Utilizing natural language processing to analyze student narrative reflections for medical curriculum improvement.

MOTIVATION: Medical curricula improvement is an ongoing process to keep material relevant and improve the student's learning experience to better prepare them for patient care. Many programs utilize end-of-year evaluations, but these frequently have low response rates and lack actionable feedback. We hypothesized that student reflections written during a fourth year Sub-Internship could be used retrospectively to mine additional information as feedback for future curriculum adjustments. However, reflections contain a large amount of narrative content that would require a cumbersome and essentially infeasible manual review process for busy medical education faculty. METHODS: We developed a Natural Language Processing (NLP) pipeline to automatically identify common themes and topics present in the set of reflective writings that could be used to improve the curriculum. The dataset contains required responses to a faculty issued question submitted between August 2016 and July 2018 about challenges experienced during the medical students fourth year Sub-Internship. RESULTS: Eleven distinct topics were identified, with several being subsequently addressed in future iterations of the curriculum. CONCLUSION: Utilizing NLP on reflective writings was able to identify areas of curriculum improvement, and the NLP results provided a quick and easy way to explore the main themes and challenges expressed by students.

The use and potential impact of digital health tools at the community level: results from a multi-country survey of community health workers.

BACKGROUND: Community health workers (CHWs) are increasingly viewed as a critical workforce to address health system strengthening and sustainable development goals. Optimizing and widening the capacity of this workforce through digital technology is currently underway, though there is skepticism regarding CHWs' willingness and optimism to engage in digital health. We sought to understand CHWs' perceptions on the use of digital health tools in their work. METHODS: We obtained survey data from 1,141 CHWs from 28 countries with complete study information. We conducted regression analyses to explore the relationship between CHWs' training and perceived barriers to digital health access with current use of digital devices/tools and belief in digital impact while adjusting for demographic factors. RESULTS: Most of the CHWs worked in Kenya (n = 502, 44%) followed by the Philippines (n = 308, 27%), Ghana (n = 107, 9.4%), and the United States (n = 70, 6.1%). There were significant, positive associations between digital tools training and digital device/tool use (Adjusted Odds Ratio (AOR) = 2.92, 95% CI = 2.09-4.13) and belief in digital impact (AORhigh impact = 3.03, 95% CI = 2.04-4.49). CHWs were significantly less likely to use digital devices for their work if they identified cost as a perceived barrier (AORmobile service cost = 0.68, 95% CI = 0.49-0.95; AORphone/device cost = 0.66, 95% CI = 0.47-0.92). CHWs who were optimistic about digital health, were early adopters of technology in their personal lives, and found great value in their work believed digital health helped them to have greater impact. Older age and greater tenure were associated with digital device/tool use and belief in digital impact, respectively. CONCLUSIONS: CHWs are not an obstacle to digital health adoption or use. CHWs believe that digital tools can help them have more impact in their communities regardless of perceived barriers. However, cost is a barrier to digital device/tool use; potential solutions to cost constraints of technological access will benefit from further exploration of reimbursement models. Digital health tools have the potential to increase CHW capacity and shape the future of community health work.

Understanding the preference for receiving mental health and substance use support in African Americans 50 and older.

OBJECTIVE: This study aims to determine whether current tobacco and/or alcohol use is associated with setting preferences for seeking support for substance use (SU) and mental health (MH) services to African Americans ages 50 and older. METHODS: Data from 368 African American individuals (aged 50+) who participated in a community-based needs assessment survey were used. Preferences included community-based (e.g., health centers) and traditional settings (e.g., doctor's office). SU was measured as a categorical variable detailing past-month use of conventional cigarettes and alcohol graded by risk levels. Logistic regression models tested the associations between SU and setting preference before and after adjusting for the influence of self-reported MH diagnoses. RESULTS: Prior to adjustment for the influence of MH outcomes, high-risk use of tobacco and alcohol in the past month was associated with a lower odds of preferring MH/SU support in traditional settings (OR = 0.23, 95% CI = 0.06-0.85) compared to participants engaged in no-/low- risk substance use. This association was no longer significant after accounting for the influence of mental health symptoms and covariates. DISCUSSION: These results provide preliminary evidence that mental health outcomes mediate the association between substance use and setting preference for seeking MH/SU support in traditional settings. TRANSLATIONAL SIGNIFICANCE: This exploratory study encourages additional investigation of the association between substance use, setting preferences, and the likelihood of seeking treatment in community health centers using larger sample sizes. Additional opportunities to offer mental health/substance use support to African American older adults within clinical settings should be explored.

