قانون خطر معکوس: فشار خون، آزار و اذیت جنسی، تبعیض نژادی، آزار در محل کار و مواجهات شغلی سیاه و سفید کم درآمد در آمریکا، کارگران سفید و لاتین

Abstract Research on societal determinants of health suggests the existence of an “inverse hazard law,” which we define as: “The accumulation of health hazards tends to vary inversely with the power and resources of the populations affected.” Yet, little empirical research has systematically investigated this topic, including in relation to workplace exposures. We accordingly designed the United for Health study (Greater Boston Area, Massachusetts, 2003–2004) to investigate the joint distribution and health implications of workplace occupational hazards (dust, fumes, chemical, noise, ergonomic strain) and social hazards (racial discrimination, sexual harassment, workplace abuse). Focusing on blood pressure as our health outcome, we found that among the 1202 low-income multi-racial/ethnic working class participants in our cohort – of whom 40% lived below the US poverty line – 79% reported exposure to at least one social hazard and 82% to at least one high-exposure occupational hazard. Only sexual harassment, the least common social hazard, was associated with elevated systolic blood pressure (SBP) among the women workers. By contrast, no statistically significant associations were detectable between the other additional highly prevalent social and occupational hazards and SBP; we did, however, find suggestive evidence of an association between SBP and response to unfair treatment, implying that in a context of high exposure, differential susceptibility to the exposure matters. These results interestingly contrast to our prior findings for this same cohort, in which we found associations between self-reported experiences of racial discrimination and two other health outcomes: psychological distress and cigarette smoking. Likely explanations for these contrasting findings include: (a) the differential etiologic periods and pathways involving somatic health, mental health, and health behaviors, and (b) the high prevalence of adverse exposures, limiting the ability to detect significant associations. As clarified by the “inverse hazard law,” to understand health inequities, research is needed that contrasts exposures and health status population-wide, not just among those most inequitably exposed.

Results Table 1 shows the distribution of the outcome and selected exposures and key covariates among the 1202 members of the United for Health cohort, overall and by race/ethnicity, prior to imputing missing values. As presented in more detail in Supplemental Table 1 and in our previous papers ( Barbeau et al., 2007, Krieger et al., 2005 and Krieger et al., 2006), the findings reveal high levels of economic deprivation and of exposure to social and occupational hazards (79% to at least 1 social hazard, 82% to at least 1 occupational hazard) among this cohort of predominantly low-income multi-racial/ethnic working class women and men, of whom fully 40% were below the US poverty line (24% of whites vs. 40–50% among the black, Latino, and other workers of color). Among all workers, the average