A data-driven approach to implementing the HPTN 094 complex intervention INTEGRA in local communities.

BACKGROUND: HIV burden in the US among people who inject drugs (PWID) is driven by overlapping syndemic factors such as co-occurring health needs and environmental factors that synergize to produce worse health outcomes among PWID. This includes stigma, poverty, and limited healthcare access (e.g. medication to treat/prevent HIV and for opioid use disorder [MOUD]). Health services to address these complex needs, when they exist, are rarely located in proximity to each other or to the PWID who need them. Given the shifting drug use landscapes and geographic heterogeneity in the US, we evaluate a data-driven approach to guide the delivery of such services to PWID in local communities. METHODS: We used a hybrid, type I, embedded, mixed method, data-driven approach to identify and characterize viable implementation neighborhoods for the HPTN 094 complex intervention, delivering integrated MOUD and HIV treatment/prevention through a mobile unit to PWID across five US cities. Applying the PRISM framework, we triangulated geographic and observational pre-implementation phase data (epidemiological overdose and HIV surveillance data) with two years of implementation phase data (weekly ecological assessments, study protocol meetings) to characterize environmental factors that affected the viability of implementation neighborhoods over time and across diverse settings. RESULTS: Neighborhood-level drug use and geographic diversity alongside shifting socio-political factors (policing, surveillance, gentrification) differentially affected the utility of epidemiological data in identifying viable implementation neighborhoods across sites. In sites where PWID are more geographically dispersed, proximity to structural factors such as public transportation and spaces where PWID reside played a role in determining suitable implementation sites. The utility of leveraging additional data from local overdose and housing response systems to identify viable implementation neighborhoods was mixed. CONCLUSIONS: Our findings suggest that data-driven approaches provide a contextually relevant pragmatic strategy to guide the real-time implementation of integrated care models to better meet the needs of PWID and help inform the scale-up of such complex interventions. This work highlights the utility of implementation science methods that attend to the impact of local community environmental factors on the implementation of complex interventions to PWID across diverse drug use, sociopolitical, and geographic landscapes in the US. TRIAL REGISTRATION: ClincalTrials.gov, Registration Number: NCT04804072 . Registered 18 February 2021.

Leveraging pooled medical examiner records to surveil complex and emerging patterns of polysubstance use in the United States.

BACKGROUND: The United States (US) is an extreme global outlier for drug-related death rates. However, data describing drug-related deaths are generally available only on an 8-13-month lag. Furthermore, granular details about substance-involvement are often not available, which particularly stymies efforts to track fatal polysubstance and novel psychoactive substance use. Detailed medical examiner records provide a powerful source of information for drug-related death surveillance, but have been underutilized. METHODS: We pooled medical examiner data from five US states and 14 counties that together comprise 18% of the US population to examine demographic, geographic, and drug-specific trends in polysubstance drug-related deaths. We employed mixed effects logistic regression to identify demographic factors associated with polysubstance rather than single substance drug-related deaths. We assessed the correlations between drug classes and described geographic variation in the prevalence of specific drugs and the presence of novel and emerging psychoactive substances. RESULTS: Our sample included 73,077 drug-related deaths from 2012 through early 2022. Nearly two-thirds of drug-related deaths were polysubstance-involved, with the number and percentage growing annually. High percentages of polysubstance drug-related deaths were observed in both urban and rural jurisdictions. After adjusting for year and jurisdiction, female, American Indian and Alaska Native, and White individuals had the most elevated odds of polysubstance drug-related deaths. Drug-related deaths involving benzodiazepines or opioids, whether pharmaceutical or illicit, and other pharmaceutical drugs were most likely to have polysubstance involvement, while methamphetamine-involved deaths were least likely to involve multiple substances. Strong correlations were observed between prescription opioids and prescription benzodiazepines, fentanyl and xylazine, and designer benzodiazepines and novel synthetic opioids. CONCLUSIONS: Analysis of detailed medical examiner records reveals the breadth and complexity of polysubstance drug-related deaths in the US. Future efforts to use this unique resource can improve population-based surveillance of drug-related deaths to better tailor interventions and solutions to this critical health crisis.

