Evaluation of respondent-driven sampling in seven studies of people who use drugs from rural populations: findings from the Rural Opioid Initiative.

BACKGROUND: Accurate prevalence estimates of drug use and its harms are important to characterize burden and develop interventions to reduce negative health outcomes and disparities. Lack of a sampling frame for marginalized/stigmatized populations, including persons who use drugs (PWUD) in rural settings, makes this challenging. Respondent-driven sampling (RDS) is frequently used to recruit PWUD. However, the validity of RDS-generated population-level prevalence estimates relies on assumptions that should be evaluated. METHODS: RDS was used to recruit PWUD across seven Rural Opioid Initiative studies between 2018-2020. To evaluate RDS assumptions, we computed recruitment homophily and design effects, generated convergence and bottleneck plots, and tested for recruitment and degree differences. We compared sample proportions with three RDS-adjusted estimators (two variations of RDS-I and RDS-II) for five variables of interest (past 30-day use of heroin, fentanyl, and methamphetamine; past 6-month homelessness; and being positive for hepatitis C virus (HCV) antibody) using linear regression with robust confidence intervals. We compared regression estimates for the associations between HCV positive antibody status and (a) heroin use, (b) fentanyl use, and (c) age using RDS-1 and RDS-II probability weights and no weights using logistic and modified Poisson regression and random-effects meta-analyses. RESULTS: Among 2,842 PWUD, median age was 34 years and 43% were female. Most participants (54%) reported opioids as their drug of choice, however regional differences were present (e.g., methamphetamine range: 4-52%). Many recruitment chains were not long enough to achieve sample equilibrium. Recruitment homophily was present for some variables. Differences with respect to recruitment and degree varied across studies. Prevalence estimates varied only slightly with different RDS weighting approaches, most confidence intervals overlapped. Variations in measures of association varied little based on weighting approach. CONCLUSIONS: RDS was a useful recruitment tool for PWUD in rural settings. However, several violations of key RDS assumptions were observed which slightly impacts estimation of proportion although not associations.

Predictors of engagement in screening for a hepatitis C virus (HCV) treatment trial in a rural Appalachian community.

An HCV treatment trial was initiated in September 2019 to address the opioid/hepatitis C virus (HCV) syndemic in rural Kentucky. The focus of the current analysis is on participation in diagnostic screening for the trial. Initial eligibility (≥18 years of age, county resident) was established by phone followed by in-person HCV viremia testing. 900 rural residents met the inclusion criteria and comprised the analytic sample. Generalized linear models were specified to estimate the relative risk of non-attendance at the in-person visit determining HCV eligibility. Approximately one-quarter (22.1%) of scheduled participants were no-shows. People who inject drugs were no more likely than people not injecting drugs to be a no-show; however, participants ≤35 years of age were significantly less likely to attend. While the median time between phone screening and scheduled in-person screening was only 2 days, each additional day increased the odds of no-show by 3% (95% confidence interval: 2%-3%). Finally, unknown HCV status predicted no-show even after adjustment for age, gender, days between screenings and injection status. We found that drug injection did not predict no-show, further justifying expanded access to HCV treatment among people who inject drugs. Those 35 years and younger were more likely to no-show, suggesting that younger individuals may require targeted strategies for increasing testing and treatment uptake. Finally, streamlining the treatment cascade may also improve outcomes, as participants in the current study were more likely to attend if there were fewer days between phone screening and scheduled in-person screening.

Kentucky Outreach Service Kiosk (KyOSK) Study protocol: a community-level, controlled quasi-experimental, type 1 hybrid effectiveness study to assess implementation, effectiveness and cost-effectiveness of a community-tailored harm reduction kiosk on HIV, HCV and overdose risk in rural Appalachia.

