling basis. Among the survey questions is one that asks participants to evaluate their own mental health status. Responses of "not good" were the current study's primary outcome measure. Among all African American respondents in the 2013-2015 data, the mean was 4.1 days of "not good" mental health during the previous month.

Data on police killings came primarily from the , which in turn draws on other non-authoritative sources such as with follow-up fact-checking.

"We note that our analysis defines as exposed any person living in the state where the police killing occurred, including some people who might not have been substantively exposed to the police killing (or even aware of it)," Bor and colleagues noted. "Our causal estimates therefore have an intention-to-treat interpretation, similar to a randomized trial with non-compliance to treatment assignment."

As a result, they said, their analysis "probably underestimate[s] the effect of police killings on the mental health of subpopulations most directly exposed."

Analyses of mental-health status responses from white respondents showed no relationship to previous killings of unarmed blacks. Also, no association between mental health status of white or black participants and reported killings of armed African Americans was seen. Finally, stratifying the number of killings of unarmed blacks to which respondents might have been exposed -- analyzed as one versus two or more -- did not show what could be termed a dose response.

Although these analyses were based on survey data and nominally only showed associations, Bor and colleagues argued their "quasi-experimental methodology ... allows for a causal interpretation of these estimates."

That interpretation, of course, was that police killings of unarmed blacks is detrimental to mental health in the broader African American population. In particular, the data imply that such killings "could contribute 1.7 additional poor mental health days per person per year," or 55 million poor mental health days overall in the population. A similar calculation using previous studies of diabetes' mental health burden suggests that condition adds 75 million poor mental health days per year among black Americans.

Bor and colleagues also said that their methodology may underestimate the effect: many of the killings in 2013-2015 were nationally publicized and thus would presumably affect African Americans throughout the country, but the analysis only included respondents in the same states where the killings occurred.

In addition, the researchers argued, they could have missed some police killings that should have been included, because many such killings go unrecorded in public databases. Also, race/ethnicity of victims was unknown for 7% of those they did consider for inclusion.

"Our results point to the importance of structural racism as a driver of population health disparities," Bor and colleagues concluded, adding that their results "provide rare causal evidence" about the impact of police killings on the black community.

The researchers did, however, admit to a number of limitations, including lack of information on how much BRFSS participants actually knew about police killings, the self-reports of mental health status, and the possibility of unmeasured confounders.

Comment