Mitigating Sociodemographic Bias in Opioid Use Disorder Prediction: Fairness-Aware Machine Learning Framework.

BACKGROUND: Opioid use disorder (OUD) is a critical public health crisis in the United States, affecting >5.5 million Americans in 2021. Machine learning has been used to predict patient risk of incident OUD. However, little is known about the fairness and bias of these predictive models. OBJECTIVE: The aims of this study are two-fold: (1) to develop a machine learning bias mitigation algorithm for sociodemographic features and (2) to develop a fairness-aware weighted majority voting (WMV) classifier for OUD prediction. METHODS: We used the 2020 National Survey on Drug and Health data to develop a neural network (NN) model using stochastic gradient descent (SGD; NN-SGD) and an NN model using Adam (NN-Adam) optimizers and evaluated sociodemographic bias by comparing the area under the curve values. A bias mitigation algorithm, based on equality of odds, was implemented to minimize disparities in specificity and recall. Finally, a WMV classifier was developed for fairness-aware prediction of OUD. To further analyze bias detection and mitigation, we did a 1-N matching of OUD to non-OUD cases, controlling for socioeconomic variables, and evaluated the performance of the proposed bias mitigation algorithm and WMV classifier. RESULTS: Our bias mitigation algorithm substantially reduced bias with NN-SGD, by 21.66% for sex, 1.48% for race, and 21.04% for income, and with NN-Adam by 16.96% for sex, 8.87% for marital status, 8.45% for working condition, and 41.62% for race. The fairness-aware WMV classifier achieved a recall of 85.37% and 92.68% and an accuracy of 58.85% and 90.21% using NN-SGD and NN-Adam, respectively. The results after matching also indicated remarkable bias reduction with NN-SGD and NN-Adam, respectively, as follows: sex (0.14% vs 0.97%), marital status (12.95% vs 10.33%), working condition (14.79% vs 15.33%), race (60.13% vs 41.71%), and income (0.35% vs 2.21%). Moreover, the fairness-aware WMV classifier achieved high performance with a recall of 100% and 85.37% and an accuracy of 73.20% and 89.38% using NN-SGD and NN-Adam, respectively. CONCLUSIONS: The application of the proposed bias mitigation algorithm shows promise in reducing sociodemographic bias, with the WMV classifier confirming bias reduction and high performance in OUD prediction.

Self-help groups and opioid use disorder treatment: An investigation using a machine learning-assisted robust causal inference framework.

OBJECTIVES: This study investigates the impact of participation in self-help groups on treatment completion among individuals undergoing medication for opioid use disorder (MOUD) treatment. Given the suboptimal adherence and retention rates for MOUD, this research seeks to examine the association between treatment completion and patient-level factors. Specifically, we evaluated the causal relationship between self-help group participation and treatment completion for patients undergoing MOUD. METHODS: We used the Substance Abuse and Mental Health Services Administration's (SAMHSA) Treatment Episode Data Set: Discharges (TEDS-D) from 2015 to 2019. The data are filtered by the patient's opioid use history, demographics, treatment modality, and other relevant information. In this observational study, machine learning models (Lasso Regression, Decision Trees, Random Forest, and XGBoost) were developed to predict treatment completion. Outcome Adaptive Elastic Net (OAENet) was used to select confounders and outcome predictors, and the robust McNemars test was used to evaluate the causal relationship between self-help group participation and MOUD treatment completion. RESULTS: The machine-learning models showed a strong association between participation in self-help groups and treatment completion. Our causal analysis demonstrated an average treatment effect on treated (ATT) of 0.260 and a p-value < 0.0001 for the robust McNemars test. CONCLUSIONS: Our study demonstrates the importance of participation in self-help groups for MOUD treatment recipients. We found that participation in MOUD along with self-help groups caused higher chances of treatment completion than MOUD alone. This suggests that policymakers should consider further integrating self-help groups into the treatment for OUD to improve the adherence and completion rate.

An explainable machine learning framework for predicting the risk of buprenorphine treatment discontinuation for opioid use disorder among commercially insured individuals.