INTRODUCTION: Many rural communities bear a disproportionate share of drug-related harms. Innovative harm reduction service models, such as vending machines or kiosks, can expand access to services that reduce drug-related harms. However, few kiosks operate in the USA, and their implementation, impact and cost-effectiveness have not been adequately evaluated in rural settings. This paper describes the Kentucky Outreach Service Kiosk (KyOSK) Study protocol to test the effectiveness, implementation outcomes and cost-effectiveness of a community-tailored, harm reduction kiosk in reducing HIV, hepatitis C and overdose risk in rural Appalachia. METHODS AND ANALYSIS: KyOSK is a community-level, controlled quasi-experimental, non-randomised trial. KyOSK involves two cohorts of people who use drugs, one in an intervention county (n=425) and one in a control county (n=325). People who are 18 years or older, are community-dwelling residents in the target counties and have used drugs to get high in the past 6 months are eligible. The trial compares the effectiveness of a fixed-site, staffed syringe service programme (standard of care) with the standard of care supplemented with a kiosk. The kiosk will contain various harm reduction supplies accessible to participants upon valid code entry, allowing dispensing data to be linked to participant survey data. The kiosk will include a call-back feature that allows participants to select needed services and receive linkage-to-care services from a peer recovery coach. The cohorts complete follow-up surveys every 6 months for 36 months (three preceding kiosk implementation and four post-implementation). The study will test the effectiveness of the kiosk on reducing risk behaviours associated with overdose, HIV and hepatitis C, as well as implementation outcomes and cost-effectiveness. ETHICS AND DISSEMINATION: The University of Kentucky Institutional Review Board approved the protocol. Results will be disseminated in academic conferences and peer-reviewed journals, online and print media, and community meetings. TRIAL REGISTRATION NUMBER: NCT05657106.

Validation of dried blood spots for capturing hepatitis C virus diversity for genomic surveillance.

Dried blood spots (DBS) have emerged as a promising alternative to traditional venous blood for hepatitis C virus (HCV) testing. However, their capacity to accurately reflect the genetic diversity of HCV remains poorly understood. We employed deep sequencing and advanced phylogenetic analyses on paired plasma and DBS samples from two common subtypes to evaluate the suitability of DBS for genomic surveillance. Results demonstrated that DBS captured equivalent viral diversity compared to plasma with no phylogenetic discordance observed. The ability of DBS to accurately reflect the profile of viral genetic diversity suggests it may be a promising avenue for future surveillance efforts to curb HCV outbreaks.

Correlates of Recent HIV Testing Among People Who Inject Drugs in Rural Areas: A Multi-site Cross-Sectional Study, 2018-2020.

The Rural Opioid Initiative surveyed 2693 people who inject drugs (PWID) in eight rural U.S. areas in 2018-2020 about self-reported HIV testing in the past 6 months. Correlates of interest included receipt of any drug-related services, incarceration history, and structural barriers to care (e.g., lack of insurance, proximity to syringe service programs [SSP]). Overall, 20% of participants reported receiving an HIV test within the past 6 months. Multivariable generalized estimating equations showed that attending substance use disorder (SUD) treatment (OR 2.11, 95%CI [1.58, 2.82]), having health insurance (OR 1.42, 95%CI [1.01, 2.00]) and recent incarceration (OR 1.49, 95%CI [1.08, 2.04]) were positively associated with HIV testing, while experiencing a resource barrier to healthcare (inability to pay, lack of transportation, inconvenient hours, or lack of child care) had inverse (OR 0.73, 95%CI [0.56, 0.94]) association with HIV testing. We found that the prevalence of HIV testing among rural PWID is low, indicating an unmet need for testing. While SUD treatment or incarceration may increase chances for HIV testing for rural PWID, other avenues for expanding HIV testing, such as SSP, need to be explored.

Latent Class Analysis of Online Platforms for Partner-Seeking and Sexual Behaviors Among Men Who Have Sex with Men from Central Kentucky.

Little is known whether engagement in sexual behaviors associated with potential HIV risks differs by subgroups of men who have sex with men (MSM), who are distinct regarding patterns of use of online tools for partner-seeking. Using latent class analysis, we revealed four classes of app-using MSM (n = 181, 18-34 y.o., 82.4% identified as White and non-Hispanic) residing in Central Kentucky: the Grindr/Tinder class; the Poly App Use class of MSM-oriented apps; the General Social Media class, and the Bumble class. Unadjusted penalized logistic regressions showed associations of the Poly App Use class with increased numbers of receptive anal sex partners and reporting condomless receptive anal sex. Adjusting for other covariates, poly app users versus others were more likely to be older (25-34 vs. 18-24, AOR = 3.81, 95%CI = 1.70-9.03), to report past six-month illicit drug use (AOR = 2.93, 95%CI = 1.25-7.43) and to have ever used pre-exposure prophylaxis (AOR = 2.79, 95%CI = 1.10-7.12). Poly app users had behavior profiles associated with an elevated HIV risk and also reported HIV-related protective behaviors likely indicating increased risk awareness among this class. Our findings warrant differentiation of behavior profiles by patterns of app use and suggest not to generalize sexual behaviors associated with potential HIV risks to all app-using MSM.

