Research Opportunities

 

Bowdoin St. is looking to partner with investigators who share our mission for increasing representation in science of marginalized communities and for conducting research that incorporates community input.

To propose a research collaboration, please fill out this intake survey and we will get back to you promptly.

Our research review process is below:  Research review process

 

Things Happening at Bowdoin St.

More coming soon. For research activities happening at Bowdoin go here

Recent Publications

  • Background: Health disparities in colorectal cancer screening, diagnosis, and treatment related to socioeconomic status and insurance are well documented. These disparities often intersect with self-identified race, ethnicity, and language. We sought to identify equity-relevant associations with completion of a colonoscopy order within a Medicaid population.

    Methods: We identified all colonoscopy orders for patients with Medicaid insurance seen at a hospital-based practice and an affiliated community health center during a three-year period (2018-2021.) Time between ordering and completion of colonoscopy was the outcome of interest in a Cox proportional hazards model. The model was adjusted by equity-specific variables and other potential confounders.

    Results: In this Medicaid population, 48.8% (972/1997) colonoscopy orders were not completed within 365 days. Patients seen at a community health center and those using a patient portal had a higher hazard ratio of colonoscopy completion, when controlling for other potential confounders and equity-specific variables (p =.002, p =.005.) Patients with depression and those whose colonoscopy was ordered after a telehealth visit had lower HR of completing colonoscopy (p=.002 and p<.0001, respectively.) Self-reported race, Hispanic ethnicity, and language were not independently associated with the outcome.

    Conclusions: Successful completion within 365 days of a colonoscopy order was very low overall (48.8%). Care at a community health center and documented use of the patient portal increased the hazards of a colonoscopy order, whereas a preceding telehealth appointment and a diagnosis of depression lowered the hazards of success. Common practices at the CHC, such as additional scheduling support, might explain its protective effect, suggesting important resiliency factors that warrant further exploration.

  • Salant, Talya. 2024. Symptom Tracking in Primary Care: Creating and Testing a Safety-Net System​. SGIM National Meeting 5/2024.

    Background: Tracking of worrisome symptoms in primary care can provide a valuable safety-net for preventing missed or delayed diagnoses. However, few reliable symptom tracking systems exist within primary care, leading to potential patient harm and provider burnout. We combined systems engineering and userbased design to create and test a worrisome symptom tracking system at a community-based primary care practice in Boston.

    Methods: Chart reviews of 86 patients established baseline rates of worrisome symptom follow-up in usual primary care. Based on findings from the chart review and input from a PCP survey, a team of systems engineers, PCPs, nurses, and software engineers met over a period of four months to iterate a semiautomated symptom tracking system prototype and to ensure feasibility within a busy primary care practice. The system was designed to provide a structured “test of time” through a referral template, timed nurse-led outreach, reminder flags, and communication back to the PCP. Work is ongoing to further automate the outreach through texting and embedding the system within the EHR.

    Results: In usual care, only 60.5% of worrisome symptoms are followed up to diagnosis or resolution (i.e. diagnostic loop closed). Over a period of 10 months (1/10/2023-11/10/2023), 46 patients were referred by eight providers for worrisome symptom monitoring. Of these, 85% were successfully followed up (“closed”) and 33.3% required additional evaluation. The average time to documented resolution of symptoms using the symptom tracking system was 20 days compared to 103 days in usual care.

    Conclusions: A symptom tracking system is feasible and highly effective in improving the rate of symptom follow-up and in reducing the time to symptom follow-up within a community-based primary care practice. Broader implementation and study of symptom tracking systems in diverse clinical settings is needed to further demonstrate their benefit as operationalized “tests of time”.