Hepatitis C Treatment Among Primary Care and Specialty Providers: A Single Center Study, 2015 to 2022.

INTRODUCTION: Despite national goals to eliminate Hepatitis C (HCV) and the advancement of curative, well-tolerated direct-acting antiviral (DAAs) regimens, rates of HCV treatment have declined nationally since 2015. Current HCV guidelines encourage treatment of HCV by primary care providers (PCPs). Payors have reduced restrictions to access DAAs nationally and in California however it remains unclear if the removal of these restrictions has impacted the proportion of PCPs prescribing DAAs at a health system level. Our objective was to examine the proportion of DAAs prescribed by PCPs and specialists and to describe the population receiving treatment in a single health system from 2015 to 2022. METHODS: We examined the proportion of DAAs prescribed by PCPs and specialists and the population receiving treatment through a retrospective analysis of claims data in the University of California, Los Angeles (UCLA) Health System from 2015 to 2022. We described number of prescriptions for HCV medication prescribed by PCPs and specialists by year, medication type, and physician specialty. We also described numbers of prescriptions by patient demographics and comorbidities. RESULTS: A total of 1515 adult patients received a prescription for HCV medication through the UCLA Health System between 2015 and 2022. The proportion of patients receiving prescriptions for PCPs peaked at 19% in 2016, yet decreased to 5.7% in 2022, an average of 13% across all years. Median age of patients receiving treatment was 60 years old, and 56% of patients receiving HCV treatment had commercial insurance as their primary payer. CONCLUSIONS: HCV treatment declined from 2015 to 2022 among specialists and PCPs in our health system. Older patients comprised the majority of patients receiving treatment, suggesting a need for novel approaches to reach patients under 40, an age group with significant increases in HCV transmission.

Perspectives on benefits and risks of creation of an "injection drug use" billing code.

People with substance use disorder (SUD) face barriers to prevention and treatment services, increasing risk for hospitalization and death. Injection drug use (IDU) can lead to an increased risk of overdose and infections. However, identifying people who inject drugs (PWID) within healthcare systems is challenging. International Classification of Disease (ICD-10) codes are used for billing and tracking healthcare utilization. In this commentary, experts in the field weigh the benefits and risks of creating an IDU-specific ICD-10 code. Potential benefits include earlier identification, better access to health services, and improved systems of resource allocation. Potential risks include further stigmatization of PWID and, if not tied to financial reimbursement, low rates of code utilization. As the current systems of identifying PWID are lacking, we feel that a guided operationalization of an ICD code to identify PWID could improve quantitative and epidemiological research accuracy and, therefore, support the health and well-being of PWID.

Impact of Potential Case Misclassification by Administrative Diagnostic Codes on Outcome Assessment of Observational Study for People Who Inject Drugs.

Introduction: Initiation of medications for opioid use disorder (MOUD) within the hospital setting may improve outcomes for people who inject drugs (PWID) hospitalized because of an infection. Many studies used International Classification of Diseases (ICD) codes to identify PWID, although these may be misclassified and thus, inaccurate. We hypothesized that bias from misclassification of PWID using ICD codes may impact analyses of MOUD outcomes. Methods: We analyzed a cohort of 36 868 cases of patients diagnosed with Staphylococcus aureus bacteremia at 124 US Veterans Health Administration hospitals between 2003 and 2014. To identify PWID, we implemented an ICD code-based algorithm and a natural language processing (NLP) algorithm for classification of admission notes. We analyzed outcomes of prescribing MOUD as an inpatient using both approaches. Our primary outcome was 365-day all-cause mortality. We fit mixed-effects Cox regression models with receipt or not of MOUD during the index hospitalization as the primary predictor and 365-day mortality as the outcome. Results: NLP identified 2389 cases as PWID, whereas ICD codes identified 6804 cases as PWID. In the cohort identified by NLP, receipt of inpatient MOUD was associated with a protective effect on 365-day survival (adjusted hazard ratio, 0.48; 95% confidence interval, .29-.81; P < .01) compared with those not receiving MOUD. There was no significant effect of MOUD receipt in the cohort identified by ICD codes (adjusted hazard ratio, 1.00; 95% confidence interval, .77-1.30; P = .99). Conclusions: MOUD was protective of all-cause mortality when NLP was used to identify PWID, but not significant when ICD codes were used to identify the analytic subjects.