RESUMEN: Poco se sabe si la participación en comportamientos sexuales asociados con riesgos potenciales de VIH difiere según distintos subgrupos de hombres que tienen sexo con hombres (HSH) en base a patrones en el uso de herramientas online para la búsqueda de pareja. Mediante el análisis de clases latentes, identificamos cuatro clases de HSH que usan aplicaciones (n = 181, 18­34 años, 82.4% identificados como Blancos y no-Hispanos) y residen en Kentucky Central: la clase Grindr/Tinder, la clase Poly App Use de aplicaciones orientadas a HSH, la clase General Social Media y la clase Bumble. Las regresiones logísticas penalizadas no ajustadas mostraron asociaciones de la clase Poly App Use con un alza en el número de parejas receptivas en relaciones sexuales anales y con reportes de sexo anal receptivo sin condón. Al ajustar por otras covariables, los miembros de la clase Poly App Use, con respecto a las otras clases, tuvieron más probabilidades de ser mayores (25­34 vs. 18­24, ORA = 3,81; IC95%=1,70 − 9,03), de informar el uso de drogas ilícitas en los últimos 6 meses (ORA = 2,93; IC95%=1,25 − 7,43) y haber utilizado alguna vez profilaxis-preexposición (ORA = 2,79; IC95%=1,10 − 7,12). Así mismo, los miembros de esta clase tuvieron perfiles de comportamiento asociado con el riesgo elevado de VIH y también informaron comportamientos de protección relacionados con el VIH que probablemente indican una mayor conciencia del riesgo entre ellos. Nuestros hallazgos justifican la diferenciación de perfiles de comportamiento en base a patrones de uso de aplicaciones y sugieren no generalizar los comportamientos sexuales asociados con riesgos potenciales de VIH a todos los HSH que las usan.

Willingness to utilize a mobile treatment unit in five counties at the epicenter of the US rural opioid epidemic.

INTRODUCTION: US federal policies are evolving to expand the provision of mobile treatment units (MTUs) offering medications for opioid use disorder (MOUD). Mobile MOUD services are critical for rural areas with poor geographic access to fixed-site treatment providers. This study explored willingness to utilize an MTU among a sample of people who use opioids in rural Eastern Kentucky counties at the epicenter of the US opioid epidemic. METHODS: The study analyzed Cross-sectional survey data from the Kentucky Communities and Researchers Engaging to Halt the Opioid Epidemic (CARE2HOPE) study covering five rural counties in the state. Logistic regression models investigated the association between willingness to utilize an MTU providing buprenorphine and naltrexone and potential correlates of willingness, identified using the Behavioral Model for Vulnerable Populations. RESULTS: The analytic sample comprised 174 people who used opioids within the past six months. Willingness to utilize an MTU was high; 76.5 % of participants endorsed being willing. Those who had recently received MOUD treatment, compared to those who had not received any form of treatment or recovery support services, had six-fold higher odds of willingness to use an MTU. However, odds of being willing to utilize an MTU were 73 % lower among those who were under community supervision (e.g., parole, probation) and 81 % lower among participants who experienced an overdose within the past six months. CONCLUSIONS: There was high acceptability of MTUs offering buprenorphine and naltrexone within this sample, highlighting the potential for MTUs to alleviate opioid-related harms in underserved rural areas. However, the finding that people who were recently under community supervision or had overdosed were significantly less willing to seek mobile MOUD treatment suggest barriers (e.g., stigma) to mobile MOUD at individual and systemic levels, which may prevent improving opioid-related outcomes in these rural communities given their high rates of criminal-legal involvement and overdose.