OBJECTIVES: Buprenorphine is an effective evidence-based medication for opioid use disorder (OUD). Yet premature discontinuation undermines treatment effectiveness, increasing the risk of mortality and overdose. We developed and evaluated a machine learning (ML) framework for predicting buprenorphine care discontinuity within 12 months following treatment initiation. METHODS: This retrospective study used United States (US) 2018-2021 MarketScan commercial claims data of insured individuals aged 18-64 who initiated buprenorphine between July 2018 and December 2020 with no buprenorphine prescriptions in the previous six months. We measured buprenorphine prescription discontinuation gaps of ≥30 days within 12 months of initiating treatment. We developed predictive models employing logistic regression, decision tree classifier, random forest, extreme gradient boosting, Adaboost, and random forest-extreme gradient boosting ensemble. We applied recursive feature elimination with cross-validation to reduce dimensionality and identify the most predictive features while maintaining model robustness. For model validation, we used several statistics to evaluate performance, such as C-statistics and precision-recall curves. We focused on two distinct treatment stages: at the time of treatment initiation and one and three months after treatment initiation. We employed SHapley Additive exPlanations (SHAP) analysis that helped us explain the contributions of different features in predicting buprenorphine discontinuation. We stratified patients into risk subgroups based on their predicted likelihood of treatment discontinuation, dividing them into decile subgroups. Additionally, we used a calibration plot to analyze the reliability of the models. RESULTS: A total of 30,373 patients initiated buprenorphine and 14.98% (4551) discontinued treatment. C-statistic varied between 0.56 and 0.76 for the first-stage models including patient-level demographic and clinical variables. Inclusion of proportion of days covered (PDC) measured after one month and three months following treatment initiation significantly increased the models' discriminative power (C-statistics: 0.60 to 0.82). Random forest (C-statistics: 0.76, 0.79 and 0.82 with baseline predictors, one-month PDC and three-months PDC, respectively) outperformed other ML models in discriminative performance in all stages (C-statistics: 0.56 to 0.77). Most influential risk factors of discontinuation included early stage medication adherence, age, and initial days of supply. CONCLUSION: ML algorithms demonstrated a good discriminative power in identifying patients at higher risk of buprenorphine care discontinuity. The proposed framework may help healthcare providers optimize treatment strategies and deliver targeted interventions to improve buprenorphine care continuity.

Investigation of Simulated Adherence in Long-Term Buprenorphine/Naloxone Treatment Patients.

BACKGROUND: Buprenorphine is a medication that is used to treat opioid use disorder by reducing withdrawal symptoms and cravings for opioids. Patients with poor adherence are at higher risk of relapse and overdose. Providers often test adherence through urine testing but are not aware of simulated adherence, where patients may directly add buprenorphine to the urine samples. As of now, there exists no literature on the simulated adherence practices for patients who stayed in the treatment for more than three months. METHODS: This study is a cross-sectional analysis of simulated adherence through urine toxicology results of 3950 patients undergoing buprenorphine/naloxone treatment. Simulated adherence was measured by the ratio of norbuprenorphine and buprenorphine <0.02 in the urine sample. Descriptive statistics as well as multivariate analysis was conducted to examine the relationship between patient information and outcomes. RESULTS: Out of 3950 patients, 411 (10.4%) had a history of one or more simulated adherence. On average, patients with multiple simulated adherences had 48.1% of their tests simulated, while on the contrary, patients with a single occurrence of simulated adherence had 17.6% of their tests simulated. Weekly testing and visit number of over 15 were associated with a higher likelihood of simulated adherence. CONCLUSION: The study demonstrates that simulated adherence is a recurring phenomenon among buprenorphine/naloxone treatment patients regardless of the duration in the treatment. Utilization of quantitative urine toxicology to identify simulated adherence will enable healthcare providers to formulate a more precise and effective treatment plan tailored to support individual patient needs.

Treatment Experiences for Patients Receiving Buprenorphine/Naloxone for Opioid Use Disorder: A Qualitative Study of Patients' Perceptions and Attitudes.