Delivering integrated strategies from a mobile unit to address the intertwining epidemics of HIV and addiction in people who inject drugs: the HPTN 094 randomized controlled trial protocol (the INTEGRA Study).

BACKGROUND: Persons with opioid use disorders who inject drugs (PWID) in the United States (US) face multiple and intertwining health risks. These include interference with consistent access, linkage, and retention to health care including medication for opioid use disorder (MOUD), HIV prevention using pre-exposure prophylaxis (PrEP), and testing and treatment for sexually transmitted infections (STIs). Most services, when available, including those that address substance misuse, HIV prevention, and STIs, are often provided in multiple locations that may be difficult to access, which further challenges sustained health for PWID. HPTN 094 (INTEGRA) is a study designed to test the efficacy of an integrated, "whole-person" strategy that provides integrated HIV prevention including antiretroviral therapy (ART), PrEP, MOUD, and STI testing and treatment from a mobile health delivery unit ("mobile unit") with peer navigation compared to peer navigation alone to access these services at brick and mortar locations. METHODS: HPTN 094 (INTEGRA) is a two-arm, randomized controlled trial in 5 US cities where approximately 400 PWID without HIV are assigned either to an experimental condition that delivers 26 weeks of "one-stop" integrated health services combined with peer navigation and delivered in a mobile unit or to an active control condition using peer navigation only for 26 weeks to the same set of services delivered in community settings. The primary outcomes include being alive and retained in MOUD and PrEP at 26 weeks post-randomization. Secondary outcomes measure the durability of intervention effects at 52 weeks following randomization. DISCUSSION: This trial responds to a need for evidence on using a "whole-person" strategy for delivering integrated HIV prevention and substance use treatment, while testing the use of a mobile unit that meets out-of-treatment PWID wherever they might be and links them to care systems and/or harm reduction services. Findings will be important in guiding policy for engaging PWID in HIV prevention or care, substance use treatment, and STI testing and treatment by addressing the intertwined epidemics of addiction and HIV among those who have many physical and geographic barriers to access care. TRIAL REGISTRATION: ClinicalTrials.gov NCT04804072 . Registered on 18 March 2021.

Decreasing Stigma Toward People Who Inject Drugs: Harm Reduction Training for First-Year Medical Students.

PROBLEM: Stigma in health care toward people who inject drugs (PWID) is a well-described, significant barrier to quality care, resulting in poor health outcomes. Harm reduction offers a person-centered counter-framework for minimizing harm for people who use drugs. Despite the evidence in support of harm reduction, medical students typically receive minimal training on harm reduction and the care of PWID. APPROACH: To fill this gap, medical students at the University of California, Los Angeles organized around the principles of harm reduction to improve the medical school curriculum related to PWID. Students screened lectures for stigmatizing language and collaborated with faculty to improve lecture materials. They partnered with a community organizer and hosted a mandatory 1-hour lecture and 30-minute discussion introducing the principles of harm reduction within an overdose prevention, recognition, and response training for first-year medical students during medical school orientation in August 2022. An anonymous online pretest and posttest survey, assessing student attitudes toward PWID, was used to evaluate the effects of the training. OUTCOMES: A total of 156 students completed the pretest survey, and 107 students completed the pretest and posttest survey (68.5% response rate). The overall posttest mean stigma score was 1.8 (standard deviation [SD] = 0.5) and was significantly lower than the pretest mean of 2.1 (SD = 0.7; P < .0001), indicating a reduction in stigma among medical student attitudes after the course. There was statistically significant improvement in attitudes for 7 of 13 component measures. NEXT STEPS: This analysis demonstrated that the mandatory class has the capacity to improve medical student attitudes toward PWID. The authors plan to further evaluate the program's effectiveness through measuring and reporting outcomes for future student cohorts. The authors are working with curriculum directors to further incorporate harm reduction principles into other lectures and problem-based learning exercises.

