Internet is becoming an increasingly common tool for survey research, particularly among "hidden" or vulnerable populations, such as men who have sex with men (MSM). Web-based research has many ...advantages for participants and researchers, but fraud can present a significant threat to data integrity.
The purpose of this analysis was to evaluate fraud detection strategies in a Web-based survey of young MSM and describe new protocols to improve fraud detection in Web-based survey research.
This study involved a cross-sectional Web-based survey that examined individual- and network-level risk factors for HIV transmission and substance use among young MSM residing in 15 counties in Central Kentucky. Each survey entry, which was at least 50% complete, was evaluated by the study staff for fraud using an algorithm involving 8 criteria based on a combination of geolocation data, survey data, and personal information. Entries were classified as fraudulent, potentially fraudulent, or valid. Descriptive analyses were performed to describe each fraud detection criterion among entries.
Of the 414 survey entries, the final categorization resulted in 119 (28.7%) entries identified as fraud, 42 (10.1%) as potential fraud, and 253 (61.1%) as valid. Geolocation outside of the study area (164/414, 39.6%) was the most frequently violated criterion. However, 33.3% (82/246) of the entries that had ineligible geolocations belonged to participants who were in eligible locations (as verified by their request to mail payment to an address within the study area or participation at a local event). The second most frequently violated criterion was an invalid phone number (94/414, 22.7%), followed by mismatching names within an entry (43/414, 10.4%) and unusual email addresses (37/414, 8.9%). Less than 5% (18/414) of the entries had some combination of personal information items matching that of a previous entry.
This study suggests that researchers conducting Web-based surveys of MSM should be vigilant about the potential for fraud. Researchers should have a fraud detection algorithm in place prior to data collection and should not rely on the Internet Protocol (IP) address or geolocation alone, but should rather use a combination of indicators.
Abstract Purpose The purpose of this review was to systematically summarize research on nonmedical use of prescription medications (NMUPM) among U.S. adolescents, with specific focus on scheduled ...medications falling into one of the following drug classes: pain relievers, stimulants, sedatives, or tranquilizers. Methods Databases were searched for peer-reviewed primary quantitative research published between January 2000 and June 2011 on NMUPM among out-of-treatment U.S. adolescents aged 12–17 years (or age 18 if enrolled in high school). Results Thirty publications met inclusion criteria. A total of 25 studies were represented; 15 involved nationally representative samples. The prevalence and correlates of NMUPM varied across studies and by drug class. Nonmedical use of pain relievers was more prevalent than for stimulants, sedatives, and tranquilizers. Female gender was generally associated with pain reliever use and, to a lesser degree, with tranquilizer use. White adolescents also appeared to have a higher prevalence of NMUPM, although there was some evidence to the contrary. Older age, illicit drug use, and delinquency were consistently associated with NMUPM across studies. Conclusions This review identified several areas for further research, including that of racially/ethnically diverse samples of adolescents, more focus on sedative and tranquilizer use, and longitudinal research to examine temporal patterns in NMUPM and other illicit drug use, delinquency, and substance abuse and dependence.
Timely data is key to effective public health responses to epidemics. Drug overdose deaths are identified in surveillance systems through ICD-10 codes present on death certificates. ICD-10 coding ...takes time, but free-text information is available on death certificates prior to ICD-10 coding. The objective of this study was to develop a machine learning method to classify free-text death certificates as drug overdoses to provide faster drug overdose mortality surveillance.
Using 2017-2018 Kentucky death certificate data, free-text fields were tokenized and features were created from these tokens using natural language processing (NLP). Word, bigram, and trigram features were created as well as features indicating the part-of-speech of each word. These features were then used to train machine learning classifiers on 2017 data. The resulting models were tested on 2018 Kentucky data and compared to a simple rule-based classification approach. Documented code for this method is available for reuse and extensions: https://github.com/pjward5656/dcnlp.
The top scoring machine learning model achieved 0.96 positive predictive value (PPV) and 0.98 sensitivity for an F-score of 0.97 in identification of fatal drug overdoses on test data. This machine learning model achieved significantly higher performance for sensitivity (p<0.001) than the rule-based approach. Additional feature engineering may improve the model's prediction. This model can be deployed on death certificates as soon as the free-text is available, eliminating the time needed to code the death certificates.
Machine learning using natural language processing is a relatively new approach in the context of surveillance of health conditions. This method presents an accessible application of machine learning that improves the timeliness of drug overdose mortality surveillance. As such, it can be employed to inform public health responses to the drug overdose epidemic in near-real time as opposed to several weeks following events.
ABSTRACT
Aim The study's objectives were to characterize initiation of injection drug use, examine the independent association of specific substance use with injection drug use and determine factors ...associated with rates of transition from first illicit drug use to first injection among a sample of rural Appalachian drug users.
