Secondary Analysis of the National Elder Mistreatment Study: Exploration of Prevalence, Risk, and Protective Factors within American Indian and Alaska Native Populations

Limited research on elder abuse among American Indians and Alaska Natives suggests higher rates of abuse. However, no research has used a nationally representative sample to measure elder abuse prevalence among both American Indian and Alaska Native men and women. Using data from the National Elder Mistreatment Survey (NEMS), comparisons were made between American Indians and Alaska Natives, Black, and White respondents.

Descriptive statistics were calculated within each of the three racial groups. Comparison between the three race groups consisted of chi-square test of independence or Fisher’s exact tests for categorical variables, and non-parametric Mann-Whitney U or Kruskal Wallis tests for continuous variables. Bivariate, unadjusted logistic regression was conducted with twenty-four independent variables. Multiple logistic regression was conducted for six abuse types using a stepwise selection method that incorporated significant variables for the American Indian and Alaska Native group. Replication of the final American Indian and Alaska Native group’s model identified via stepwise logistic regression was conducted for other race groups as the last step.

There were differences in the prevalence of multiple abuse types and also demographic, socioeconomic, social, and health status of American Indian and Alaska Native elders, White, and Black respondents. We also found that American Indian and Alaska Native respondents had more similarities in demographic and socioeconomic characteristics compared with Black respondents than White, though significant differences still existed between the two samples. The cumulative prevalence of emotional, physical, and sexual mistreatment in the past year; neglect; and financial abuse by a family member for the American Indian and Alaska Native group was 33%. This is almost double that of the overall findings (17.1%) reported in the original NEMS study. Lifetime prevalence of mistreatment and prevalence of abuse since the age of 60 for American Indians and Alaska Natives were generally higher than for White respondents.

No single set of bivariate predictors was the same for any abuse type between the three race groups. Logistic regression models constructed based on predictors specific to the American Indian and Alaska Native group contained some similar variables in the models constructed for the original study, most specifically social support. Models built to American Indian and Alaska Native group’s specification were not all significant nor was there good model fit for the Black and White groups for all models. The predictive capacity and ability to classify abuse cases was better for the American Indian and Alaska Native group’s models. There was much room for improvement in the predictive capacity of all final models.

This dissertation addresses gaps in elder abuse literature for the American Indian and Alaska Native population by identifying prevalence and predictors that incorporated large comparison groups and consistently measured abuse types. Furthermore, it revealed that the complex context and how it intersects to shape abuse outcomes among the American Indian and Alaska Native population must be considered. There is a need for the development of more advanced predictive modeling to aid health care providers and others who work with elders in the screening and detection of abuse. These gaps identified among the American Indian and Alaska Native population mirror those of the larger elder abuse field. Researchers, health care providers, tribal leaders, and other policy makers must take notice and then act to aid in reducing morbidity, mortality, and the overall impact of violence perpetrated against American Indian and Alaska Native elders.