Longitudinal changes in co-involved drugs, comorbidities, and demographics of methamphetamine-related deaths in Los Angeles County.

INTRODUCTION: We conducted a population-based observational study of all medical examiner cases in Los Angeles County from January 2012 through June 2021 in which methamphetamine was listed as a cause of or contributing factor to death (n = 6125). We aimed to characterize demographics, comorbidities, and co-involved substances in methamphetamine-related deaths longitudinally in Los Angeles County, California. METHODS: We used detailed death record data to manually classify fatalities by involvement of each organ system, opioids, alcohol, cocaine, other drugs or medications, and external/traumatic causes. Primary outcomes included: the number of methamphetamine-involved deaths, demographics of decedents, percentage of methamphetamine deaths also involving other drugs, and percentage of methamphetamine deaths involving different organ systems. We performed Mann Kendall tests of trends to identify statistically significant longitudinal changes. RESULTS: During the study period, the percentage of methamphetamine-related deaths involving opioids significantly increased from 16 % in 2012 to 54 % in 2021 (p < 0.001). Concurrently, the percentage involving cardiovascular causes significantly decreased from 47 % to 26 % (p < 0.05). Methamphetamine-related deaths in LAC increasingly affected people experiencing homelessness, for whom the percentage tripled from 13 % in 2012 to 35 % in 2021. The share of decedents under 40 years old increased from 33 % to 41 %. The percentage of Black or African American decedents increased over five-fold from 3 % to 17 %. CONCLUSIONS: Methamphetamine-related deaths involving opioids more than tripled in Los Angeles County from 2012 to 2021, reflecting the drug supply's shift to illicit fentanyl. More than a quarter involved cardiovascular causes. These findings have implications for treatment and prevention, including scaling up contingency management, distributing naloxone to people who primarily use stimulants, and including cardiovascular care alongside these interventions directly targeted to reduce harms of methamphetamine use.

Implementing a decentralized opioid overdose prevention strategy in Mexico, a pending public policy issue.

The public health crisis due to opioid overdose is worsening in Mexico's northern region due to the introduction of illicitly manufactured fentanyl into the local drug supply. Though there is an increase in overdose deaths, there is no accurate report of overdoses by Mexican government agencies and no comprehensive opioid overdose prevention strategy. There is currently only an anti-drug marketing strategy which is likely insufficient to mitigate the growing epidemic. In order to address the growing opioid overdose crisis in the country, it is necessary to create and implement a decentralized prevention strategy, that includes naloxone distribution, expanded treatment services in regions most in need, and create active dialogue with community organisations already implementing harm reduction actions. Decisive action must be taken by the Mexican government to ensure the health and wellbeing of the Mexican citizens, especially those at high risk for opioid overdose.

Intersections between syndemic conditions and stages along the continuum of overdose risk among women who inject drugs in Mexicali, Mexico.

BACKGROUND: Research on women who inject drugs is scarce in low- and middle-income countries. Women experience unique harms such as sexism and sexual violence which translate into negative health outcomes. The present work aims to provide insight into the experiences of women who inject drugs at the US-Mexico border to identify social and health-related risk factors for overdose to guide harm reduction interventions across the Global South. METHODS: We recruited 25 women ≥ 18 years of age accessing harm reduction and sexual health services at a non-governmental harm reduction organization, "Verter", in Mexicali, Mexico. We employed purposeful sampling to recruit women who inject drugs who met eligibility criteria. We collected quantitative survey data and in-depth interview data. Analyses of both data sources involved the examination of descriptive statistics and thematic analysis, respectively, and were guided by the syndemic and continuum of overdose risk frameworks. RESULTS: Survey data demonstrated reports of initiating injection drug use at a young age, experiencing homelessness, engaging in sex work, being rejected by family members, experiencing physical violence, injecting in public spaces, and experiencing repeated overdose events. Interview data provided evidence of stigma and discrimination toward women, a lack of safe spaces and support systems, risk of overdose-related harms, sexual violence, and the overall need for harm reduction services. CONCLUSION: Women who inject drugs in Mexicali describe experiences of violence, overdose, and public injecting. Women are particularly vulnerable in the Mexicali context, as this area faces a noticeable lack of health and social services. Evidenced-based harm reduction strategies such as safe consumption sites and overdose prevention strategies (e.g., naloxone distribution and training) may benefit this population. Evidence from local organizations could help close the gap in service provision in low-resource settings like Mexico, where government action is almost nonexistent.