Design Interview‐administered questionnaires were administered to a sample of drug users recruited via respondent‐driven sampling.
Setting Appalachian Kentucky.
Participants Injection drug users (IDUs) (n = 394) and non‐IDUs (n = 109).
Measurements Data were collected on substance use and years from age at initiation of illicit substance use to ‘event’ (initiation of injection or date of baseline interview for non‐IDUs). Logistic regression and Cox regression were used to identify factors associated with life‐time injection drug use and transition time to injection, respectively.
Findings OxyContin® was involved in nearly as many initiations to injection (48%), as were stimulants, other prescription opioids and heroin combined; for participants who initiated with OxyContin®, the median time from which they began OxyContin® use to their first injection of OxyContin® was 3 years. Adjusting for demographics, five prescription drugs (benzodiazepines, illicit methadone, oxycodone, OxyContin® and other opiates) were associated with an increased hazard for transitioning from first illicit drug use to first injection drug use (each at P < 0.01).
Conclusions In Appalachia, in the United States, the prescription opioid OxyContin® is widely used non‐medically and appears to show a particularly high risk of rapid transition to injection compared with the use of other illicit drugs.
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.
Buprenorphine is a cornerstone to curbing opioid epidemics, but emerging data suggest that rural pharmacists in the US sometimes refuse to dispense this medication. We conducted a case study to ...explore buprenorphine dispensing practices in 12 rural Appalachian Kentucky counties, and analyze whether and how they were shaped by features of the rural risk environment.
In this case study, we conducted one-on-one semi-structured interviews with 14 pharmacists operating 15 pharmacies in these counties to explore buprenorphine dispensing practices and perceived influences on these practices. Thematic analyses of the resulting transcripts revealed three features of the rural risk environment that shaped dispensing. To explore these three risk environment features, we analyzed policy documents (e.g., Attorney General lawsuits) and administrative databases (e.g., incarceration data). Textual documents were analyzed using thematic analyses and administrative data were analyzed using descriptive statistics; memoes explored relationships among risk environment features and dispensing practices.
Twelve of the 15 pharmacies limited dispensing, by refusing to serve new patients; limiting dispensing to known patients or prescribers; or refusing to dispense buprenorphine altogether. Concerns about exceeding a “Drug Enforcement Administration (DEA) cap” on opioid dispensing stifled dispensing. A legacy of aggressive and fraudulent marketing of opioid analgesics (OAs) by pharmaceutical companies and physician OA overprescribing undermined pharmacist trust in buprenorphine and in its prescribers. The escalating local war on drugs may have undermined dispensing by reinforcing stigma against people who use drugs.
Initiatives to increase buprenorphine prescribing must be accompanied by policy changes to increase dispensing. Specifically, buprenorphine should be removed from opioid monitoring systems; efforts to de-escalate the war on drugs should be extended to encompass rural areas; initiatives to dismantle aggressive OA marketing should be strengthened; and efforts to re-build pharmacist trust in physicians are needed.
Though overdose rates have been increasing in US rural areas for two decades, little is known about the rural risk environment for overdoses. This qualitative study explored the risk environment for ...overdoses among young adults in Eastern Kentucky, a rural epicenter of the US opioid epidemic.
Participants were recruited via community-based outreach. Eligibility criteria included living in one of five rural Eastern Kentucky counties; being aged 18–35; and using opioids to get high in the past 30 days. Semi-structured interviews explored the rural risk environment, and strategies to prevent overdose and dying from an overdose. Interviews were transcribed verbatim and analyzed using constructivist grounded-theory methods.
In this sample (N = 19), participants reported using in a range of locations, including homes and outdoor settings; concerns about community stigma and law enforcement shaped the settings where participants used opioids and the strategies they deployed in these settings to prevent an overdose, and to survive an overdose. Almost half of participants reported using opioids in a “trap house” or other dealing locations, often to evade police after buying drugs, and reported that others present pressed them to use more than usual. If an overdose occurred in this setting, however, these same people might refuse to call EMS to protect themselves from arrest. Outdoor settings presented particular vulnerabilities to overdose and dying from an overdose. Most participants reported using opioids outdoors, where they skipped overdose prevention steps to reduce their risk of arrest; they worried that no one would find them if they overdosed, and that cell phone coverage would be too weak to summon EMS.
Findings suggest that initiatives to reduce overdoses in Eastern Kentucky would be strengthened by de-escalating the War on Drugs and engaging law enforcement in initiatives to protect the health of people who use opioids.
•Greater network stability over time for reciprocal and transitive relationships.•Stability of particular network structures can inform network-based interventions.•Selection and influence shape ...injection drug use initiation, cessation and maintenance.•Mechanisms shaping drug-use behaviors appear consistent over time.