Recent Incarceration, Substance Use, Overdose, and Service Use Among People Who Use Drugs in Rural Communities.

Importance: Drug use and incarceration have a substantial impact on rural communities, but factors associated with the incarceration of rural people who use drugs (PWUD) have not been thoroughly investigated. Objective: To characterize associations between recent incarceration, overdose, and substance use disorder (SUD) treatment access among rural PWUD. Design, Setting, and Participants: For this cross-sectional study, the Rural Opioid Initiative research consortium conducted a survey in geographically diverse rural counties with high rates of overdose across 10 US states (Illinois, Wisconsin, North Carolina, Oregon, Kentucky, West Virginia, Ohio, Massachusetts, New Hampshire, and Vermont) between January 25, 2018, and March 17, 2020, asking PWUD about their substance use, substance use treatment, and interactions with the criminal legal system. Participants were recruited through respondent-driven sampling in 8 rural US regions. Respondents who were willing to recruit additional respondents from their personal networks were enrolled at syringe service programs, community support organizations, and through direct community outreach; these so-called seed respondents then recruited others. Of 3044 respondents, 2935 included participants who resided in rural communities and reported past-30-day injection of any drug or use of opioids nonmedically via any route. Data were analyzed from February 8, 2022, to September 15, 2023. Exposure: Recent incarceration was the exposure of interest, defined as a report of incarceration in jail or prison for at least 1 day in the past 6 months. Main Outcomes and Measures: The associations between PWUD who were recently incarcerated and main outcomes of treatment use and overdose were examined using logistic regression. Results: Of 2935 participants, 1662 (56.6%) were male, 2496 (85.0%) were White; the mean (SD) age was 36 (10) years; and in the past 30 days, 2507 (85.4%) reported opioid use and 1663 (56.7%) reported injecting drugs daily. A total of 1224 participants (41.7%) reported recent incarceration, with a median (IQR) incarceration of 15 (3-60) days in the past 6 months. Recent incarceration was associated with past-6-month overdose (adjusted odds ratio [AOR], 1.38; 95% CI, 1.12-1.70) and recent SUD treatment (AOR, 1.62; 95% CI, 1.36-1.93) but not recent medication for opioid use disorder (MOUD; AOR, 1.03; 95% CI, 0.82-1.28) or currently carrying naloxone (AOR, 1.02; 95% CI, 0.86-1.21). Conclusions and Relevance: In this cross-sectional study of PWUD in rural areas, participants commonly experienced recent incarceration, which was not associated with MOUD, an effective and lifesaving treatment. The criminal legal system should implement effective SUD treatment in rural areas, including MOUD and provision of naloxone, to fully align with evidence-based SUD health care policies.

Houselessness and syringe service program utilization among people who inject drugs in eight rural areas across the USA: a cross-sectional analysis.

BACKGROUND: Research conducted in urban areas has highlighted the impact of housing instability on people who inject drugs (PWID), revealing that it exacerbates vulnerability to drug-related harms and impedes syringe service program (SSP) use. However, few studies have explored the effects of houselessness on SSP use among rural PWID. This study examines the relationship between houselessness and SSP utilization among PWID in eight rural areas across 10 states. METHODS: PWID were recruited using respondent-driven sampling for a cross-sectional survey that queried self-reported drug use and SSP utilization in the prior 30 days, houselessness in the prior 6 months and sociodemographic characteristics. Using binomial logistic regression, we examined the relationship between experiencing houselessness and any SSP use. To assess the relationship between houselessness and the frequency of SSP use, we conducted multinomial logistic regression analyses among participants reporting any past 30-day SSP use. RESULTS: Among 2394 rural PWID, 56.5% had experienced houselessness in the prior 6 months, and 43.5% reported past 30-day SSP use. PWID who had experienced houselessness were more likely to report using an SSP compared to their housed counterparts (adjusted odds ratio [aOR] = 1.24 [95% confidence intervals [CI] 1.01, 1.52]). Among those who had used an SSP at least once (n = 972), those who experienced houselessness were just as likely to report SSP use two (aOR = 0.90 [95% CI 0.60, 1.36]) and three times (aOR = 1.18 [95% CI 0.77, 1.98]) compared to once. However, they were less likely to visit an SSP four or more times compared to once in the prior 30 days (aOR = 0.59 [95% CI 0.40, 0.85]). CONCLUSION: This study provides evidence that rural PWID who experience houselessness utilize SSPs at similar or higher rates as their housed counterparts. However, housing instability may pose barriers to more frequent SSP use. These findings are significant as people who experience houselessness are at increased risk for drug-related harms and encounter additional challenges when attempting to access SSPs.