A network approach to substance use, internalizing, and externalizing comorbidity in U.S. adults.

INTRODUCTION: Use of conventional cigarettes (CIG), alcohol, marijuana, and sedatives [i.e., benzodiazepines and barbiturates]) commonly co-occur with internalizing and externalizing disorders. It is unclear how these relationships extend to electronic cigarettes (ECIGs) and prescription drugs not prescribed (i.e., sedatives, tranquilizers, and painkillers [PDNP]), and whether they differ by gender. METHODS: Adult data (N = 30,211) from Wave 1 (2013-2014) of the Population Assessment of Tobacco and Health Study were used to estimate a network of current or past-month use for six substances, experiencing four internalizing symptoms in the past month, and experiencing seven externalizing symptoms in the past month. Visual comparisons, global strength invariance, network structure invariance, and edge strength invariance were tested to detail substance use and internalizing/externalizing symptom networks. RESULTS: Overall, networks were consistent between men and women. The strongest substance use/mental health symptom connections estimated as edge-weights (EW) were between marijuana with lying (EW = 0.60, 95% CI = 0.49; 0.70), marijuana with engaging in fights (EW = 0.54, 95% CI = 0.27; 0.81), PDNP with having trouble sleeping (EW = 0.53, 95% CI = 0.40; 0.66), and alcohol and impulsivity (EW = 0.48, 95% CI = 0.42; 0.53). DISCUSSION: There were many weak connections throughout the substance use and internalizing/externalizing network. A few important connections were identified and encourage future study. In particular, PDNP was most strongly associated with internalizing symptoms. Marijuana, alcohol and PDNP use were most strongly associated with externalizing symptoms.

Effects of family history of alcohol problems on alcohol consumption: Stronger for medically underserved men.

Family history (FH), informed by genetics and family environment, can be used by practitioners for risk prediction. This study compares the associations of FH with alcohol outcomes for medically underserved (MUS) men and women with the associations for non-underserved individuals to assess the utility of FH as a screening tool for this high-priority group. Data were from 29,993 adult lifetime drinkers in the Wave 1 (2001-2002) and Wave 2 (2004-2005) National Epidemiologic Survey on Alcohol and Related Conditions. All variables except FH were measured at Wave 2. Dependent variables were 12-month alcohol consumption and alcohol use disorder (AUD). FH scores (FH-SCORE) measured the proportion of first- and second-degree biological relatives with alcohol problems. MUS status was defined by household income at or below 100% of the federal poverty line and participants reporting no usual source of health care. Multivariate linear and logistic regression models tested main and interaction effects. Models showed a significant interaction of FH-SCORE with MUS status (p < .01), with a stronger effect of FH on alcohol consumption for the MUS group. This moderating effect was weaker for women than for men (FH-SCORE x MUS x Sex three-way interaction: p < .01). AUD models showed a significant positive association with FH-SCORE (p < .001) but no association with MUS status and no significant interaction effects. In this sample of lifetime drinkers, FH was associated with higher alcohol consumption, especially for MUS men. These results encourage additional validation of FH scores to prioritize MUS adults at high risk for alcohol problems to receive preventive interventions.

The Association Between Health Literacy and Tobacco Use: Results from a Nationally Representative Survey.