Prior studies have shown that individuals and their peers often have similar substance use behaviors, but the mechanisms driving these similarities – particularly in rural settings, are not well understood. The primary objectives of this analysis are to (1) identify factors that contribute to relationship turnover and maintenance within a rural network of persons who use drugs (PWUD), (2) determine whether assimilation and/or homophily shape participants use of injection drugs, heroin, and stimulants (methamphetamine and cocaine), and (3) assess the extent that these mechanisms influence networks ties and/or behaviors and whether these effects vary across time.
Sociometric network data were collected from a cohort of PWUD in rural Eastern Kentucky at baseline (2008–2010) and at four follow-up visits conducted approximately semiannually. Stochastic actor-oriented models (SAOMS) were used to model network structure and participant behaviors as jointly dependent variables and to identify characteristics associated with the maintenance, dissolution, and formation of network ties and changes in drug use behaviors.
Findings suggest (1) greater network stability over time for reciprocal and transitive relationships, (2) both homophily and assimilation played a greater role in shaping injection drug use (IDU) initiation and cessation than they did in shaping heroin and stimulant use, and (3) the importance of these mechanisms appeared consistent over time.
Given the stability of particular network structures and evidence of both homophily and assimilation with respect to drug-use behaviors, interventions that leverage social networks could be used to motivate health-promoting behaviors.
•Qualitative interviews assessed ethical considerations for network data collection.•Willingness to provide partner name detail varied by relationship-level factors.•Perspectives varied more for sex ...partners than for drug use partners.•Privacy concerns were noted for screennames, phone numbers, tattoos, and scars.•Collecting reasons for omitting data could lend insight on network completeness.
Studies aiming to construct risk networks have historically collected network members' names, demographic characteristics and relational data (i.e., type, strength, duration, frequency of interaction, and HIV-related risk behaviors between the pair). Due to difficulties in constructing risk networks stemming from partner anonymity and the use of nicknames, some studies also collect network members’ screen names, phone numbers, physical attributes, and scars/tattoos to assist with entity resolution. In-depth interviews with 20 men who have sex with men and transgender women in Kentucky assessed privacy/confidentiality concerns as well as accuracy/recall issues associated with providing these details. Most preferred providing alters’ nicknames/first names only or a first name with the first letter of the last name. Many perceived screen names, phone numbers, and scars/tattoos to be too personal/identifying. Willingness to provide more detailed information varied by relationship type/strength, which could influence the validity of the resulting network.
People who inject drugs (PWID) experience a high prevalence of incarceration and might be at high risk of HIV and hepatitis C virus (HCV) infection during or after incarceration. We aimed to assess ...whether incarceration history elevates HIV or HCV acquisition risk among PWID.
In this systematic review and meta-analysis, we searched MEDLINE, Embase, and PsycINFO databases for studies in any language published from Jan 1, 2000 until June 13, 2017 assessing HIV or HCV incidence among PWID. We included studies that measured HIV or HCV incidence among community-recruited PWID. We included only studies reporting original results and excluded studies that evaluated incident infections by self-report. We contacted authors of cohort studies that met the inclusion or exclusion criteria, but that did not report on the outcomes of interest, to request data. We extracted and pooled data from the included studies using random-effects meta-analyses to quantify the associations between recent (past 3, 6, or 12 months or since last follow-up) or past incarceration and HIV or HCV acquisition (primary infection or reinfection) risk among PWID. We assessed the risk of bias of included studies using the Newcastle-Ottawa Scale. Between-study heterogeneity was evaluated using the I2 statistic and the P-value for heterogeneity.
We included published results from 20 studies and unpublished results from 21 studies. These studies originated from Australasia, western and eastern Europe, North and Latin America, and east and southeast Asia. Recent incarceration was associated with an 81% (relative risk RR 1·81, 95% CI 1·40–2·34) increase in HIV acquisition risk, with moderate heterogeneity between studies (I2=63·5%; p=0·001), and a 62% (RR 1·62, 95% CI 1·28–2·05) increase in HCV acquisition risk, also with moderate heterogeneity between studies (I2=57·3%; p=0·002). Past incarceration was associated with a 25% increase in HIV (RR 1·25, 95% CI 0·94–1·65) and a 21% increase in HCV (1·21, 1·02–1·43) acquisition risk.
Incarceration is associated with substantial short-term increases in HIV and HCV acquisition risk among PWID and could be a significant driver of HCV and HIV transmission among PWID. These findings support the need for developing novel interventions to minimise the risk of HCV and HIV acquisition, including addressing structural risks associated with drug laws and excessive incarceration of PWID.
Engineering and Physical Sciences Research Council, National Institute for Health Research, National Institutes of Health.