Co-prescription network reveals social dynamics of opioid doctor shopping

Network analysis of U.S. non-fatal opioid-involved overdose journeys, 2018-2023

We present a nation-wide network analysis of non-fatal opioid-involved overdose journeys in the United States. Leveraging a unique proprietary dataset of Emergency Medical Services incidents, we construct a journey-to-overdose geospatial network capturing nearly half a million opioid-involved overdose events spanning 2018-2023. We analyze the structure and sociological profiles of the nodes, which are counties or their equivalents, characterize the distribution of overdose journey lengths, and investigate changes in the journey network between 2018 and 2023. Our findings include that authority and hub nodes identified by the HITS algorithm tend to be located in urban areas and involved in overdose journeys with particularly long geographical distances.

Social network analysis for medical narcotics in South Korea: focusing on patients and healthcare organizations

BACKGROUND

Medical narcotics must be administered under medical supervision because of their potential for misuse and abuse, leading to more dangerous and addictive substances. The control of medical narcotics requires close monitoring to ensure that they remain safe and effective. This study proposes a methodology that can effectively identify the overprescription of medical narcotics in hospitals and patients.

METHODS

Social network analysis (SNA) was applied to prescription networks for medical narcotics. Prescription data were obtained from the Narcotics Information Management System in South Korea, which contains all data on narcotic usage nationwide. Two-mode networks comprising hospitals and patients were constructed based on prescription data from 2019 to 2021 for the three most significant narcotics: appetite suppressants, zolpidem, and propofol. Two-mode networks were then converted into one-mode networks for hospitals. Network structures and characteristics were analyzed to identify hospitals suspected of overprescribing.

RESULTS

The SNA identified hospitals that overprescribed medical narcotics. Patients suspected of experiencing narcotic addiction seek treatment in such hospitals. The structure of the network was different for the three narcotics. While appetite suppressants and propofol networks had a more centralized structure, zolpidem networks showed a less centralized but more fragmented structure. During the analysis, two types of hospitals caught our attention: one with a high degree, meaning that potential abusers have frequently visited the hospital, and the other with a high weighted degree, meaning that the hospital may overprescribe. For appetite suppressants, these two types of hospitals matched 84.6%, compared with 30.0% for propofol. In all three narcotics, clinics accounted for the largest share of the network. Patients using appetite suppressants were most likely to visit multiple locations, whereas those using zolpidem and propofol tended to form communities around their neighborhoods.

CONCLUSIONS

The significance of this study lies in its analysis of nationwide narcotic use reports and the differences observed across different types of narcotics. The social network structure between hospitals and patients varies depending on the composition of the medical narcotics. Therefore, these characteristics should be considered when controlling medication with narcotics. The results of this study provide guidelines for controlling narcotic use in other countries.

Graph-based machine learning improves just-in-time defect prediction

The increasing complexity of today's software requires the contribution of thousands of developers. This complex collaboration structure makes developers more likely to introduce defect-prone changes that lead to software faults. Determining when these defect-prone changes are introduced has proven challenging, and using traditional machine learning (ML) methods to make these determinations seems to have reached a plateau. In this work, we build contribution graphs consisting of developers and source files to capture the nuanced complexity of changes required to build software. By leveraging these contribution graphs, our research shows the potential of using graph-based ML to improve Just-In-Time (JIT) defect prediction. We hypothesize that features extracted from the contribution graphs may be better predictors of defect-prone changes than intrinsic features derived from software characteristics. We corroborate our hypothesis using graph-based ML for classifying edges that represent defect-prone changes. This new framing of the JIT defect prediction problem leads to remarkably better results. We test our approach on 14 open-source projects and show that our best model can predict whether or not a code change will lead to a defect with an F1 score as high as 77.55% and a Matthews correlation coefficient (MCC) as high as 53.16%. This represents a 152% higher F1 score and a 3% higher MCC over the state-of-the-art JIT defect prediction. We describe limitations, open challenges, and how this method can be used for operational JIT defect prediction.