Dual use of conventional cigarettes and electronic cigarettes presents an emerging public health issue. Previous research has demonstrated a negative relationship between health literacy and conventional cigarette (CIG) use. However, the relationship between health literacy and e-cigarette (ECIG) use remains unclear. This studies examines the possible association of health literacy and CIG, ECIG, or dual use. A multinomial regression was used to model the association between health literacy and current CIG use, current ECIG use, or dual tobacco use status using state-optional data from the 2016 Behavioral Risk Factor Surveillance System (BRFSS; N = 40,404). One-third of the sample (N = 13,478; 33.3%) had initiated tobacco use. Approximately 36.6% of participants exclusively used cigarettes. A smaller proportion of participants were dual users of ECIG and CIGs (7.0%) and e-cigarette exclusive users (4.5%). After adjusting for covariates, higher levels of oral health literacy was associated with lower odds of current dual use. However, there was no significant association between written HL and either conventional cigarette use or electronic cigarette use or after adjusting for covariates. Oral messaging around the dangers of CIG use may be effective at lowering odds of CIG or dual use, especially for those with higher levels of HL. Further research is needed to examine how to best disseminate information regarding the health risks of ECIGs.

Association between caregiver depression and child after-school program participation.

Depressive symptoms in parents and caregivers to children are associated with adverse biopsychosocial outcomes for caregivers themselves and the children in their custody. Higher overall and parenting-related stress, including stress over children's unsupervised after-school time, is associated with increased caregiver depression risk. Child after-school program participation is a form of social support that may mitigate parenting-related stress and reduce caregiver depression risk. This study tested for the association between child after-school program participation and caregiver depression in a sample of 486 caregivers in Richmond, Virginia. Child after-school program participation was associated with a significant reduction in the likelihood of a past caregiver depression diagnosis (OR = 0.58, 95% CI = 0.39 - 0.86, p = 0.007). This relationship remained significant after adjusting for the influence of caregiver anxiety, stress, financial hardship, and sociodemographic characteristics (OR = 0.49, 95% CI = 0.27 - 0.86, p = 0.015). Child after-school program participation may function as a protective factor that reduces caregiver depression risk. More research is needed to determine whether the observed association is causal in nature and dosage dependent. Findings from this and future studies may be used to inform evaluation of the impact of after-school programs at the family-level.

The association between internalizing and externalizing severity with current use of cigarettes, e-cigarettes, and alcohol in adults: Wave 1 of the Population Assessment of Tobacco and Health (PATH) study.

INTRODUCTION: Concurrent tobacco/alcohol use is common in adults, and associated with the severity of symptoms experienced by those with mental health disorders. However, few studies have explored this relationship across different combinations of tobacco products [i.e., conventional cigarette (CC) and electronic cigarette (EC)] and alcohol. METHODS: Data from the Wave 1 (2013-2014) Population Assessment of Tobacco and Health study were used. A total of 15,947 adults aged 18 years or older with complete study information were included. Multinomial logistic regression analyses were performed to determine the relationship between lifetime internalizing/externalizing severity and past 30-day use of tobacco and alcohol, adjusting for nicotine dependence (ND), sex, age, race, education, and income. RESULTS: Internalizing severity was more strongly associated with CC and alcohol use (moderate AOR = 1.47, 95% CI = 1.22-1.77; high AOR = 1.29, 95% CI = 1.03-1.61) as well as alcohol-exclusive use (moderate AOR = 1.58, 95% CI = 1.27-1.96; high AOR = 1.31, 95% CI = 1.05-1.64) while externalizing severity was more strongly associated with EC and alcohol use (high AOR = 2.97, 95% CI = 1.84-4.81, moderate AOR = 2.29, 95% CI = 1.53-3.43) when accounting for ND compared to none. The relationship between externalizing severity with EC use was dependent on alcohol being used with EC. CONCLUSIONS: The associations between psychopathology (internalizing vs. externalizing severity) varies by different combinations of alcohol, CC, and EC. Further, these relationships may be mediated through ND. Future investigations into the comorbidity between mental disorder symptoms with tobacco and alcohol use should consider use of specific substances as well as their combination.

