Introduction

Project Overview

Our goals are to: 1) examine the social and health determinants of racial disparities in sepsis mortality; 2) test the hypothesis that socioeconomic factors, health behaviors and health status will explain the racial disparity in sepsis mortality; and 3) examine if healthcare utilization is associated with decreased sepsis mortality in fully adjusted multivariate, multivariable models. We will use Cox Proportional Hazards modeling strategies to examine and rank the effects of each socioeconomic and health variable on all-cause and cause-specific age- and sex-adjusted mortality. Then each variable will be assessed for its confounding and/or mediation effects the Black race risk ratio for all-cause and cause-specific mortality. Finally, these analyses will inform the development of multivariate, multivariable models. The results will fill a critical knowledge gap in the field of sepsis epidemiology by providing the first and largest nation-level longitudinal cohort study examining the risk factors for sepsis mortality and the racial disparity in sepsis mortality. This will address an important US public health concern by potentially identifying modifiable risk factors for this prevalent disease.

Dataset

We utilize the 1999-2005 National Health Interview Surveys (NHIS) linked at the individual-level to outcomes in the National Death Index (NDI), years 1999-2011. The outcomes of interest in this analysis is combined septicemia and influenza/pneumonia mortality, defined using the National Vital Statistics ICD-10 coding schema. these codes can occur in any position on the death record. The competing outcome of interest is ‘other cause’ mortality, deinfed as any death without an ICD-10 code for septicemia, influenza or pneumonia among the listed caused odeath on the death record.

Age- & Sex-Adjusted Mortality Hazard for Race

This report presents the results of age- and sex-adjusted Cox proportional hazard ratios for sespsis, influenza and pneumonia mortality in Blacks compared to Whites. All statistics are computed using the SAS 9.2 SURVEY family of procedures and account for the weighted, complex, multi-stage survey design of the NHIS. Missing observations are treated as missing “not completely at random”


Base Model

This forest plot represents the results of an age-group, sex, and race/ethnicity Cox PH model for sepsis/influenza/pneumonia death and censoring for other causes of death. This gives us our baseline age- and sex- adjusted hazard ratio for Blacks compared to Whites of 1.477.

Hazard Ratios for Black Compared to White Mortality for Sepsis, Influenza & Pneumonia Death Base Model
variable Value Probt
17-24 years NA (NA-NA)
25-34 years 2.573 (1.276-5.187) 0.0084
35-44 years 6.109 (3.135-11.905) <.0001
45-54 years 14.766 (7.724-28.225) <.0001
55-64 years 39.815 (21.043-75.333) <.0001
65-74 years 97.635 (52.088-183.009) <.0001
75 years and over 308.108 (163.648-580.089) <.0001
Non-Hispanic White NA (NA-NA)
Hispanic 1.067 (0.918-1.239) 0.3976
Other 0.942 (0.72-1.233) 0.6632
Non-Hispanic Black 1.477 (1.303-1.673) <.0001
Female NA (NA-NA)
Male 1.44 (1.33-1.559) <.0001

Health Status

Comorbid Medical Conditions

Age- and Sex-Adjusted Hazard Ratios for Black Compared to White Mortality for Sepsis, Influenza & Pneumonia Death When Controlling for Given Covariate
Label Value Probt
Asthma 1.47 (1.3-1.67) <.0001
Hay Fever 1.48 (1.3-1.67) <.0001
Severe headache/migrain3(3m)? 1.48 (1.3-1.68) <.0001
Angina 1.49 (1.31-1.68) <.0001
Cancer 1.53 (1.35-1.73) <.0001
Chronic bronchitis 1.48 (1.31-1.68) <.0001
Diabetes mellitus 1.39 (1.22-1.57) <.0001
Emphysema 1.51 (1.33-1.71) <.0001
Heart condition 1.51 (1.33-1.72) <.0001
Hypertension 1.41 (1.24-1.6) <.0001
Weak or failing kidneys 1.44 (1.27-1.64) <.0001
Liver condition 1.48 (1.3-1.67) <.0001
Myocardial infarction 1.5 (1.32-1.7) <.0001
Neck pain 1.49 (1.31-1.69) <.0001
Limb pain 1.49 (1.31-1.69) <.0001
Sinusitis(12m) 1.48 (1.31-1.68) <.0001
Stroke 1.46 (1.29-1.65) <.0001
Stomach or intenstinal ulcer 1.49 (1.31-1.69) <.0001