Covid and the coalfield: Covid-19 vaccine hesitance in Wales and Appalachia.

BACKGROUND: Vaccine hesitancy is a barrier to Covid-19 vaccine uptake and displays a social gradient, compounding health disparities. While social gradients are a vital concept in health, they flatten distinctions between types of disadvantaged community. This paper focuses on vaccine hesitance in post-industrial and de-industrialising coalfields. The social consequences of the decline of coal mining may present barriers to vaccine uptake. METHODS: We ran parallel surveys in Wales (N = 4187) and US states overlapping with central Appalachia (N = 4864), to examine whether vaccine attitudes and uptake varied between areas with different coal mining histories. These surveys were accompanied by qualitative interviews of 36 residents of these coalfields to explore vaccination decisions and triangulate with survey data. RESULTS: Factor analysis identified four axes of attitudes in the survey data: vaccine confidence, covid scepticism, vaccine individualism, and concerned confusion. These themes were echoed in the interviews. Vaccine confidence was lower; and covid scepticism, vaccine individualism, and concerned confusion higher, in residents of areas of Wales with greater mining extent and where pits closed during certain periods. Residents of former US coal counties had lower vaccine confidence and higher covid scepticism, while those in current coal counties had greater vaccine individualism and concerned confusion. In former US coal counties and Welsh areas where pits closed since 1980, vaccine uptake was lower. Differences could not be explained by respondents' income and education. In the interviews, norms of social solidarity were often invoked by vaccinated respondents, while unvaccinated respondents did not frame decisions in the context of the industrial history of their areas. DISCUSSION: The legacy of coal-mining's decline presents barriers to public health campaigns. We show evidence of this across two historically significant coalfields. Attention is needed to avert negative public health consequences of global energy transition.

Associations of hepatitis C virus (HCV) antibody positivity with opioid, stimulant, and polysubstance injection among people who inject drugs (PWID) in rural U.S. communities.

BACKGROUND: People who inject drugs (PWID) in the rural U.S. often inject stimulants, alone or with opioids. The impact of these substance use patterns may influence HCV risk behaviors. This analysis examines the associations of HCV antibody positivity with injecting only opioids, only stimulants (methamphetamine/cocaine), and opioids and stimulants together among rural PWID. METHODS: The Rural Opioid Initiative (ROI) consists of eight research sites that enrolled people who use drugs in rural communities in ten U.S. states from 2018 to 2020. This cross-sectional analysis included adult participants who resided in a study area and injected any drug in the past 30 days. The primary outcome was HCV antibody positivity. The exposure of interest was injection drug use classified as only opioids, only stimulants, separate injections of opioids and stimulants, and same-syringe injection of both in the past 30 days. We used multivariable log-binomial regression with generalized linear mixed models to generate prevalence ratios (P.R.) adjusted for demographics, injection history, health insurance, and substance use treatment. RESULTS: Among 3,084 participants enrolled in the ROI, 1,982 met inclusion criteria. Most participants injected opioids and stimulants in the same syringe (34%) or separately (21%), followed by injecting only stimulants (26%), and injecting only opioids (19%). Half (51%) were HCV antibody positive. Compared to people who injected only stimulants, HCV antibody positivity was more prevalent among people who injected opioids alone (aPR=1.62, 95% CI:(1.29-2.03)), injected both opioids and stimulants separately (aPR=1.61, 95% CI:(1.32-1.95)), and in the same syringe (aPR=1.54, 95% CI:(1.28-1.85)). CONCLUSION: HCV antibody positivity, indicating prior exposure, was highest among those who had recently injected opioids, alone or with stimulants. Additional nucleic acid testing is necessary to confirm active infection. More research is needed to determine the underlying causes of HCV antibody positivity by injection use.