Background: Although buprenorphine/naloxone has been demonstrated to be an effective treatment for patients with opioid use disorder (OUD), treatment retention has been a challenge. This study extends what is presently a limited literature regarding patients' experiences with this medication and the implications for treatment retention. Methods: The study was conducted as a qualitative investigation of patients in treatment for OUD at the time of the study. Forty-three patients (27 men, 15 women, mean age 34.7) were recruited from three clinical settings, a community health center, an academically-based treatment site, and an independent substance abuse treatment facility. Most patients had returned to use in the past after attempts to become abstinent. Results: Patients generally reported positive experiences with this medication noting it helped to reduce opioid cravings quickly. As important considerations for treatment retention, patients emphasized a firm commitment to achieving abstinence when beginning treatment and a prescriber who is informed about and attentive to their emotional state. Diverging attitudes did exist regarding treatment duration as some patients were accepting of long-term treatment while others desired a relatively brief option. Among patients who had returned to use, potentially important issues emerged pertaining to the absence of patient outreach for missed medication appointments and inadequate discharge planning following stays at rehabilitation facilities. Conclusions: While results regarding the importance of patient motivation and strong patient-prescriber relationships have been noted in previous studies, other findings regarding opportunities to improve patient outreach and coordination of care have received relatively less attention and warrant further consideration.

Long-term patient outcomes following buprenorphine/naloxone treatment for opioid use disorder: a retrospective analysis in a commercially insured population.

Background: While buprenorphine/naloxone (buprenorphine) has been demonstrated to be an effective medication for treating opioid use disorder (OUD), an important question exists about how long patients should remain in treatment.Objective: To examine the relationship between treatment duration and patient outcomes for individuals with OUD who have been prescribed buprenorphine.Methods: We conducted a retrospective, longitudinal study using the Massachusetts All Payer Claims Database, 2013 to 2017. The study comprised over 2,500 patients, approximately one-third of whom were female, who had been prescribed buprenorphine for OUD. The outcomes were hospitalizations and emergency room (ER) visits at 36 months following treatment initiation and 12 months following treatment discontinuation. Patients were classified into four groups based on treatment duration and medication adherence: poor adherence, duration <12 months; good adherence, duration <6 months; good adherence, duration 6 to 12 months, and good adherence, duration >12 months. We conducted analyses at the patient level of the relationship between duration and outcomes.Results: Better outcomes were observed for patients whose duration was greater than 12 months. Patients in the other groups had higher odds of hospitalization at 36 months following treatment initiation: poor adherence (2.71), <6 months (1.53), and 6 to 12 months (1.42). They also had higher odds of ER visits: poor adherence (1.69), <6 months (1.51), and 6 to 12 months (1.30). Similar results were observed following treatment discontinuation.Conclusions: OUD treatment with buprenorphine should be continued for at least 12 months to reduce hospitalizations and ED visits.

A machine learning based two-stage clinical decision support system for predicting patients' discontinuation from opioid use disorder treatment: retrospective observational study.

BACKGROUND: Buprenorphine is a widely used treatment option for patients with opioid use disorder (OUD). Premature discontinuation from this treatment has many negative health and societal consequences. OBJECTIVE: To develop and evaluate a machine learning based two-stage clinical decision-making framework for predicting which patients will discontinue OUD treatment within less than a year. The proposed framework performs such prediction in two stages: (i) at the time of initiating the treatment, and (ii) after two/three months following treatment initiation. METHODS: For this retrospective observational analysis, we utilized Massachusetts All Payer Claims Data (MA APCD) from the year 2013 to 2015. Study sample included 5190 patients who were commercially insured, initiated buprenorphine treatment between January and December 2014, and did not have any buprenorphine prescription at least one year prior to the date of treatment initiation in 2014. Treatment discontinuation was defined as at least two consecutive months without a prescription for buprenorphine. Six machine learning models (i.e., logistic regression, decision tree, random forest, extreme-gradient boosting, support vector machine, and artificial neural network) were tested using a five-fold cross validation on the input data. The first-stage models used patients' demographic information. The second-stage models included information on medication adherence during the early phase of treatment based on the proportion of days covered (PDC) measure. RESULTS: A substantial percentage of patients (48.7%) who started on buprenorphine discontinued the treatment within one year. The area under receiving operating characteristic curve (C-statistic) for the first stage models varied within a range of 0.55 to 0.59. The inclusion of knowledge regarding patients' adherence at the early treatment phase in terms of two-months and three-months PDC resulted in a statistically significant increase in the models' discriminative power (p-value < 0.001) based on the C-statistic. We also constructed interpretable decision classification rules using the decision tree model. CONCLUSION: Machine learning models can predict which patients are most at-risk of premature treatment discontinuation with reasonable discriminative power. The proposed machine learning framework can be used as a tool to help inform a clinical decision support system following further validation. This can potentially help prescribers allocate limited healthcare resources optimally among different groups of patients based on their vulnerability to treatment discontinuation and design personalized support systems for improving patients' long-term adherence to OUD treatment.