Fentanyl, heroin, and methamphetamine-based counterfeit pills sold at tourist-oriented pharmacies in Mexico: An ethnographic and drug checking study.

BACKGROUND: Fentanyl- and methamphetamine-based counterfeit prescription drugs have driven escalating overdose death rates in the US, however their presence in Mexico has not been assessed. Our ethnographic team has conducted longitudinal research focused on illicit drug markets in Northern Mexico since 2018. In 2021-2022, study participants described the arrival of new, unusually potent tablets sold as ostensibly controlled substances, without a prescription, directly from pharmacies that cater to US tourists. AIMS: To characterize the availability of counterfeit and authentic controlled substances at pharmacies in Northern Mexico available to English-speaking tourists without a prescription. METHODS: We employed an iterative, exploratory, mixed methods design. Longitudinal ethnographic data was used to characterize tourist-oriented micro-neighborhoods and guide the selection of n=40 pharmacies in n=4 cities in Northern Mexico. In each pharmacy, samples of "oxycodone", "Xanax", and "Adderall" were sought as single pills, during English-language encounters, after which detailed ethnographic accounts were recorded. We employed immunoassay-based testing strips to check each pill for the presence of fentanyls, benzodiazepines, amphetamines, and methamphetamines. We used Fourier-Transform Infrared Spectroscopy to further characterize drug contents. RESULTS: Of n=40 pharmacies, one or more of the requested controlled substances could be obtained with no prescription (as single pills or in bottles) at 28 (70.0%) and as single pills at 19 (47.5%). Counterfeit pills were obtained at 11 pharmacies (27.5%). Of n=45 samples sold as one-off controlled substances, 18 were counterfeit. 7 of 11 (63.6%) samples sold as "Adderall" contained methamphetamine, 8 of 27 (29.6%) samples sold as "Oxycodone" contained fentanyl, and 3 "Oxycodone" samples contained heroin. Pharmacies providing counterfeit drugs were uniformly located in tourist-serving micro-neighborhoods, and generally featured English-language advertisements for erectile dysfunction medications and "painkillers". Pharmacy employees occasionally expressed concern about overdose risk and provided harm reduction guidance. DISCUSSION: The availability of fentanyl-, heroin-, and methamphetamine-based counterfeit medications in tourist-oriented independent pharmacies in Northern Mexico represents a public health risk, and occurs in the context of 1) the normalization of medical tourism as a response to rising unaffordability of healthcare in the US, 2) plummeting rates of opioid prescription in the US, affecting both chronic pain patients and the availability of legitimate pharmaceuticals on the unregulated market, 3) the rise of fentanyl-based counterfeit opioids as a key driver of the fourth, and deadliest-to-date, wave of the opioid crisis. It was not possible to distinguish counterfeit medications based on appearance of pills or geography of pharmacies, because identically-appearing authentic and counterfeit versions were often sold in close geographic proximity. Nevertheless, people who consume drugs may be more trusting of controlled substances purchased directly from pharmacies. Due to Mexico's limited opioid overdose surveillance infrastructure, the current death rate from these substances remains unknown.

US drug overdose mortality rose faster among hispanics than non-hispanics from 2010 to 2021.