Methamphetamine use and utilization of medications for opioid use disorder among rural people who use drugs.

BACKGROUND: Methamphetamine use is common among persons with opioid use disorder. This study evaluated associations between methamphetamine use and treatment with agonist medications for opioid use disorder (MOUD, specifically buprenorphine, and/or methadone) in U.S. rural communities. METHODS: The Rural Opioid Initiative (ROI) is a consortium spanning 10 states and 65 rural counties that included persons who reported past 30-day use of opioids and/or injection drug use between 1/2018 and 3/2020. Analyses were restricted to participants who had ever used opioids and had data on past 30-day methamphetamine use. Multivariable models examined the relationship between methamphetamine use and utilization of agonist MOUD. RESULTS: Among 2899 participants, 2179 (75.2%) also reported recent methamphetamine use. Persons with methamphetamine use compared to those without were younger, more likely to have injected drugs, be unhoused, criminal justice involved, and less likely to have health insurance. Adjusted for age, sex, race, and study site, recent methamphetamine use was associated with lower relative odds of past 30-day methadone treatment (aOR=0.66; 95% CI: 0.45-0.99) and fewer methadone treatment days (aIRR=0.76; 0.57-0.99), but not past 30-day buprenorphine receipt (aOR=0.90; 0.67-1.20), buprenorphine treatment days in past 6 months: aIRR=0.88; 0.69-1.12) or perceived inability to access buprenorphine (aOR=1.12; 0.87-1.44) or methadone (aOR=1.06; 0.76-1.48). CONCLUSION: Methamphetamine use is common among persons who use opioids in rural U.S. areas and negatively associated with current treatment and retention on methadone but not buprenorphine. Future studies should examine reasons for this disparity and reduce barriers to methadone for persons who use opioids and methamphetamine.

Validation of Dried Blood Spots for Capturing Hepatitis C Virus Diversity for Genomic Surveillance.

Dried blood spots (DBS) have emerged as a promising alternative to traditional venous blood for HCV testing. However, their capacity to accurately reflect the genetic diversity of HCV remains poorly understood. We employed deep sequencing and advanced phylogenetic analyses on paired plasma and DBS samples to evaluate the suitability of DBS for genomic surveillance. Results demonstrated that DBS captured equivalent viral diversity compared to plasma with no phylogenetic discordance observed. The ability of DBS to accurately reflect the profile of viral genetic diversity suggests it may be a promising avenue for future surveillance efforts to curb HCV outbreaks.

Contraception and Healthcare Utilization by Reproductive-Age Women Who Use Drugs in Rural Communities: a Cross-Sectional Survey.

BACKGROUND: Women who use drugs (WWUD) have low rates of contraceptive use and high rates of unintended pregnancy. Drug use is common among women in rural U.S. communities, with limited data on how they utilize reproductive, substance use disorder (SUD), and healthcare services. OBJECTIVE: We determined contraceptive use prevalence among WWUD in rural communities then compared estimates to women from similar rural areas. We investigated characteristics of those using contraceptives, and associations between contraceptive use and SUD treatment, healthcare utilization, and substance use. DESIGN: Rural Opioids Initiative (ROI) - cross-sectional survey using respondent-driven sampling (RDS) involving eight rural U.S. regions (January 2018-March 2020); National Survey on Family Growth (NSFG) - nationally-representative U.S. household reproductive health survey (2017-2019). PARTICIPANTS: Women aged 18-49 with prior 30-day non-prescribed opioid and/or non-opioid injection drug use; fecundity determined by self-reported survey responses. MAIN MEASURES: Unweighted and RDS-weighted prevalence estimates of medical/procedural contraceptive use; chi-squared tests and multi-level linear regressions to test associations. KEY RESULTS: Of 855 women in the ROI, 36.8% (95% CI 33.7-40.1, unweighted) and 38.6% (95% CI 30.7-47.2, weighted) reported contraceptive use, compared to 66% of rural women in the NSFG sample. Among the ROI women, 27% had received prior 30-day SUD treatment via outpatient counseling or inpatient program and these women had increased odds of contraceptive use (aOR 1.50 [95% CI 1.08-2.06]). There was a positive association between contraception use and recent medications for opioid use disorder (aOR 1.34 [95% CI 0.95-1.88]) and prior 6-month primary care utilization (aOR 1.32 [95% CI 0.96-1.82]) that did not meet the threshold for statistical significance. CONCLUSION: WWUD in rural areas reported low contraceptive use; those who recently received SUD treatment had greater odds of contraceptive use. Improvements are needed in expanding reproductive and preventive health within SUD treatment and primary care services in rural communities.