General Health Condition

Age- and Sex-Adjusted Hazard Ratios for Black Compared to White Mortality for Sepsis, Influenza & Pneumonia Death When Controlling for Given Covariate
Label Value Probt
Compared to 1 yr age, health is… 1.46 (1.29-1.65) <.0001
Days health kept you in bed(12m)? 1.42 (1.26-1.61) <.0001
Body-mass Index 1.46 (1.29-1.65) <.0001
Any functional limitation 1.46 (1.29-1.65) <.0001
Limited in any way 1.33 (1.17-1.5) <.0001
Health in general is… 1.17 (1.03-1.32) 0.0151
Need help with ADLS 1.35 (1.19-1.54) <.0001
Need help with IADLS 1.32 (1.16-1.5) <.0001
Health prevents working 1.28 (1.14-1.45) <.0001
Health problem requires equipment? 1.34 (1.18-1.51) <.0001

Health Behaviors

Age- and Sex-Adjusted Hazard Ratios for Black Compared to White Mortality for Sepsis, Influenza & Pneumonia Death When Controlling for Given Covariate
Label Value Probt
Alcohol Use 1.36 (1.2-1.55) <.0001
Frequency alcohol(12m)? 1.42 (1.25-1.61) <.0001
Chances of AIDS virus 1.48 (1.3-1.68) <.0001
Cigarette Use 1.49 (1.31-1.69) <.0001
Light or moderate activity 1.35 (1.19-1.53) <.0001
Tobacco smoking status 1.49 (1.31-1.68) <.0001
Muscle strengthening activity? 1.46 (1.28-1.65) <.0001
Vigorous activity 1.4 (1.23-1.58) <.0001

Healthcare Utilization

Age- and Sex-Adjusted Hazard Ratios for Black Compared to White Mortality for Sepsis, Influenza & Pneumonia Death When Controlling for Given Covariate
Label Value Probt
Seen health professional(12m)? 1.47 (1.29-1.67) <.0001
Seen a general provider(12m)? 1.48 (1.3-1.68) <.0001
Visits to ER(12m)? 1.4 (1.23-1.59) <.0001
Overnight hospital stays(12m)? 1.45 (1.28-1.64) <.0001
Received healthcare > 10 times(12m)? 1.45 (1.28-1.64) <.0001
Seen a healthcare professional(2w)? 1.46 (1.29-1.66) <.0001
Spoken to healthcare professional(2w)? 1.49 (1.32-1.69) <.0001
Overnight in hospital(12m)? 1.46 (1.29-1.66) <.0001
Had flu shot(12m)? 1.51 (1.33-1.72) <.0001
Ever had a pneumonia shot? 1.5 (1.32-1.71) <.0001
A usual place for healtchare? 1.48 (1.3-1.68) <.0001

Socioeconomic

Material Capital

Age- and Sex-Adjusted Hazard Ratios for Black Compared to White Mortality for Sepsis, Influenza & Pneumonia Death When Controlling for Given Covariate
Label Value Probt
Not afford medicines(12m)? 1.44 (1.27-1.64) <.0001
Healthcare spending(12m)? 1.45 (1.27-1.65) <.0001
W/o health insurance(12m)? 1.48 (1.3-1.68) <.0001
House ownership arrangement 1.33 (1.18-1.51) <.0001
No health insurance 1.48 (1.31-1.68) <.0001
Delayed healthcare bc of cost(12m)? 1.47 (1.29-1.66) <.0001
Foregone healthcare bc of cost(12m) 1.45 (1.28-1.64) <.0001
A telephone number 1.48 (1.3-1.68) <.0001

Age- and Sex-Adjusted Hazard Ratios for Black Compared to White Mortality for Sepsis, Influenza & Pneumonia Death When Controlling for Given Covariate
Label Value Probt
Earnings (12m) 1.41 (1.24-1.6) <.0001
Family Income(12m) 1.22 (1.07-1.39) 0.0026
Family Income(12m) 1.31 (1.15-1.49) <.0001
Income-poverty ratio 1.16 (1.01-1.34) 0.037

Human Capital

Age- and Sex-Adjusted Hazard Ratios for Black Compared to White Mortality for Sepsis, Influenza & Pneumonia Death When Controlling for Given Covariate
Label Value Probt
Education achieved 1.31 (1.15-1.48) <.0001
Language of the interview 1.46 (1.29-1.65) <.0001
Born in the US 1.48 (1.31-1.68) <.0001

Social Capital

Age- and Sex-Adjusted Hazard Ratios for Black Compared to White Mortality for Sepsis, Influenza & Pneumonia Death When Controlling for Given Covariate
Label Value Probt
US Citizen 1.48 (1.3-1.68) <.0001
Highest education within family 1.31 (1.16-1.49) <.0001
Family composition 1.44 (1.27-1.63) <.0001
Significant other 1.35 (1.19-1.53) <.0001