BACKGROUND: Historically, overdose mortality rates among Hispanics have been lower than non-Hispanics. The purpose of this analysis was to characterize the U.S. overdose crisis among Hispanics compared to non-Hispanics. METHODS: We used the Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiological Research (WONDER) platform to obtain drug overdose mortality rates per 100,000 population between 2010 and 2021 for Hispanics and non-Hispanics. We examined the relative percent change and specific drug involvement (2010-2021) and state-level disparities (2010-2020) among Hispanics versus non-Hispanics. We calculated rate ratios by state and annual percent change in total and for each specific drug. Statistical analyses were performed using R software version 4.0.3 (R Project for Statistical Computing). RESULTS: Nationally, from 2010 to 2021, Hispanic overdose rates rose from 5.6 to 21.7 per 100,000, an increase of 287.5 % compared to 13.5-35.1 per 100,000, an increase of 160 % among non-Hispanics. The average annual percent change was 12 % for Hispanics and 9 % for non-Hispanics. The three most common drug classes involved in overdose deaths among both groups included: Fentanyls and synthetic opioids; cocaine; and prescription opioids. Hispanic overdose rates were higher than non-Hispanic rates in New Mexico, Colorado, Massachusetts, and Pennsylvania in 2020, versus only Michigan in 2010. CONCLUSIONS: We observed disparities in overdose mortality growth among Hispanics compared to non-Hispanics from 2010 to 2021. These disparities highlight the urgency to develop community-centered solutions that take into consideration the social and structural inequalities that exacerbate the effects of the opioid overdose crisis on Hispanic communities.

Fentanyl, Heroin, and Methamphetamine-Based Counterfeit Pills Sold at Tourist-Oriented Pharmacies in Mexico: An Ethnographic and Drug Checking Study.

Background: Fentanyl- and methamphetamine-based counterfeit prescription drugs have driven escalating overdose death rates in the US, however their presence in Mexico has not been assessed. Our ethnographic team has conducted longitudinal research focused on illicit drug markets in Northern Mexico since 2018. In 2021-2022, study participants described the arrival of new, unusually potent tablets sold as ostensibly controlled substances, without a prescription, directly from pharmacies that cater to US tourists. Aims: To characterize the availability of counterfeit and authentic controlled substances at pharmacies in Northern Mexico available to English-speaking tourists without a prescription. Methods: We employed an iterative, exploratory, mixed methods design. Longitudinal ethnographic data was used to characterize tourist-oriented micro-neighborhoods and guide the selection of n=40 pharmacies in n=4 cities in Northern Mexico. In each pharmacy, samples of "oxycodone", "Xanax", and "Adderall" were sought as single pills, during English-language encounters, after which detailed ethnographic accounts were recorded. We employed immunoassay-based testing strips to check each pill for the presence of fentanyls, benzodiazepines, amphetamines, and methamphetamines. We used Fourier-Transform Infrared Spectroscopy to further characterize drug contents. Results: Of n=40 pharmacies, one or more of the requested controlled substance could be obtained with no prescription (as single pills or in bottles) at 28 (70.0%) and as single pills at 19 (47.5%). Counterfeit pills were obtained at 11 pharmacies (27.5%). Of n=45 samples sold as one-off controlled substances, 18 were counterfeit. 7 of 11 (63.6%) samples sold as "Adderall" contained methamphetamine, 8 of 27 (29.6%) samples sold as "Oxycodone" contained fentanyl, and 3 "Oxycodone" samples contained heroin. Pharmacies providing counterfeit drugs were uniformly located in tourist-serving micro-neighborhoods, and generally featured English-language advertisements for erectile dysfunction medications and "painkillers". Pharmacy employees occasionally expressed concern about overdose risk and provided harm reduction guidance. Discussion: The availability of fentanyl-, heroin-, and methamphetamine-based counterfeit medications in tourist-oriented independent pharmacies in Northern Mexico represents a public health risk, and occurs in the context of 1) the normalization of medical tourism as a response to rising unaffordability of healthcare in the US, 2) plummeting rates of opioid prescription in the US, affecting both chronic pain patients and the availability of legitimate pharmaceuticals on the unregulated market, 3) the rise of fentanyl-based counterfeit opioids as a key driver of the fourth, and deadliest-to-date, wave of the opioid crisis. It was not possible to distinguish counterfeit medications based on appearance of pills or geography of pharmacies, because identically-appearing authentic and counterfeit versions were often sold in close geographic proximity. Nevertheless, drug consumers may be more trusting of controlled substances purchased directly from pharmacies. Due to Mexico's limited opioid overdose surveillance infrastructure, the current death rate from these substances remains unknown.