Patterns of patient discontinuation from buprenorphine/naloxone treatment for opioid use disorder: A study of a commercially insured population in Massachusetts.

BACKGROUND: Research has shown buprenorphine/naloxone to be an effective medication for treating individuals with opioid use disorder. At the same time, treatment discontinuation rates are reportedly high though much of the extant evidence comes from studies of the Medicaid population. OBJECTIVES: To examine the pattern and determinants of buprenorphine/naloxone treatment discontinuation in a population of commercially insured individuals. RESEARCH DESIGN: We performed a retrospective observational analysis of Massachusetts All Payer Claims Data (MA APCD) covering years 2013 through 2017. We defined treatment discontinuation as a gap of 60 consecutive days without a prescription for buprenorphine/naloxone within a time frame of 24 months from the initiation of treatment. A mixed-effect Cox proportional hazard model examined the associated risk of discontinuing treatment with baseline predictors. SUBJECTS: A total of 5134 individuals who were commercially insured during the study period. MEASURES: Buprenorphine/naloxone treatment discontinuation. RESULTS: Overall 75% of individuals had discontinued treatment within two years of initiating treatment, and median time to discontinuation was 300 days. Patients aged between 18 and 24 years (HR = 1.436, 95%, CI = 1.240-1.663) and receiving treatment from prescribers with high panel-size (HR = 1.278, 95% CI = 1.112-1.468) had higher risk of discontinuing treatment. On the contrary, patients receiving treatment from multiple prescribers had lower associated risk of treatment discontinuation. CONCLUSIONS: A substantial percentage of patients discontinue treatment well before they can typically meet criteria for sustained remission. Further investigations should assess the clinical outcomes following premature discontinuation and identify strategies for retaining patients in treatment.

Patterns of buprenorphine/naloxone prescribing: an analysis of claims data from Massachusetts.

Background: The brand name Suboxone and its generic formulation buprenorphine/naloxone is a medication for treating opioid use disorder. While this medication has been shown to be effective, little research has examined the extent to which it is being prescribed and under what circumstances.Objective: This study examined patterns of prescription claims for buprenorphine/naloxone in terms of volume and associated clinical conditions.Methods: The study was conducted using a statewide database comprising pharmacy and medical claims that were covered by commercial health insurance plans in Massachusetts between 2011 and 2015. Trends in prescription volume for buprenorphine/naloxone were assessed based on the annual number of patients with a prescription for buprenorphine/naloxone. To examine clinical conditions associated with buprenorphine/naloxone prescriptions, patients' pharmacy claims were linked to their medical claims within the prior three months. For patients with common pain-related conditions, the odds they were prescribed buprenorphine/naloxone rather than oxycodone, a widely used opioid for pain management, were also examined.Results: The number of patients with a buprenorphine/naloxone prescription increased substantially during the study period, from approximately 25,000 in 2011 to over 39,000 in 2015. The most common clinical condition associated with buprenorphine/naloxone prescribing was opioid use disorder, but a substantial percentage of prescriptions were preceded by diagnoses that included pain or were for pain alone.Conclusion: A substantial increase in the number of patients with a prescription for buprenorphine/naloxone was observed. While buprenorphine/naloxone is most frequently prescribed for opioid use disorder, clinicians also appear to prescribe it for pain, particularly for patients who may be at elevated risk for opioid use disorder.