Developing community-based health education strategies with family history: Assessing the association between community resident family history and interest in health education.

BACKGROUND: Family history (FH) is an underutilized genetically informative tool that can influence disease prevention and treatment. It is unclear how FH fits into the development of community-based health education. This study examines the role that FH plays in perceived threat and health education related to mental and chronic physical conditions in the context of the health belief model. METHODS: Data were collected from 1,048 adult participants aged 18-90 years. Approximately 76% of participants indicated African-American race/ethnicity and 35% had less than high school level education. Self-report data were collected on FH of four disorders: anxiety, depression, diabetes, and high blood pressure. Interest in receiving information regarding prevention as well as future testing efforts was assessed broadly. A series of logistic regressions examined the association between FH for each of the disorders and interest in receiving information on (1) prevention of diseases in general and (2) testing for diseases in general. These associations were also analyzed after accounting for the influence of perceived threat of conditions. RESULTS: Interest in receiving general health education was significantly associated with FH of depression (OR = 2.72, 95% CI = 1.74-4.25), anxiety (OR = 2.26, 95% CI = 1.45-3.22), and high blood pressure (OR = 2.54, 95% CI = 1.05-6.12). After adjustment for perceived threat, the magnitude of these associations was reduced substantially. The associations between perceived threat and either interest in receiving information on disease testing or receiving general health education were strong and significant across all conditions (OR = 2.11-3.74). DISCUSSION: These results provide evidence that perceived threat mediates the association between FH and engagement with health education. Currently available health education programs may benefit from considering the role of FH in an individual's motivation for participation in health education activities alongside other factors.

Assessment of Co-Occurring Substance Use During Opiate Treatment Programs in the United States.

The effectiveness of opiate treatment programs (OTPs) can be significantly influenced by co-occurring substance use, yet there are no standardized guidelines for assessing the influence of co-occurring substance use on treatment outcomes. In this review, we aim to provide an overview on the status of the assessment of co-occurring substance use during participation in OTPs in the United States. We searched 4 databases-MEDLINE/PubMed, EMBASE, PsychINFO, and the Cumulative Index to Nursing and Allied Health Literature (CINAHL)-from database inception to November 2018 to select relevant publications on OTPs that assessed participants' co-occurring substance use. We used a standardized protocol to extract study, intervention, and co-occurring substance use characteristics. Methodological quality was assessed using the Quality in Prognosis Studies tool. Of the 3,219 titles screened, 614 abstracts and 191 full-text original publications were assessed, leaving 85 eligible articles. Co-occurring substance use was most often assessed during opioid treatments using combined (pharmacological and behavioral) (n = 57 studies) and pharmacological (n = 25 studies) interventions. Cocaine, alcohol, marijuana, and benzodiazepines were frequently measured, while amphetamines and tobacco were rarely assessed. Great variation existed between studies in the timing and measurement of co-occurring substance use, as well as definitions for substances and polysubstance/polydrug use. Inconsistencies in the investigation of co-occurring substance use make comparison of results across studies challenging. Standardized measures and consensus on research on co-occurring substance use is needed to produce the evidence required to develop personalized treatment programs for persons using multiple substances and to inform best-practice guidelines for addressing polydrug use during participation in OTPs.

The Influence of Co-Occurring Substance Use on the Effectiveness of Opiate Treatment Programs According to Intervention Type.