systolic blood pressure equaled 134.8 mm Hg and about 21% of the workers were hypertensive (ranging from 6% among Latinas to 28% among white men), with a similar proportion taking medication for hypertension. Table 2 presents Pearson correlations between the social and occupational hazards. Statistically significant correlations (p < 0.05) were especially evident between: (a) racial discrimination and workplace abuse, among the workers of color; (b) racial discrimination and sexual harassment, among the women workers; and (c) workplace abuse and high exposure to occupational hazards, among all workers. Only among black men was the correlation significant between high exposure to occupational hazards and racial discrimination. In all cases, the significant correlations ranged between 0.2 and 0.5 (indicating an association, but not co-linearity), with correlations above 0.35 occurring only among workers of color. Table 2. Pearson correlations between the social and occupational hazards,a by racial/ethnic-gender group: ρ (p-value*), United for Health study (Boston, MA, 2003–2004) Hazard Race/ethnicity Gender Workplace abuse Sexual harassment Racial discrimination High exposure to occupational hazards Workplace abuse Black Men 1.00 Women 1.00 Latino Men 1.00 Women 1.00 White Men 1.00 Women 1.00 Additional race/ethnicity Men 1.00 Women 1.00 Sexual harassment Black Men 0.17 (<0.01) 1.00 Women 0.24 (<0.01) 1.00 Latino Men 0.07 (0.43) 1.00 Women 0.26 (0.01) 1.00 White Men 0.06 (0.43) 1.00 Women 0.12 (0. 23) 1.00 Additional race/ethnicity Men −0.07 (0.56) 1.00 Women 0.05 (0.77) 1.00 Racial discrimination Black Men 0.36 (<0.01) 0.07 (0.24) 1.00 Women 0.38 (<0.01) 0.28 (<0.01) 1.00 Latino Men 0.44 (<0.01) 0.10 (0.24) 1.00 Women 0.49 (<0.01) 0.39 (<0.01) 1.00 White Men 0.20 (0.01) 0.28 (<0.01) 1.00 Women 0.08 (0.44) 0.35 (<0.01) 1.00 Additional race/ethnicity Men 0.52 (<0.01) 0.32 (0.01) 1.00 Women 0.46 (<0.01) 0.33 (0.03) 1.00 High exposure to occupational hazards Black Men 0.26 (<0.01) 0.05 (0.40) 0.35 (<0.01) 1.00 Women 0.19 (0.02) 0.03 (0.75) 0.13 (0.12) 1.00 Latino Men 0.27 (<0.01) −0.02 (0.77) 0.16 (0.04) 1.00 Women 0.22 (0.02) 0.00 (0.96) 0.13 (0.15) 1.00 White Men 0.29 (<0.01) −0.09 (0.24) 0.11 (0.16) 1.00 Women 0.07 (0.51) 0.15 (0.14) 0.11 (0.27) 1.00 Additional race/ethnicity Men 0.38 (<0.01) 0.04 (0.76) 0.22 (0.06) 1.00 Women 0.30 (0.05) 0.10 (0.51) 0.24 (0.11) 1.00 *Correlations with p < 0.05 noted in bold. a The correlations are based on the summary score for each of the social hazards and the composite high-exposure occupational hazard score. Table options Table 3 presents the analytic results for associations between SBP and the specified variables, overall and by race/ethnicity. Key findings for the total population are as follows. First, as expected, adjusting for covariates notably altered the bivariate results, including slightly increasing the difference in SBP among blacks compared to whites (by 1.3 points), further decreasing the SBP of women compared to men (by 2.5 points), and reducing the Latinos' initially statistically significant 5.3 point lower SBP compared to whites to 1.0, a non-significant difference. Covariates