Can AI Mitigate Bias in Writing Letters of Recommendation?

Letters of recommendation play a significant role in higher education and career progression, particularly for women and underrepresented groups in medicine and science. Already, there is evidence to suggest that written letters of recommendation contain language that expresses implicit biases, or unconscious biases, and that these biases occur for all recommenders regardless of the recommender's sex. Given that all individuals have implicit biases that may influence language use, there may be opportunities to apply contemporary technologies, such as large language models or other forms of generative artificial intelligence (AI), to augment and potentially reduce implicit biases in the written language of letters of recommendation. In this editorial, we provide a brief overview of existing literature on the manifestations of implicit bias in letters of recommendation, with a focus on academia and medical education. We then highlight potential opportunities and drawbacks of applying this emerging technology in augmenting the focused, professional task of writing letters of recommendation. We also offer best practices for integrating their use into the routine writing of letters of recommendation and conclude with our outlook for the future of generative AI applications in supporting this task.

Infectious Vaginitis, Cervicitis, and Pelvic Inflammatory Disease.

Vaginal symptoms are one of the most common reasons women consult with physicians and can significantly impact quality of life. The differential diagnosis of vaginal discharge includes physiologic discharge, vaginitis, cervicitis, and pelvic inflammatory disease (PID). Vaginitis is inflammation of the vagina, most commonly caused by bacterial vaginosis (BV), vulvovaginal candidiasis, and trichomoniasis infections. Cervicitis is an inflammation of the cervix and typically caused by Chlamydia trachomatis and Neisseria gonorrhoeae. PID is infection of the female upper genital tract, involving the uterus, fallopian tubes, ovaries, and/or pelvic peritoneum and usually caused by Chlamydia trachomatis, Neisseria gonorrhoeae, and bacterial vaginosis-associated pathogens. A pelvic exam should be performed for any woman presenting with vaginal discharge to confirm the diagnosis and rule out an upper tract infection. BV and vulvovaginal candidal infections only require treatment if symptomatic and do not require partner therapy, whereas treatment and partner therapy is recommended for sexually transmitted illnesses, such as trichomoniasis, chlamydia and gonorrhea. Vaginitis may be uncomfortable, but rarely leads to serious long-term consequence, but pelvic inflammatory disease can lead to serious long-term sequelae, including increased risk for ectopic pregnancy, infertility, and chronic pelvic pain.

A Closer Look at Women's Health Centers: Historical Lessons and Future Aims.

Women's Health Centers (WHC) have evolved over the last few decades as comprehensive centers for women's health care. This article reviews the history and evaluation of WHC, as well as opportunities for women's health training. Prior studies comparing WHC with traditional primary care and obstetrics/gynecology clinics have found that WHC offer at least similar levels of preventative care, may increase access to care for a more diverse patient population, and improve patient/provider relationship satisfaction. WHC also increase women's health providers' education and research opportunities. There is still a gap in women's health education and training, although residency and fellowship programs have aimed to address this through women's health tracks and fellowships. The coronavirus disease 2019 (COVID-19) pandemic and its negative impact on women's access to care have further highlighted the potential of WHC to meet women's health care demands. WHC can provide comprehensive, convenient, and single-site care for women. The increased opportunities for women's health training through WHC give rise to more representation in leadership and investment in women's health. New research is needed to reassess and further evaluate health outcomes of WHC compared with traditional care models.

Let's talk about sex: Development and evaluation of a sexual history and counseling curriculum for internal medicine interns.

OBJECTIVE: We developed a curriculum to increase internal medicine interns' proficiency in sexual history taking and sexually transmitted illness (STI) counseling. METHODS: The 4-h curriculum included didactics and interactive components, and was evaluated with matched pre- and post-surveys. RESULTS: Twenty-one interns completed the curriculum. Mean knowledge score improved from 59% to 76% from pre- to post-curriculum (P = 0.004). Median comfort score (Likert scale 1-5) with obtaining a sexual history improved from 3.8 [IQR 3.0, 4.0] to 3.8 [IQR 3.6, 4.6] and 3.8 [IQR 3.6, 4.0] to 4.1 [IQR 3.9, 4.8] for male (P = 0.05) and female patients (P = 0.007), respectively. Median frequency score for obtaining a sexual history improved from 2.9 [IQR 2.7, 3.0] to 3.1 [IQR 2.8, 3.4] and 3.2 [IQR 2.8, 3.7] to 3.4 [IQR 3.2, 4.0] for male (P = 0.16) and female patients (P = 0.008), respectively. Pre- and post- curriculum, interns reported significantly higher comfort and frequency in obtaining sexual histories from female vs. male patients. Post- curriculum, interns reported significantly higher comfort with positive STI counseling. PRACTICE IMPLICATIONS: Our curriculum improved interns' knowledge and comfort in sexual history taking and STI counseling. Future interventions should address sex disparities in sexual history taking.