This systematic review describes the influence of co-occurring substance use on the effectiveness of opiate treatment programs. MEDLINE/PubMed, Embase, PsychINFO, and the Cumulative Index to Nursing and Allied Health Literature were searched from database inception to November 28, 2018, to identify eligible opioid treatment studies in the United States that assessed the relationship between co-occurring substance use and treatment outcome (i.e., opioid abstinence and treatment retention). A total of 34 eligible studies were included. Overall, co-occurring substance use was associated with negative treatment outcomes regardless of intervention type. However, patterns varied by substance and intervention type. In particular, co-occurring use of cocaine or marijuana with opioids was associated with reduced treatment retention and opioid abstinence regardless of intervention type. Co-occurring use of amphetamines, compared with no use or reduced use of amphetamines, decreased treatment retention. Co-occurring use of alcohol was both positively and negatively associated with treatment outcomes. One study reported a significant positive association between sedative use and opioid abstinence. Generally, findings suggest that combined interventions reported better health outcomes compared with pharmacological or behavioral intervention studies alone. The findings of this review emphasize the need to comprehensively study and address co-occurring substance use to improve opiate treatment programs.

Nurturing the Student, Sustaining the Mission: 20 Years of the International/Inner-City/Rural Preceptorship Program.

BACKGROUND AND OBJECTIVES: Specialized medical school educational tracks aim to increase the primary care workforce. The International/Inner-City/Rural Preceptorship (I2CRP) Program is unique in addressing multiple communities, a large cohort and applying the Self Determination Theory framework. This study examined program impact by analyzing the numbers of graduates matched into primary care and practicing in medically underserved communities. METHODS: We compared the match list of I2CRP graduates between 2000 and 2017 (n=204) to non-I2CRP Virginia Commonwealth University School of Medicine (VCU SOM) graduates (n=3,037). We analyzed the matches into primary care, National Health Service Corps (NHSC) priority specialties, and NHSC priority plus general surgery. We searched a federal database to determine which graduates are practicing in workforce shortage areas. RESULTS: Many more I2CRP graduates matched to primary care (71.1%), compared to non-I2CRP graduates (38.2%; P<.001). Within primary care, I2CRP graduates matched to family medicine more frequently than non-I2CRP graduates (36.3% vs 8.4%). Eighteen percent of posttraining I2CRP graduates work in rural areas and 41% work in medically underserved areas. CONCLUSIONS: I2CRP graduates are more likely to match to family medicine and primary care. I2CRP curriculum nurtures new medical students' interest in primary care, and self-determination theory provides a framework to organize the program curriculum. The program's impact endures as evidenced by participants' continued work in underserved areas after residency. Increasing support for such programs may help address the primary care physician shortage in medically underserved areas.

Developing Comprehensive Strategies to Evaluate Medical School Curricula.

Evaluation of medical school curriculum is important to document outcomes, effectiveness of learning, engagement in quality improvement, and to meet accreditation compliance. This monograph provides a roadmap and resource for medical schools to meaningfully evaluate their curriculum based on specific metrics. The method of evaluation includes an examination of Kirkpatrick's levels of outcomes including reactions, learning, behavior, and impact. It is important that student outcomes are mapped in relation to curricular objectives. There are specific outcomes that may be utilized to determine if the curriculum has met the institution's goals. The first is comparison to national metrics (United States Medical Licensing Examinations and American Association of Medical Colleges Graduation Questionnaire). Second, medical schools collect internal program metrics, which include specific student performance metrics, such as number of students graduating, attrition, and matching to specialty. Further, schools may examine student performance and surveys in the preclerkship and clinical phases (e.g., grades, failing courses, survey responses about the curriculum), including qualitative responses on surveys or focus groups. As the learning environment is critical to learning, a deep dive to understand the environment and mistreatment may be important for program evaluation. This may be performed by specifically examining the Graduation Questionnaire, internal surveys, and mistreatment reporting. Finally, there are numerous attitudinal instruments that may help medical schools understand their students' development at one point or over time. These include measurements of stress, wellness, burnout, lifelong learning, and attitudes toward patient safety. Together, examining the composite of outcomes helps to understand and improve the medical school curriculum.