Short report
Hispanic victims being killed at higher rates than White victims and contributing significantly more YLL (figure 1 and table 1). The age distribution among unarmed victims was younger than the overall victim pool (table 1). DISCUSSION By capturing the persistent and longitudinal trend of dispropor- tionate mortality and YLL due to fatal police shootings among Black, Hispanic and Native American populations over the last 5 years, this study extends previous work that demonstrated this inequity based on narrower time windows. 1 2 Further, this study is novel in its demonstration that mortality and YLL burden for BIPOC have not decreased, and that it exists even among unarmed victims. This study is not without limitations. Our analysis relies upon news reports curated by the Washington Post , necessa- rily excluding incidents that did not attract media attention. However, previous studies have shown that media data on fatal police encounters are more complete than centralised federal databases. 7 Second, our data did not capture gender minority status, which may obscure an intersectional mod- erator of police violence. Also, serious injury can occur with- out firearms as was the case with George Floyd, who died of asphyxiation, and Emerald Black, a visibly pregnant woman whose violent arrest reportedly lead her to miscarry in police custody. 8 Lastly, we excluded missing data from our analyses which could potentially bias our results. However, given the strength of the association measures observed here, it is unlikely that the severe inequities in fatal police shootings among Black and Indigenous People of Colour can be com- pletely explained by potential bias attributable to missing data. We found an annual average of 31 960 YLL due to fatal police shootings in the USA, which is 83% of that due to
This study uses publicly available data and was therefore exempt from IRB approval. RESULTS There were 5367 fatal police shootings reported by the Washington Post from 2015 to May 2020. A total of 627 were excluded from the analysis because race was unknown (n=580) or marked as ‘ other ’ (n=47), leaving a total of 4740 deaths for analysis. Of those, age was not reported for 87 victims, leaving 4653 deaths included in the YLL analysis. Most victims were White (51%), followed by Black (26.7%), Hispanic (18.8%), Asian (2%) and Native American (1.6%). The median age of victims was 34 and varied across groups with younger Black victims (median age=30) and older White victims (median age=38) (table 1). There was a small, but statistically significant decline in death rate for White victims with a rate ratio of 0.99 (95% CI 0.98 to 0.995), indicating an estimated 1% decline in deaths based on the GLMM. For all other race/ethnic groups there were no significant trends in death rates. No group showed a significant trend in YLL (figure 1). By group, mean deaths pqpm were highest among Native Americans (1.74 pqpm), followed by Blacks (1.49 pqpm), Hispanics (0.74 pqpm), Whites (0.57 pqpm) and Asians (0.25 pqpm) (table 1). Native Americans, Blacks and Hispanics had significantly higher rates than Whites with RRs of 3.06 (95% CI 2.42 to 3.86), 2.62 (95% CI 2.41 to 2.86) and 1.29 (95% CI 1.18 to 1.42) respectively, while Asians had a death rate significantly lower than Whites (RR=0.44, 95% CI 0.36 to 0.55) (table 1). Similar relative trends were noted for YLL (figure 1 and table 1). A subset of 753 (15.9%) fatal police shootings had unarmed victims. There were no significant trends in deaths or YLL for any group. Relative trends by race/ethnicity in the death and YLL rates were consistent with the overall analysis, with Black and
Table1 Fatal police shootings by race/ethnicity, 2015 – 2020 All fatal police shootings Deaths
Years of life lost
Age
Mean quarterly rate per million
Rate ratio [95%CI]
Mean quarterly rate per million
Rate ratio [95%CI]
Race/ethnicity N [%]
N [%]
Median [IQR]
Asian Black
93 [2.0]
0.25 1.49 0.74 1.74
0.44 [0.36, 0.55] 2.62 [2.41, 2.86] 1.29 [1.18, 1.42] 3.06 [2.42, 3.86]
3408 [2.0]
9.09 61.5
0.49 [0.38, 0.62] 35.0 [27.5 – 44.5] 3.29 [2.59, 4.17] 30.0 [24.0 – 38.0] 1.55 [1.22, 1.97] 33.0 [25.0 – 40.0] 3.95 [3.11, 5.02] 31.0 [25.0 – 36.0]
1265 [26.7] 889 [18.8]
52 186 [30.1] 35 098 [20.3]
Hispanic Native American
29.01 73.75
77 [1.6]
3266 [1.9]
White
2416 [51.0]
0.57 0.71
1.00 [Reference]
79 158 [45.7]
18.72 25.84
1.00 [Reference] 38.0 [30.0 – 50.0]
All
4740
173 116
34.0 [27.0 – 45.0]
–
–
Fatal shootings with unarmed victim Deaths
Years of life lost
Age
Mean biennial rate per million
Rate ratio [95%CI]
Mean biennial rate per million
Rate ratio [95%CI]
Race/ethnicity N [%]
N [%]
Median [IQR] 30.0 [25.2 – 43.0]
Asian*
11 [1.5]
410 [1.4]
–
–
–
–
Black
218 [29.7] 146 [19.9]
0.52 0.24
3.18 [2.63, 3.85] 1.45 [1.17, 1.79]
9107 [31.3] 5930 [20.4]
22.00
3.49 [2.74, 4.45] 30.0 [23.0 – 37.0] 1.55 [1.21, 1.98] 32.0 [24.0 – 40.0]
Hispanic
9.76
Native American*
8 [1.1]
346 [1.2]
29.5 [25.0 – 37.5]
–
–
–
–
White
352 [47.9]
0.16 0.21
1.00 [Reference]
13 306 [45.7]
6.29 8.51
1.00 [Reference] 34.0 [26.0 – 44.0]
All
735
29 099
32.0 [25.0 – 41.0]
–
–
All rates are crude estimates and all rate ratios are estimated using negative binomial generalised linear-mixed models. *Asian and Native American victims were excluded from the unarmed trend analysis due to low counts.
Lett E, et al. J Epidemiol Community Health 2020; 0 :1 – 4. doi:10.1136/jech-2020-215097
2
Powered by FlippingBook