Comparing measures of centrality in bipartite patient-prescriber networks: A study of drug seeking for opioid analgesics

Visiting multiple prescribers is a common method for obtaining prescription opioids for nonmedical use and has played an important role in fueling the United States opioid epidemic, leading to increased drug use disorder and overdose. Recent studies show that centrality of the bipartite network formed by prescription ties between patients and prescribers of opioids is a promising indicator for drug seeking. However, node prominence in bipartite networks is typically estimated with methods that do not fully account for the two-mode topology of the underlying network. Although several algorithms have been proposed recently to address this challenge, it is unclear how these algorithms perform on real-world networks. Here, we compare their performance in the context of identifying opioid drug seeking behaviors by applying them to massive bipartite networks of patients and providers extracted from insurance claims data. We find that two variants of bipartite centrality are significantly better predictors of subsequent opioid overdose than traditional centrality estimates. Moreover, we show that incorporating non-network attributes such as the potency of the opioid prescriptions into the measures can further improve their performance. These findings can be reproduced on different datasets. Our results demonstrate the potential of bipartiteness-aware indices for identifying patterns of high-risk behavior.

Defining "Doctor Shopping" with Dispensing Data: A Scoping Review

BACKGROUND

"Doctor shopping" typically refers to patients that seek controlled substance prescriptions from multiple providers with the presumed intent to obtain these medications for non-medical use and/or diversion. The purpose of this scoping review is to document and examine the criteria used to identify "doctor shopping" from dispensing data in the United States.

METHODS

A scoping review was conducted on "doctor shopping" or analogous terminology from January 1, 2000, through December 31, 2020, using the Web of Science Core Collection (seven citation indexes). Our search was limited to the United States only, English-language, peer-reviewed and US federal government studies. Studies without explicit "doctor shopping" criteria were excluded. Key components of these criteria included the number of prescribers and dispensers, dispensing period, and drug class (e.g., opioids).

RESULTS

Of 9,845 records identified, 95 articles met the inclusion criteria and our pool of studies ranged from years 2003 to 2020. The most common threshold-based or count definition was (≥4 Prescribers [P] AND ≥4 Dispensers [D]) (n = 12). Thirty-three studies used a 365-day detection window. Opioids alone were studied most commonly (n = 69), followed by benzodiazepines and stimulants (n = 5 and n = 2, respectively). Only 39 (41%) studies provided specific drug lists with active ingredients.

CONCLUSION

Relatively simple P x D criteria for identifying "doctor shopping" are still the dominant paradigm with the need for ongoing validation. The value of P x D criteria may change through time with more diverse methods applied to dispensing data emerging.

Analytic Advances in Social Networks and Health in the Twenty-First Century

The study of social networks is increasingly central to health research for medical sociologists and scholars in other fields. Here, we review the innovations in theory, substance, data collection, and methodology that have propelled the study of social networks and health from a niche subfield to the center of larger sociological and scientific debates. In particular, we contextualize the broader history of network analysis and its connections to health research, concentrating on work beginning in the late 1990s, much of it in this journal. Using bibliometric and network visualization approaches, we examine the subfield's evolution over this period in terms of topics, trends, key debates, and core insights. We conclude by reflecting on persistent challenges and areas of innovation shaping the study of social networks and health and its intersection with medical sociology in the coming years.

New means, new measures: assessing prescription drug-seeking indicators over 10 years of the opioid epidemic

BACKGROUND AND AIMS

Prescription drug-seeking (PDS) from multiple prescribers is a primary means of obtaining prescription opioids; however, PDS behavior has probably evolved in response to policy shifts, and there is little agreement about how to operationalize it. We systematically compared the performance of traditional and novel PDS indicators.

DESIGN

Longitudinal study using a de-identified commercial claims database.

