Automatic Extraction of Research Themes in Epidemiological Criminology From PubMed Abstracts From 1946 to 2020: Text Mining Study
George Karystianis,Paul Simpson,Wilson Lukmanjaya,Natasha Ginnivan,Goran Nenadic,Iain Buchan,Tony Butler
DOI: https://doi.org/10.2196/49721
2023-09-22
Abstract:Background: The emerging field of epidemiological criminology studies the intersection between public health and justice systems. To increase the value of and reduce waste in research activities in this area, it is important to perform transparent research priority setting considering the needs of research beneficiaries and end users along with a systematic assessment of the existing research activities to address gaps and harness opportunities. Objective: In this study, we aimed to examine published research outputs in epidemiological criminology to assess gaps between published outputs and current research priorities identified by prison stakeholders. Methods: A rule-based method was applied to 23,904 PubMed epidemiological criminology abstracts to extract the study determinants and outcomes (ie, "themes"). These were mapped against the research priorities identified by Australian prison stakeholders to assess the differences from research outputs. The income level of the affiliation country of the first authors was also identified to compare the ranking of research priorities in countries categorized by income levels. Results: On an evaluation set of 100 abstracts, the identification of themes returned an F1-score of 90%, indicating reliable performance. More than 53.3% (11,927/22,361) of the articles had at least 1 extracted theme; the most common was substance use (1533/11,814, 12.97%), followed by HIV (1493/11,814, 12.64%). The infectious disease category (2949/11,814, 24.96%) was the most common research priority category, followed by mental health (2840/11,814, 24.04%) and alcohol and other drug use (2433/11,814, 20.59%). A comparison between the extracted themes and the stakeholder priorities showed an alignment for mental health, infectious diseases, and alcohol and other drug use. Although behavior- and juvenile-related themes were common, they did not feature as prison priorities. Most studies were conducted in high-income countries (10,083/11,814, 85.35%), while countries with the lowest income status focused half of their research on infectious diseases (47/91, 52%). Conclusions: The identification of research themes from PubMed epidemiological criminology research abstracts is possible through the application of a rule-based text mining method. The frequency of the investigated themes may reflect historical developments concerning disease prevalence, treatment advances, and the social understanding of illness and incarcerated populations. The differences between income status groups are likely to be explained by local health priorities and immediate health risks. Notable gaps between stakeholder research priorities and research outputs concerned themes that were more focused on social factors and systems and may reflect publication bias or self-publication selection, highlighting the need for further research on prison health services and the social determinants of health. Different jurisdictions, countries, and regions should undertake similar systematic and transparent research priority-setting processes.