significantly associated with elevated SBP in the multivariable analyses thus included: race/ethnicity (but only for black vs. white comparisons (3.5 mm Hg, 95% CI 0.3, 6.7)); gender (men higher than women by 8.8 mm Hg, 95% CI 5.7, 11.8); age (by year, 0.4 mm Hg, 95% CI 0.3, 0.5); BMI (by unit, 0.7, 95% CI 0.5, 0.8); hypertension medication (7.4 mm Hg, 95% CI 4.7, 10.0); and alcohol (5.1 mm Hg; 95% CI 1.9, 8.4). The interaction of sexual harassment and gender was statistically significant (higher for women, by 2.8 mm Hg, 95% CI 0.6, 5.0). A protective effect for workers in transportation compared to retail (by 3.4 mm Hg, 95% CI 0.5, 6.4) was also evident. Weaker associations with elevated SBP occurred for lower vs. higher education, and for responding to unfair treatment by accepting it as a fact of life and keeping it to oneself, vs. taking action and talking to others. No significant associations were evident for poverty, racial discrimination, workplace abuse, occupational hazards, social desirability, nativity, job control, co-worker gender, or co-worker race/ethnicity. Given the lack of a main effect for the occupational hazards, we did not model the interaction of social and occupational hazards on SBP. Table 3. Association of systolic blood pressure with social and occupational hazards: United for Health study (Boston, MA, 2003–2004) Variable Systolic blood pressure (difference in mm Hg): View the MathML sourceβˆ parameter estimate (95% confidence interval) Total population (N = 1202) White (N = 300) Black (N = 493) Latino (N = 286) Bivariate Multivariablea Multivariablea Multivariablea Multivariablea Race/ethnicity White (referent [ref]) Black 2.20 (−0.36, 4.76) 3.52 (0.34, 6.23) Latino −5.29 (−8.37, −2.22) −1.05 (−4.57, 2.47) Other 0.30 (−3.45, 4.04) 3.08 (−0.86, 7.03) Gender Men (ref) Women −6.29 (−8.39, −4.19) −8.75 (−11.78, −5.71) −8.14 (−14.54, −1.74) −8.76 (−13.91, −3.62) −6.92 (−13.14, −0.70) Poverty No (≥100% poverty) (ref) Yes (<100% poverty) 1.42 (−0.79, 3.63) −0.68 (−2.81, 1.45) 0.80 (−3.64, 5.23) −0.43 (−3.72, 2.87) 0.62 (−4.11, 5.35) Education <High school (HS) 1.36 (−1.18, 3.90) 1.83 (−0.56, 4.21) 5.53 (0.83, 10.22) 1.09 (−2.89, 5.07) 0.27 (−4.49, 5.02) ≥HS, <4 yrs college (ref) 4+ yrs college −0.43 (−3.94, 3.09) 0.12 (−3.10, 3.35) −2.09 (−7.94, 3.75) 1.05 (−4.35, 6.44) −0.08 (−7.75, 7.59) Sexual harassment 0.22 (−0.91, 1.36) −0.46 (−1.78, 0.87) −5.35 (−8.86, −1.85) 0.20 (−1.87, 2.28) 1.14 (−2.22, 4.50) Sexual harassment × gender 2.75 (0.56, 4.95) 6.30 (0.81, 11.79) 2.50 (−1.02, 6.01) 2.55 (−2.47, 7.57) Racial discrimination never (ref) 1–2 exposures 1.69 (−0.94, 4.32) 1.48 (−1.00, 3.97) −0.61 (−4.85, 3.64) 3.82 (−0.63, 8.27) 1.49 (−3.56, 6.53) 3+ exposures 0.16 (−2.25, 2.57) −1.19 (−3.72, 1.35) 2.43 (−4.19, 9.06) 1.52 (−2.46, 5.50) −3.74 (−9.73, 2.26) Workplace abuse −0.13 (−0.38, 0.12) −0.16 (−0.47, 0.16) 0.16 (−0.39, 0.72) −0.39 (−0.93, 0.16) 0.07 (−0.71, 0.86) Workplace abuse × gender 0.24 (−0.26, 0.75) −0.01 (−1.01, 0.99) 0.43 (−0.43, 1.30) −0.32 (−1.47, 0.83) Occupational hazards 0 exposures (ref) 1–2 exposures −1.09 (−4.16, 1.97) −0.76 (−3.63, 2.10) 1.74 (−4.67, 8.14) −2.23 (−6.25, 1.79) 