SETTING

United States, 2009-18.

PARTICIPANTS

A total of 318 million provider visits from 21.5 million opioid-prescribed patients.

MEASUREMENTS

We applied binary classification and generalized linear models to compare predictive accuracy and average marginal effect size predicting future opioid use disorder (OUD), overdose and high morphine milligram equivalents (MME). We compared traditional indicators of PDS to a network centrality measure, PageRank, that reflects the prominence of patients in a co-prescribing network. Analyses used the same data and adjusted for patient demographics, region, SES, diagnoses and health services.

FINDINGS

The predictive accuracy of a widely used traditional measure (N + unique doctors and N + unique pharmacies in 90 days) on OUD, overdose and MME decreased between 2009 and 2018, and performed no better than chance (50% accuracy) after 2015. Binarized PageRank measures however exhibited higher predictive accuracy than the traditional binary measures throughout 2009-2018. Continuous indicators of PDS performed better than binary thresholds, with days of Rx performing best overall with 77-93% predictive accuracy. For example, days of Rx had the highest average marginal effects on overdose and OUD: a 1 standard deviation increase in days of Rx was associated with a 6-8% [confidence intervals (CIs) = 0.058-0.061 and 0.078-0.082] increase in the probability of overdose and a 4-5% (CIs = 0.038-0.043 and 0.047-0.053) increase in the probability of OUD. PageRank performed nearly as well or better than traditional indicators of PDS, with predictive performance increasing after 2016.

CONCLUSIONS

In the United States, network-based measures appear to have increasing promise for identifying prescription opioid drug-seeking behavior, while indicators based on quantity of providers or pharmacies appear to have decreasing utility.

Characteristics and Network Influence of Providers Involved in the Treatment of Patients With Chronic Back, Neck or Joint Pain in Arkansas

Increasing emphasis on guidelines and prescription drug monitoring programs highlight the role of healthcare providers in pain treatment. Objectives of this study were to identify characteristics of key players and influence of opioid prescribers through construction of a referral network of patients with chronic pain. A retrospective cohort study was performed and patients with commercial or Medicaid coverage with chronic back, neck, or joint pain were identified using the Arkansas All-Payer Claims-Database. A social network comprised of providers connected by patient referrals based on 12-months of healthcare utilization following chronic pain was constructed. Network measures evaluated were indegree and eigen (referrals obtained), betweenness (involvement), and closeness centrality (reach). Outcomes included influence of providers, opioid prescribers, and brokerage status. Exposures included provider demographics, specialties and network characteristics. There were 51,941 chronic pain patients who visited 8,110 healthcare providers. Primary care providers showed higher betweenness and closeness whereas specialists had higher indegree. Opioid providers showed higher centrality compared to non-opioid providers, which decreased with increasing volume of opioid prescribing. Non-pharmacologic providers showed significant brokerage scores. Findings from this study such as primary care providers having better reach, non-central positions of high-volume prescribers and non-pharmacologic providers having higher brokerage can aid interventional physician detailing. PERSPECTIVE: Opioid providers held central positions in the network aiding provider-directed interventions. However, high-volume opioid providers were at the borders making them difficult targets for interventions. Primary care providers had the highest reach, specialists received the most referrals and non-pharmacological providers and specialists acted as brokers between non-opioid and opioid prescribers.

Associations between mobility patterns and COVID-19 deaths during the pandemic: A network structure and rank propagation modelling approach

BACKGROUND

From February 2020, both urban and rural Ireland witnessed the rapid proliferation of the COVID-19 disease throughout its counties. During this period, the national COVID-19 responses included stay-at-home directives issued by the state, subject to varying levels of enforcement.

