Developing a suicide risk model for use in the Indian Health Service

A New Tool to Combat Rising Suicide Rates Among American Indians

In a groundbreaking effort to combat the alarming rise in suicide rates among American Indians, researchers have developed and evaluated an electronic health record (EHR)-based model specifically designed for this population. This innovative approach aims to reduce the risk of suicide attempts and deaths by leveraging data from patient visits to identify those at highest risk. The model, evaluated using data from over 16,800 patients, has shown promising results in predicting suicide attempts and deaths within 90 days of a visit.

The Alarming Reality

Suicide remains a devastating and all-too-frequent event among American Indians. Complex interrelated factors such as mental health disorders, substance abuse, intergenerational trauma, and community-wide issues contribute to this increased risk. Despite the strength and resilience of American Indian families and communities, the statistics are stark: in the Indian Health Service (IHS), for every 100,000 patients, 1.9% experience at least one suicide attempt, and 0.2% die by suicide.

The Need for a Targeted Approach

Traditional suicide screening methods often fall short in identifying high-risk individuals, particularly in diverse populations like American Indians. Existing models may not account for the unique cultural, social, and historical contexts that influence mental health and suicide risk. Therefore, developing a model tailored to this specific population is crucial.

How the EHR-Based Model Works

The EHR-based model incorporates a wide range of features to predict the risk of suicide attempts or deaths. These features include:

• Demographics: Age, sex, and other patient characteristics.
• Medications: Prescription information that could influence mood and behavior.
• Diagnoses: Conditions like depression, anxiety, bipolar disorder, PTSD, and substance abuse.
• Screening Tool Scores: Results from relevant screening tools used during patient visits.

Evaluating the Model

Researchers compared the predictive performance of the EHR-based model with existing suicide screening methods. The results were striking:

• Logistic Regression and Random Forest Models: These machine-learning models achieved an area under the receiver operating characteristic curve (AUROC) of 0.83, significantly outperforming enhanced screening methods (AUROC 0.64).
• Key Predictive Factors: The model identified younger age, Medicaid enrollment, diagnoses of depression, anxiety, bipolar disorder, PTSD, TBI, or suicidal ideation, alcohol or substance abuse, and past positive screens for intimate partner violence as significant risk factors.

Implications and Future Directions

The development of this EHR-based model represents a significant step forward in addressing the complex issue of suicide among American Indians. By leveraging existing patient data, healthcare providers can more effectively identify high-risk individuals and implement targeted interventions.

The model's high AUROC value indicates its robust performance in distinguishing between those at risk and those who are not. This is particularly important given that the existing screening methods failed to flag any of the suicide deaths in the study's data.

Integration into Clinical Practice

To integrate this model into daily clinical practice, healthcare providers need comprehensive training. The IHS offers various training programs aimed at enhancing the skills of direct care staff in assessing and managing suicide risk. The Zero Suicide Initiative, for instance, provides funding for projects that focus on developing a system-wide approach to reducing suicide rates.

Assessing and Managing Suicide Risk (AMSR) training for direct care staff is a half-day program that includes video demonstrations, group discussions, written and paired practice, case reviews, and expert teaching. This training is essential for ensuring that healthcare providers are equipped with the necessary skills to recognize, assess, and manage suicide risk effectively.

Collaboration and Public Health Approach

Suicide prevention is not just a healthcare issue; it requires a public health approach that involves the community, family, and society. The Connect Suicide Prevention/Intervention Training emphasizes prevention and intervention strategies that consider the individual, family, community, tribe, and societal factors. This holistic approach helps in developing more effective and culturally relevant risk identification tools.

Conclusion

The development and evaluation of the EHR-based suicide risk model for the Indian Health Service represents a significant milestone in combating the rising suicide rates among American Indians. By integrating this model into clinical practice and ensuring comprehensive training for healthcare providers, we can better identify high-risk individuals and implement targeted interventions. The collaboration between Tribal, Federal, and other partners is crucial for creating a safety net of interconnected programming that maximizes effectiveness in services and protects individuals against suicide risk.

References

• https://www.nature.com/articles/s44184-024-00088-5
• https://www.ihs.gov/sasp/mtbpp/spmodelstrainings/
• https://www.ihs.gov/zerosuicide/
• https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2824733

By developing this EHR-based model and integrating it into clinical practice, we can make significant strides in reducing the devastating impact of suicide among American Indians. The future holds promise as we continue to address this complex issue with targeted interventions and comprehensive training programs.