Hospital Readmissions to Nonindex Hospitals: Patterns and Determinants Following the Medicare Readmission Reduction Penalty Program.

OBJECTIVE: To examine patterns and determinants of nonindex readmissions for Medicare as well as non-Medicare patients both before and immediately after the adoption of Medicare's Hospital Readmission Reduction Program (HRRP) in 2012. Nonindex readmissions are readmissions to hospitals that are different from the one from which the patient was discharged. METHODS: Observational analysis of statewide database from California comprising patient-level discharge reports. Mixed-effects logistic regression models examined the association between nonindex readmissions and both hospital- and patient-level characteristics. RESULTS: Nonindex readmissions for the population studied were approximately 25%, but the percentage of such readmissions was significantly higher for non-Medicare patients than those enrolled in Medicare. Nonindex readmissions were associated with several patient- and hospital-level characteristics from which patients were discharged. The adoption of the HRRP did not have any appreciable impact on the general pattern of nonindex readmissions. CONCLUSIONS: A substantial percentage of hospital readmissions are to nonindex hospitals, but the general pattern and determinants of these events have not changed following the adoption of the HRRP. As preventable readmissions continue to gain attention as a key quality indicator for hospital care, further investigations are needed to understand the potential value of nonindex readmissions as a quality indicator for hospital care.

Robust policy evaluation from large-scale observational studies.

Under the current policy decision making paradigm we make or evaluate a policy decision by intervening different socio-economic parameters and analyzing the impact of those interventions. This process involves identifying the causal relation between interventions and outcomes. Matching method is one of the popular techniques to identify such causal relations. However, in one-to-one matching, when a treatment or control unit has multiple pair assignment options with similar match quality, different matching algorithms often assign different pairs. Since all the matching algorithms assign pairs without considering the outcomes, it is possible that with the same data and same hypothesis, different experimenters can reach different conclusions creating an uncertainty in policy decision making. This problem becomes more prominent in the case of large-scale observational studies as there are more pair assignment options. Recently, a robust approach has been proposed to tackle the uncertainty that uses an integer programming model to explore all possible assignments. Though the proposed integer programming model is very efficient in making robust causal inference, it is not scalable to big data observational studies. With the current approach, an observational study with 50,000 samples will generate hundreds of thousands binary variables. Solving such integer programming problem is computationally expensive and becomes even worse with the increase of sample size. In this work, we consider causal inference testing with binary outcomes and propose computationally efficient algorithms that are adaptable for large-scale observational studies. By leveraging the structure of the optimization model, we propose a robustness condition that further reduces the computational burden. We validate the efficiency of the proposed algorithms by testing the causal relation between the Medicare Hospital Readmission Reduction Program (HRRP) and non-index readmissions (i.e., readmission to a hospital that is different from the hospital that discharged the patient) from the State of California Patient Discharge Database from 2010 to 2014. Our result shows that HRRP has a causal relation with the increase in non-index readmissions. The proposed algorithms proved to be highly scalable in testing causal relations from large-scale observational studies.

A Systematic Review on Healthcare Analytics: Application and Theoretical Perspective of Data Mining.

The growing healthcare industry is generating a large volume of useful data on patient demographics, treatment plans, payment, and insurance coverage—attracting the attention of clinicians and scientists alike. In recent years, a number of peer-reviewed articles have addressed different dimensions of data mining application in healthcare. However, the lack of a comprehensive and systematic narrative motivated us to construct a literature review on this topic. In this paper, we present a review of the literature on healthcare analytics using data mining and big data. Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, we conducted a database search between 2005 and 2016. Critical elements of the selected studies—healthcare sub-areas, data mining techniques, types of analytics, data, and data sources—were extracted to provide a systematic view of development in this field and possible future directions. We found that the existing literature mostly examines analytics in clinical and administrative decision-making. Use of human-generated data is predominant considering the wide adoption of Electronic Medical Record in clinical care. However, analytics based on website and social media data has been increasing in recent years. Lack of prescriptive analytics in practice and integration of domain expert knowledge in the decision-making process emphasizes the necessity of future research.