Long-Acting Injectable Therapy for People with HIV: Looking Ahead with Lessons from Psychiatry and Addiction Medicine.

Long-acting injectable antiretroviral medications are new to HIV treatment. People with HIV may benefit from a treatment option that better aligns with their preferences, but could also face new challenges and barriers. Authors from the fields of HIV, substance use treatment, and mental health collaborated on this commentary on the issues surrounding equitable implementation and uptake of LAI ART by drawing lessons from all three fields. We employ a socio-ecological framework beginning at the policy level and moving through the community, organizational, interpersonal, and patient levels. We look at extant literature on the topic as well as draw from the direct experience of our clinician-authors.

Natural Language Processing and Machine Learning to Identify People Who Inject Drugs in Electronic Health Records.

Background: Improving the identification of people who inject drugs (PWID) in electronic medical records can improve clinical decision making, risk assessment and mitigation, and health service research. Identification of PWID currently consists of heterogeneous, nonspecific International Classification of Diseases (ICD) codes as proxies. Natural language processing (NLP) and machine learning (ML) methods may have better diagnostic metrics than nonspecific ICD codes for identifying PWID. Methods: We manually reviewed 1000 records of patients diagnosed with Staphylococcus aureus bacteremia admitted to Veterans Health Administration hospitals from 2003 through 2014. The manual review was the reference standard. We developed and trained NLP/ML algorithms with and without regular expression filters for negation (NegEx) and compared these with 11 proxy combinations of ICD codes to identify PWID. Data were split 70% for training and 30% for testing. We calculated diagnostic metrics and estimated 95% confidence intervals (CIs) by bootstrapping the hold-out test set. Best models were determined by best F-score, a summary of sensitivity and positive predictive value. Results: Random forest with and without NegEx were the best-performing NLP/ML algorithms in the training set. Random forest with NegEx outperformed all ICD-based algorithms. F-score for the best NLP/ML algorithm was 0.905 (95% CI, .786-.967) and 0.592 (95% CI, .550-.632) for the best ICD-based algorithm. The NLP/ML algorithm had a sensitivity of 92.6% and specificity of 95.4%. Conclusions: NLP/ML outperformed ICD-based coding algorithms at identifying PWID in electronic health records. NLP/ML models should be considered in identifying cohorts of PWID to improve clinical decision making, health services research, and administrative surveillance.

Development and Validation of Machine Models Using Natural Language Processing to Classify Substances Involved in Overdose Deaths.

Importance: Overdose is one of the leading causes of death in the US; however, surveillance data lag considerably from medical examiner determination of the death to reporting in national surveillance reports. Objective: To automate the classification of deaths related to substances in medical examiner data using natural language processing (NLP) and machine learning (ML). Design, Setting, and Participants: Diagnostic study comparing different natural language processing and machine learning algorithms to identify substances related to overdose in 10 health jurisdictions in the US from January 1, 2020, to December 31, 2020. Unstructured text from 35 433 medical examiner and coroners' death records was examined. Exposures: Text from each case was manually classified to a substance that was related to the death. Three feature representation methods were used and compared: text frequency-inverse document frequency (TF-IDF), global vectors for word representations (GloVe), and concept unique identifier (CUI) embeddings. Several ML algorithms were trained and best models were selected based on F-scores. The best models were tested on a hold-out test set and results were reported with 95% CIs. Main Outcomes and Measures: Text data from death certificates were classified as any opioid, fentanyl, alcohol, cocaine, methamphetamine, heroin, prescription opioid, and an aggregate of other substances. Diagnostic metrics and 95% CIs were calculated for each combination of feature extraction method and machine learning classifier. Results: Of 35 433 death records analyzed (decedent median age, 58 years [IQR, 41-72 years]; 24 449 [69%] were male), the most common substances related to deaths included any opioid (5739 [16%]), fentanyl (4758 [13%]), alcohol (2866 [8%]), cocaine (2247 [6%]), methamphetamine (1876 [5%]), heroin (1613 [5%]), prescription opioids (1197 [3%]), and any benzodiazepine (1076 [3%]). The CUI embeddings had similar or better diagnostic metrics compared with word embeddings and TF-IDF for all substances except alcohol. ML classifiers had perfect or near perfect performance in classifying deaths related to any opioids, heroin, fentanyl, prescription opioids, methamphetamine, cocaine, and alcohol. Classification of benzodiazepines was suboptimal using all 3 feature extraction methods. Conclusions and Relevance: In this diagnostic study, NLP/ML algorithms demonstrated excellent diagnostic performance at classifying substances related to overdoses. These algorithms should be integrated into workflows to decrease the lag time in reporting overdose surveillance data.