METHODS

In this paper, we present a new method to assess and rank the causes of Ireland COVID-19 deaths as it relates to mobility activities within each county provided by Google while taking into consideration the epidemiological confirmed positive cases reported per county. We used a network structure and rank propagation modelling approach using Personalised PageRank to reveal the importance of each mobility category linked to cases and deaths. Then a novel feature-selection method using relative prominent factors finds important features related to each county's death. Finally, we clustered the counties based on features selected with the network results using a customised network clustering algorithm for the research problem.

FINDINGS

Our analysis reveals that the most important mobility trend categories that exhibit the strongest association to COVID-19 cases and deaths include retail and recreation and workplaces. This is the first time a network structure and rank propagation modelling approach has been used to link COVID-19 data to mobility patterns. The infection determinants landscape illustrated by the network results aligns soundly with county socio-economic and demographic features. The novel feature selection and clustering method presented clusters useful to policymakers, managers of the health sector, politicians and even sociologists. Finally, each county has a different impact on the national total.

"Doctor and pharmacy shopping": A fading signal for prescription opioid use monitoring?

BACKGROUND

The term "doctor and pharmacy shopping" colloquially describes patients with high multiple provider episodes (MPEs)-a threshold count of distinct prescribers and/or pharmacies involved in prescription fulfillment. Opioid-related MPEs are implicated in the global opioid crisis and heavily monitored by government databases such as U.S. state prescription drug monitoring programs (PDMPs). We applied a widely-used MPE definition to examine U.S. trends from a large, commercially-insured population from 2010 to 2017. Further, we examined the proportion of enrollees identified as "doctor shoppers" with evidence of a cancer diagnosis to examine the risk of false positives.

METHODS

Using a large, commercially-insured population, we identified patients with opioid-related MPEs: opioid prescriptions (Schedule II-V, no buprenorphine) filled from ≥5 prescribers AND ≥ 5 pharmacies within the past 90 days ("5x5x90d"). Quarterly rates per 100,000 enrollees (two specifications) were calculated between 2010 and 2017. We examined the trend in a recently published all-payer, 7 state cohort from the U.S. Centers for Disease Control and Prevention for comparison. Cancer-related ICD-9/10-CM codes were used.

RESULTS

Quarterly MPE rates declined by approximately 73 % from 18.2-4.9 per 100,000 enrollee population with controlled substance prescriptions. In 2017, nearly one fifth of these commercially-insured enrollees identified by the 5x5x90d algorithm were diagnosed with cancer. Approximately 8% of this sample included patients with ≥ 1 buprenorphine prescriptions.

CONCLUSIONS

Opioid "shopping" flags are a long-standing but rapidly fading PDMP signal. To avoid unintended consequences, such as identifying legitimate medical encounters requiring high healthcare utilization or opioid treatment, while maintaining vigilance, more nuanced and sophisticated approaches are needed.

Systematic Evaluation of State Policy Interventions Targeting the US Opioid Epidemic, 2007-2018

Importance

In response to the increase in opioid overdose deaths in the United States, many states recently have implemented supply-controlling and harm-reduction policy measures. To date, an updated policy evaluation that considers the full policy landscape has not been conducted.

Objective

To evaluate 6 US state-level drug policies to ascertain whether they are associated with a reduction in indicators of prescription opioid abuse, the prevalence of opioid use disorder and overdose, the prescription of medication-assisted treatment (MAT), and drug overdose deaths.

Design, Setting, and Participants

This cross-sectional study used drug overdose mortality data from 50 states obtained from the National Vital Statistics System and claims data from 23 million commercially insured patients in the US between 2007 and 2018. Difference-in-differences analysis using panel matching was conducted to evaluate the prevalence of indicators of prescription opioid abuse, opioid use disorder and overdose diagnosis, the prescription of MAT, and drug overdose deaths before and after implementation of 6 state-level policies targeting the opioid epidemic. A random-effects meta-analysis model was used to summarize associations over time for each policy and outcome pair. The data analysis was conducted July 12, 2020.