Improvement in Quality Metrics by the UPMC Enhanced Care Program: A Novel Super-Utilizer Program.

The aim was to evaluate pre-post quality of care measures among super-utilizer patients enrolled in the Enhanced Care Program (ECP), a primary care intensive care program. A pre-post analysis of metrics of quality of care for diabetes, hypertension, cancer screenings, and connection to mental health care for participants in the ECP was conducted for patients enrolled in ECP for 6 or more months. Patients enrolled in ECP showed statistically significant improvements in hemoglobin A1c, retinal exams, blood pressure measurements, and screenings for colon cancer, and trends toward improvement in diabetic foot exams and screenings for cervical and breast cancer. There was a significant increase in connecting patients to mental health care. This study shows that super-utilizer patients enrolled in the ECP had significant improvements in quality metrics from those prior to enrollment in ECP.

An Innovative Framework to Improve Teratogenic Medication Risk Counseling.

Health care providers in the Veterans Affairs system are caring for a growing number of women of childbearing age. Prior research shows that a significant number of women veterans are prescribed category D or X teratogenic medications. However, the majority do not receive teratogenic risk or contraceptive counseling. In order to improve teratogenic medication prescribing practices at the Veterans Affairs Pittsburgh Healthcare System, we developed a framework called TARCC to guide primary care providers through appropriate teratogenic risk counseling. The components of TARCC are teratogen (Are you prescribing a potential teratogen?), alternative (Did you consider a nonteratogenic alternative?), risks (Did you discuss the risks of the medication?), contraception (Did you discuss contraception?), and chart (Did you chart your discussion?). We incorporated TARCC into an electronic health record reminder. Additional probes, educational tools, and options to place consults and orders are included within the reminder. The Teratogen Pharmacy E-consult prompts a pharmacist to electronically review a patient's medications and provide recommendations regarding potential alternatives and discussions of risks and benefits. This article describes the development and implementation of the TARCC framework in a clinical reminder.

Feasibility of patient navigation in resident primary care practice: a screening mammography quality improvement initiative.

Patient navigation (PN) has been used to improve breast cancer outcomes but has not been evaluated in resident practices. The objective of this study was to evaluate the feasibility of PN in resident clinics to improve biennial screening mammography. All participants whose last screening mammogram was at 18 months or after received PN. Feasibility outcomes included practicality (ability to carry out intervention), effectiveness (ability of intervention to increase biennial screening mammography rates), and acceptability (resident satisfaction with PN). Despite difficulty contacting patients, we found PN improved screening mammography adherence in resident practices serving vulnerable populations. Adherence rates increased significantly postintervention. Residents reported interest in using PN in practice.

The impact of health insurance reform on insurance instability.

We investigated the impact of the 2006 Massachusetts health care reform on insurance coverage and stability among minority and underserved women. We examined 36 months of insurance claims among 1,946 women who had abnormal cancer screening at six community health centers pre-(2004-2005) and post-(2007-2008) insurance reform. We examined frequency of switches in insurance coverage as measures of longitudinal insurance instability. On the date of their abnormal cancer screening test, 36% of subjects were publicly insured and 31% were uninsured. Post-reform, the percent ever uninsured declined from 39% to 29% (p .001) and those consistently uninsured declined from 23% to 16%. To assess if insurance instability changed between the pre- and post-reform periods, we conducted Poisson regression models, adjusted for patient demographics and length of time in care. These revealed no significant differences from the pre- to post-reform period in annual rates of insurance switches, incident rate ratio 0.98 (95%- CI 0.88-1.09). Our analysis is limited by changes in the populations in the pre- and post-reform period and inability to capture care outside of the health system network. Insurance reform increased stability as measured by decreasing uninsured rates without increasing insurance switches.

The impact of insurance coverage during insurance reform on diagnostic resolution of cancer screening abnormalities.

We examined the impact of Massachusetts insurance reform on the care of women at six community health centers with abnormal breast and cervical cancer screening to investigate whether stability of insurance coverage was associated with more timely diagnostic resolution. We conducted Cox proportional hazards models to predict time from cancer screening to diagnostic resolution, examining the impact of 1) insurance status at time of screening abnormality, 2) number of insurance switches over a three-year period, and 3) insurance history over a three-year period. We identified 1,165 women with breast and 781 with cervical cancer screening abnormalities. In the breast cohort, Medicaid insurance at baseline, continuous public insurance, and losing insurance predicted delayed resolution. We did not find these effects in the cervical cohort. These data provide evidence that stability of health insurance coverage with insurance reform nationally may improve timely care after abnormal cancer screening in historically underserved women.