Deep Neural Networks for Fine-Grained Surveillance of Overdose Mortality.

Surveillance of drug overdose deaths relies on death certificates for identification of the substances that caused death. Drugs and drug classes can be identified through the International Classification of Diseases, Tenth Revision (ICD-10), codes present on death certificates. However, ICD-10 codes do not always provide high levels of specificity in drug identification. To achieve more fine-grained identification of substances on death certificate, the free-text cause-of-death section, completed by the medical certifier, must be analyzed. Current methods for analyzing free-text death certificates rely solely on lookup tables for identifying specific substances, which must be frequently updated and maintained. To improve identification of drugs on death certificates, a deep-learning named-entity recognition model was developed, utilizing data from the Kentucky Drug Overdose Fatality Surveillance System (2014-2019), which achieved an F1-score of 99.13%. This model can identify new drug misspellings and novel substances that are not present on current surveillance lookup tables, enhancing the surveillance of drug overdose deaths.

Protocol for community-driven selection of strategies to implement evidence-based practices to reduce opioid overdoses in the HEALing Communities Study: a trial to evaluate a community-engaged intervention in Kentucky, Massachusetts, New York and Ohio.

INTRODUCTION: Opioid-involved overdose deaths continue to surge in many communities, despite numerous evidence-based practices (EBPs) that exist to prevent them. The HEALing Communities Study (HCS) was launched to develop and test an intervention (ie, Communities That HEAL (CTH)) that supports communities in expanding uptake of EBPs to reduce opioid-involved overdose deaths. This paper describes a protocol for a process foundational to the CTH intervention through which community coalitions select strategies to implement EBPs locally. METHODS AND ANALYSIS: The CTH is being implemented in 67 communities (randomised to receive the intervention) in four states in partnership with coalitions (one per community). Coalitions must select at least five strategies, including one to implement each of the following EBPs: (a) overdose education and naloxone distribution; expanded (b) access to medications for opioid use disorder (MOUD), (c) linkage to MOUD, (d) retention in MOUD and (e) safer opioid prescribing/dispensing. Facilitated by decision aid tools, the community action planning process includes (1) data-driven goal setting, (2) discussion and prioritisation of EBP strategies, (3) selection of EBP strategies and (4) identification of next steps. Following review of epidemiologic data and information on existing local services, coalitions set goals and discuss, score and/or rank EBP strategies based on feasibility, appropriateness within the community context and potential impact on reducing opioid-involved overdose deaths with a focus on three key sectors (healthcare, behavioural health and criminal justice) and high-risk/vulnerable populations. Coalitions then select EBP strategies through consensus or majority vote and, subsequently, suggest or choose agencies with which to partner for implementation. ETHICS AND DISSEMINATION: The HCS protocol was approved by a central Institutional Review Board (Advarra). Results of the action planning process will be disseminated in academic conferences and peer-reviewed journals, online and print media, and in meetings with community stakeholders. TRIAL REGISTRATION NUMBER: NCT04111939.

Association of Methamphetamine and Opioid Use With Nonfatal Overdose in Rural Communities.