0.87 (−7.23, 8.96) 3–4 exposures −2.06 (−5.20, 1.09) −0.77 (−3.77, 2.23) 1.32 (−5.15, 7.79) −0.20 (−4.78, 4.38) 0.22 (−7.62, 8.06) 5+ exposures −2.58 (−6.30, 1.13) −0.72 (−4.38, 2.93) 0.98 (−6.13, 8.09) 1.28 (−4.94, 7.50) −4.00 (−12.53, 4.53) Age 0.51 (0.41, 0.61) 0.42 (0.31, 0.53) 0.26 (0.05, 0.47) 0.58 (0.38, 0.79) 0.35 (0.09, 0.61) Body mass index 0.63 (0.46, 0.80) 0.65 (0.49, 0.82) 0.88 (0.58, 1.18) 0.50 (0.24, 0.77) 0.58 (0.15, 1.01) Hypertension medicine No (ref) Yes 11.65 (9.07, 14.24) 7.35 (4.71, 9.98) 3.26 (−1.92, 8.44) 10.26 (6.36, 14.15) 8.85 (0.52, 17.17) Smoking Never (ref) Ex 3.21 (0.34, 6.07) −1.48 (−4.28, 1.31) 0.10 (−5.00, 5.20) −0.82 (−5.57, 3.93) 1.08 (−6.30, 8.47) Current −0.38 (−2.89, 2.13) −0.42 (−2.90, 2.07) −0.97 (−4.91, 3.36) 0.51 (−4.04, 5.05) −3.81 (−9.83, 2.21) Alcohol No (not in past 24 h) (ref) Yes (in past 24 h) 4.82 (1.50, 8.14) 5.14 (1.91, 8.37) 4.02 (−0.36, 8.40) 2.18 (−3.99, 8.36) 7.60 (−2.09, 17.28) Social desirability 0.01 (−0.02, 0.04) 0.01 (−0.03, 0.04) 0.08 (0.02, 0.14) −0.03 (−0.08, 0.03) 0.03 (−0.04, 0.11) Nativity US born (ref) Foreign born −1.39 (−3.41, 0.63) −0.29 (−2.19, 2.76) 0.06 (−5.81, 5.94) 2.24 (−2.10, 6.57) −1.41 (−6.55, 3.73) Job control High (ref) Low 0.36 (−1.90, 2.61) 0.91 (−1.20, 3.02) −0.72 (−4.48, 3.05) 1.18 (−2.33, 4.69) 1.94 (−3.62, 7.49) Workplace Retail (ref) Manufacturing −1.61 (−4.16, 0.95) −0.53 (−3.39, 2.34) 0.77 (−3.48, 5.02) −3.16 (−13.91, 7.59) 2.96 (−2.70, 8.61) Transportation 4.11 (1.72, 6.49) −3.43 (−6.37, −0.48) 2.52 (−3.57, 8.61) −7.85 (−12.67, −3.03) −1.22 (−10.25, 7.81) Co-worker gender About even (ref) Mostly men 3.22 (0.94, 5.51) 1.06 (−1.35, 3.47) 0.90 (−3.54, 5.35) 1.48 (−2.27, 5.23) 0.45 (−5.81, 6.70) Mostly women −1.54 (−4.77, 1.68) −1.64 (−4.65, 1.36) −4.20 (−10.10, 1.70) −1.26 (−7.23, 4.70) −1.29 (−6.89, 4.31) Co-worker race/ethnicity Mostly same (ref) Mostly different 1.58 (−0.93, 4.09) 0.81 (−1.56, 3.17) −6.01 (−11.24, −0.78) 2.30 (−1.43, 6.03) 2.35 (−2.84, 7.54) About even 0.94 (−1.87, 3.75) 0.80 (−1.89, 3.49) −6.71 (−12.31, −1.11) 3.16 (−1.31, 7.62) 2.40 (−3.28, 8.07) Response to unfair treatment Do something/talk (ref) Do something/quiet 1.07 (−2.35, 4.48) 1.84 (−1.32, 4.99) −0.32 (−6.44, 5.80) 1.28 (−4.54, 7.11) 4.52 (−2.13, 11.16) Accept/talk 1.12 (−1.44, 3.68) −0.26 (−2.64, 2.12) 2.39 (−2.22, 7.00) −2.70 (−6.62, 1.22) 2.25 (−3.38, 7.89) Accept/quiet 3.01 (0.07, 5.96) 2.34 (−0.33, 5.01) 3.54 (−1.67, 8.75) 0.55 (−4.04, 5.14) 2.64 (−3.21, 8.48) a In multivariable models, all parameter estimates are mutually adjusted for all the other included variables. Table options Analyses stratified by race/ethnicity yielded similar results, albeit with three differences. First, the interaction of sexual harassment and gender was strongest in the white population (6.3 mm Hg, 95% CI 0.8, 11.8) but similarly weaker among the black (2.5 mm Hg, 95% CI −1.02, 6.01) and Latina (2.6 mm Hg, 95% CI −2.5, 7.8) populations. Second, the association of SBP with hypertension medication was much greater among the black and Latino compared to white workers (10.3 and 8.8 vs. 3.3 mm Hg). Third, only black workers exhibited the protective effect for transportation.