Exposures

State-level drug policy changes to address the increase of opioid-related overdose deaths included prescription drug monitoring program (PDMP) access, mandatory PDMPs, pain clinic laws, prescription limit laws, naloxone access laws, and Good Samaritan laws.

Main Outcomes and Measures

The outcomes of interests were quarterly state-level mortality from drug overdoses, known indicators for prescription opioid abuse and doctor shopping, MAT, and prevalence of drug overdose and opioid use disorder.

Results

This cross-sectional study of drug overdose mortality data and insurance claims data from 23 million commercially insured patients (12 582 378 female patients [55.1%]; mean [SD] age, 45.9 [19.9] years) in the US between 2007 and 2018 found that mandatory PDMPs were associated with decreases in the proportion of patients taking opioids (-0.729%; 95% CI, -1.011% to -0.447%), with overlapping opioid claims (-0.027%; 95% CI, -0.038% to -0.017%), with daily morphine milligram equivalent greater than 90 (-0.095%; 95% CI, -0.150% to -0.041%), and who engaged in drug seeking (-0.002%; 95% CI, -0.003% to -0.001%). The proportion of patients receiving MAT increased after the enactment of mandatory PDMPs (0.015%; 95% CI, 0.002% to 0.028%), pain clinic laws (0.013%, 95% CI, 0.005%-0.021%), and prescription limit laws (0.034%, 95% CI, 0.020% to 0.049%). Mandatory PDMPs were associated with a decrease in the number of overdose deaths due to natural opioids (-518.5 [95% CI, -728.5 to -308.5] per 300 million people) and methadone (-122.7 [95% CI, -207.5 to -37.8] per 300 million people). Prescription drug monitoring program access policies showed similar results, although these policies were also associated with increases in overdose deaths due to synthetic opioids (380.3 [95% CI, 149.6-610.8] per 300 million people) and cocaine (103.7 [95% CI, 28.0-179.5] per 300 million people). Except for the negative association between prescription limit laws and synthetic opioid deaths (-723.9 [95% CI, -1419.7 to -28.1] per 300 million people), other policies were associated with increasing overdose deaths, especially those attributed to non-prescription opioids such as synthetic opioids and heroin. This includes a positive association between naloxone access laws and the number of deaths attributed to synthetic opioids (1338.2 [95% CI, 662.5 to 2014.0] per 300 million people).

Conclusions and Relevance

Although this study found that existing state policies were associated with reduced misuse of prescription opioids, they may have the unintended consequence of motivating those with opioid use disorders to access the illicit drug market, potentially increasing overdose mortality. This finding suggests that there is no easy policy solution to reverse the epidemic of opioid dependence and mortality in the US.

Health Information Technology and Doctor Shopping: A Systematic Review

Doctor shopping is the practice of visiting multiple physicians to obtain multiple prescriptions. Health information technology (HIT) allows healthcare providers and patients to leverage records or shared information to improve effective care. Our research objective was to determine how HIT is being leveraged to control for doctor shopping. We analyzed articles that covered a 10-year time period from four databases and reported using preferred reporting items for systematic reviews and meta-analysis (PRISMA). We compared intervention, study design, and bias, in addition to showing intervention interactions with facilitators, barriers, and medical outcomes. From 42 articles published from six countries, we identified seven interventions, five facilitator themes with two individual observations, three barrier themes with six individual observations, and two medical outcome themes with four individual observations. Multiple HIT mechanisms exist to control for doctor shopping. Some are associated with a decrease in overdose mortality, but access is not universal or compulsory, and data sharing is sporadic. Because shoppers travel hundreds of miles in pursuit of prescription drugs, data sharing should be an imperative. Research supports leveraging HIT to control doctor shopping, yet without robust data sharing agreements, the efforts of the system are limited to the efforts of the entity with the least number of barriers to their goal. Shoppers will seek out and exploit that organization that does not require participation or checking of prescription drug monitoring programs (PDMP), and the research shows that they will drive great distances to exploit this weakest link.