Social labeling of people with behavioral health disorders falls under the umbrella of "stigma" and plays a key role in limiting the access to behavioral healthcare. Currently, the U.S. spends an estimated $201B on behavioral health disorders every year, making it the number one most expensive medical condition. In any given year, 43.8M Americans experience a behavioral health disorder. More than half of them receive no treatment mainly because of their fear of being socially disgraced or stigmatized against. The lack of a scalable and analytical approach to monitor stigma over time makes it difficult to compare findings across contexts. This research establishes the Stigma Index, an innovative analytical tool, that allows for (a) measuring behavioral health stigma uniformly and systematically over time, and (b) comparing the prevalence of stigma in different populations. Machine learning classification was conducted and resulted in eight questions that are used to aggregate sentiments towards individuals with behavioral health disorders. To compute the Stigma Index, the relative scores for each of the eight index questions are first derived from the percentage difference between favorable and unfavorable responses and then summed into a composite measure. To validate the tool, changes in Stigma Index were monitored and used to capture real-life differences in stigma levels across different populations. This first of its kind computational approach to standardize the measurement of stigma offers promising applications to improve policy objectives that (1) ensure the social inclusion of behavioral health consumers, and (2) promote effective population-based interventions in reducing behavioral health stigma.