COVID-19 needs no passport: the interrelationship of the COVID-19 pandemic along the U.S.-Mexico border.

OBJECTIVES: To investigate the impact of the COVID-19 pandemic along the U.S.-Mexico border region and evaluate the relationship of COVID-19 related mortality, socioeconomic status, and vaccination. METHODS: We used indirect standardization to age-adjust mortality rates and calculate standardized mortality ratios [SMR] in both countries. To examine the impact of socioeconomic factors, we calculated the Human Development Index (HDI) by county/municipality. We performed linear regression to understand the relationship between mortality, vaccination, and HDI. We used choropleth maps to visualize the trends seen in the region. RESULTS: Between January 22nd, 2020 and December 1st, 2021, surges of cases and deaths were similar in dyad cities along the U.S.-Mexico border visualizing the interconnectedness of the region. Mortality was higher in U.S. counties along the border compared to the national average (SMR 1.17, 95% CI 1.15-1.19). In Mexico, border counties had a slightly lower mortality to the national average (SMR 0.94, 95% CI 0.93-0.95). In U.S. border states, SMR was shown to negatively correlate with human development index (HDI), a socioeconomic proxy, resulting in a higher SMR in the border region compared to the rest of the counties. Conversely in Mexican border states, there was no association between SMR and HDI. Related to vaccination, U.S. counties along the border were vaccinated at a greater percentage than non-border counties and vaccination was negatively correlated with HDI. In Mexico, states along the border had a higher ratio of vaccinations per person than non-border states. CONCLUSIONS: The U.S.-Mexico border is a divide of incredible importance not only to immigration but as a region with unique social, economic, environmental, and epidemiological factors that impact disease transmission. We investigated how the COVID-19 pandemic followed trends of previously studied diseases in the corridor such as tuberculosis, HIV, and influenza H1N1. These data state how targeted intervention along the U.S.-Mexico border region is a necessity when confronting COVID-19 and have implications for future control of infectious diseases in the region.

Behavior change after fentanyl testing at a safe consumption space for women in Northern Mexico: A pilot study.

BACKGROUND: Fentanyl has led to an increased number of overdose deaths in North America. Testing substances for fentanyl may be a harm reduction strategy to prevent overdose. Little is known about behavior change after fentanyl testing and the attitudes around fentanyl knowledge and testing along the US-Mexico border in the context of a safe consumption site. METHODS: This was a pilot quantitative and qualitative study with 30 women who use drugs at an unsanctioned safe consumption site in Mexicali, Mexico. Women participated in a quantitative survey, a semi-structured interview, and fentanyl testing of substances. Injection behavior was observed after fentanyl testing results were provided. Qualitative data were collected to explore the meanings participants attributed to fentanyl and fentanyl testing. RESULTS: Half of the substances tested positive for fentanyl (n=15, 50%), and all of them were in samples of black tar heroin. Among those participants who tested positive for fentanyl, 7 (47%) subsequently used less of the intended substance, 1 did not use the intended substance, and 7 (47%) did not change their behavior (i.e., used as originally intended). In qualitative interviews, a predominant theme was a description of fentanyl as dangerous and deadly and fentanyl testing as being helpful for modifying drug use behaviors. However, participants recognized that there could be no change in behavior following a positive fentanyl test in the context of not being able to find substances free of fentanyl. CONCLUSION: We observed mixed results related to behavior change after women's intended substance for use tested positive for fentanyl. Fentanyl testing was acceptable to women, but behavior change was hampered by the inability to find substances free of fentanyl. Further research is needed to maximize the potential of fentanyl testing as a harm reduction tool especially in the context of a changing drug supply.