Importance: Overdoses continue to increase in the US, but the contribution of methamphetamine use is understudied in rural communities. Objective: To estimate the prevalence of methamphetamine use and its correlates among people who use drugs (PWUD) in rural US communities and to determine whether methamphetamine use is associated with increased nonfatal overdoses. Design, Setting, and Participants: From January 2018 through March 2020, the National Rural Opioid Initiative conducted cross-sectional surveys of PWUD in rural communities in 10 states (Illinois, Kentucky, New Hampshire, Massachusetts, North Carolina, Ohio, Oregon, Vermont, West Virginia, and Wisconsin). Participants included rural PWUD who reported any past-30-day injection drug use or noninjection opioid use to get high. A modified chain-referral sampling strategy identified seeds who referred others using drugs. Data analysis was performed from May 2021 to January 2022. Exposures: Use of methamphetamine alone, opioids alone, or both. Main Outcomes and Measures: Unweighted and weighted prevalence of methamphetamine use, any past-180-day nonfatal overdose, and number of lifetime nonfatal overdoses. Results: Among the 3048 participants, 1737 (57%) were male, 2576 (85%) were White, and 225 (7.4%) were American Indian; the mean (SD) age was 36 (10) years. Most participants (1878 of 2970 participants with any opioid or methamphetamine use [63%]) reported co-use of methamphetamine and opioids, followed by opioids alone (702 participants [24%]), and methamphetamine alone (390 participants [13%]). The estimated unweighted prevalence of methamphetamine use was 80% (95% CI, 64%-90%), and the estimated weighted prevalence was 79% (95% CI, 57%-91%). Nonfatal overdose was greatest in people using both methamphetamine and opioids (395 of 2854 participants with nonmissing overdose data [22%]) vs opioids alone (99 participants [14%]) or methamphetamine alone (23 participants [6%]). Co-use of methamphetamine and opioids was associated with greater nonfatal overdose compared with opioid use alone (adjusted odds ratio, 1.45; 95% CI, 1.08-1.94; P = .01) and methamphetamine use alone (adjusted odds ratio, 3.26; 95% CI, 2.06-5.14; P < .001). Those with co-use had a mean (SD) of 2.4 (4.2) (median [IQR], 1 [0-3]) lifetime overdoses compared with 1.7 (3.5) (median [IQR], 0 [0-2]) among those using opioids alone (adjusted rate ratio, 1.20; 95% CI, 1.01-1.43; P = .04), and 1.1 (2.9) (median [IQR], 0 [0-1]) among those using methamphetamine alone (adjusted rate ratio, 1.81; 95% CI, 1.45-2.27; P < .001). Participants with co-use most often reported having tried and failed to access substance use treatment: 827 participants (44%) for both, 117 participants (30%) for methamphetamine alone, and 252 participants (36%) for opioids alone (χ22 = 33.8; P < .001). Only 66 participants (17%) using methamphetamine alone had naloxone. Conclusions and Relevance: These findings suggest that harm reduction and substance use disorder treatment interventions must address both methamphetamine and opioids to decrease overdose in rural communities.

Extent and implications of omitted ties on network measures in a longitudinal social network survey of people who use drugs.

BACKGROUND: Previous research has revealed under-reporting of personal network members (i.e., alters) in studies involving people who use drugs (PWUD). This analysis (1) characterizes relationships that were more likely to be omitted but later recalled with prompting and (2) identifies network structural characteristics most impacted by these omissions among a sample of PWUD in rural Appalachian Kentucky, an epicenter of the opioid epidemic. METHODS: Data were collected through longitudinal assessments as part of the Social Networks Among Appalachian People (SNAP) study (2008-2017). Study participants completed interviewer-administered questionnaires that collected social network data via free-listing at baseline and six-month intervals. At visit 5, after free-listing, interviewers prompted participants with the names of previously reported alters. We used modified Poisson regression with generalized estimating equations to identify individual- and relationship-level characteristics associated with an alter being reported only after prompting. We examined the impact of including vs. excluding relationships reported after prompting on local and global sociometric network measures (i.e., betweenness centrality, bridging, density, mean degree, transitivity, cliques, and 2-cores). RESULTS: Relationships reported only after prompting were more likely to be immediate family (Adjusted Prevalence Ratio [APR]:1.29; 95% Confidence Interval [CI]: 1.03-1.63) and less likely to involve sex (APR:0.54; 95% CI: 0.43-0.67). Considerable differences were observed for participant positional rankings of betweenness centrality and bridging, and differences in network density and average degree pre- and post-prompting were statistically significant. CONCLUSION: Longitudinal network studies that aim to assess transmission dynamics, information diffusion, or peer influence should consider the effects of omitted relationships.