Hi everyone, my name is Sindhura Nimani and I'm currently a third-year medical student at Rowan Virtua School of Osteopathic Medicine in New Jersey, and I'm very excited to present our group's research abstract, Participant Attrition, Smartphone Intervention for PTSD in Primary Care. Working on this abstract with me are Danielle Schweitzer and Faith Schenck, both are clinical psychology PhD candidates, Hannah Ngo, Kaylee Dillon, and Muntaha Rana, who are all second-year medical students, and Dr. Anne Jones, who is the Assistant Dean of Student Affairs at the medical school, Associate Professor of Family Medicine, and the main project's PI. We have no financial disclosures to report, and the study has been funded through a grant from the New Jersey Health Foundation. PTSD affects 12.5% of the primary care population, and unfortunately, prolonged wait times to see a mental health expert is resulting in poor patient access. However, integrated behavioral health models have been shown to improve access to mental health through interprofessional collaboration within primary care settings. Mobile applications are an emerging trend in healthcare treatment and self-care, which expands access and meets patients where they are. In today's environment, this includes text spaces and artificial intelligence. Main study goals include recruiting 80 to 100 patients to screen positive for PTSD symptoms with the DSM-5 CAHPS-5, developing an intervention protocol for integrating PTSD coach into primary care clinics, and investigating efficacy of the application among this patient population. Specific project goals include examining the most common PTSD presentations in primary care settings, and analyzing the relationship between symptom severity and attrition. Though there is some evidence of efficacy with using smartphone interventions to manage PTSD symptoms, the predictors of participant attrition remain unclear. So our objective is to identify predictors of attrition associated with decreased follow-up rates in the study, and consequently determine potential ways in which we could modify the study to facilitate higher follow-up rates among participants. This is a prospective cohort study at the Rowan Family Medicine office in Sewell, NJ. Inclusion criteria, participants age 18 through 89, screened positive for at least one PTSD symptom as determined by the primary care provider, and access to an electronic device, preferably a smartphone. Exclusion criteria, active suicidal or homicidal ideations, substance use at the time of the visit, and inaccessibility to a device that's compatible with the application. Eligible participants were identified and introduced to the study team via a warm handoff, screened with the CAHPS-5, instructed to use PTSD Coach, and then brought back for a 30-day follow-up. Data analysis was done with a random forest using the Party Package in R, and RF aggregates fitted predictions across decision trees to determine best predictor performance. 33 participants completed baseline measures, whereas 9 completed a 30-day follow-up. Two participants withdrew, and 22 participants were lost to non-response. Based on variable importance, results indicated that the best predictors of attrition were intrusion symptom severity and quantity, followed by traumatic events including natural disasters and sexual assault. Approximately 2.7% of the variance in attrition was explained by the full model. Model visualizations suggested that increased levels of intrusion symptom severity and quantity were present within individuals who follow-up compared to those who solely completed intake. Furthermore, more participants who completed follow-up had experienced a natural disaster or sexual assault compared to those who did not complete follow-up. So in conclusion, our study found higher attrition rates compared to studies examining applications for other mental health conditions and clinician-supported applications for PTSD in primary care. Sexual assault and natural disasters were both predictive of decreased attrition rates, and our results suggested that lower levels of intrusion symptoms and specific traumatic events were most predictive of attrition compared to other symptoms of PTSD in traumatic events. These events differ with respect to personal directedness, suggesting that patients in primary care who have experienced varying traumatic events may use mobile interventions for trauma. It's possible that participants experiencing intrusion symptoms found the intervention to be useful. And third, providing reminders or additional guidance to participants may reduce attrition rates within mobile interventions. There are some limitations to consider. The small sample size limits statistical power. And despite being the gold standard for PTSD diagnosis, CAHPS-5 is noted to be lengthy and an intense screening tool for this type of evaluation. Taking these factors into consideration, this may possibly lead to an underestimation of PTSD prevalence in primary care. Future directions. For integration of mobile health into primary care clinics for trauma, alignment of research protocols to models of integrated primary care with regards to workflow and scheduling processes may improve attrition rates. Future research may examine how guidance or clinician support may facilitate adherence to standalone and blended digital interventions for trauma within primary care. And lastly, within primary care settings, providers may therefore consider recommending mobile applications without a trauma focus for individuals who are experiencing intrusion symptoms and for a range of traumatic events. Thank you for listening to this presentation.

PTSD Coach app

smartphone interventions

primary care

attrition factors

mobile interventions