Olena Mazurenko

Background: Clinical decision support (CDS) systems may improve chronic pain management by providing guideline-based recommendations within clinical workflows. However, low clinician use can limit their effectiveness. Understanding the association between CDS use and guideline-recommended care may help distinguish implementation challenges from limitations in CDS functionality. Objective: To estimate the association of OneSheet, a CDS system for chronic pain management, and guideline-recommended care when the system was used during patient encounters. Methods: We conducted a per-protocol secondary data analysis of data from a pragmatic randomized clinical trial conducted from October 2020 to May 2022 across 25 primary care clinics at 2 academic health systems in Indiana and North Carolina. The parent trial enrolled 137 primary care clinicians. This analysis included 145511 eligible chronic pain encounters and 15480 long term opioid therapy (LTOT) encounters that occurred after OneSheet implementation. OneSheet use at a patient encounter was the exposure of interest. Propensity score matching was used to create comparable groups of encounters with and without OneSheet use. Outcomes included documentation of pain-related goals, documentation of pain and function using the Pain, Enjoyment of Life and General Activity (PEG) scale, prescription drug monitoring (PDMP) review, urine drug screening (UDS) orders, and naloxone orders. Results: OneSheet was used in 959 of 145511 chronic pain encounters (0.7%) and 582 of 15480 LTOT encounters (4.0%). In matched analyses, OneSheet use was associated with increased documentation of pain-related goals (relative risk (RR)=11.70 (95% CI 4.25-32.50) and PEG score documentation (RR=9.70 (95% CI 4.14-22.31). Among LTOT encounters, OneSheet use was associated with increased UDS ordering (RR=1.52 (95% CI 1.40-1.65) and PDMP review (RR=1.06 (95% CI 1.03-1.09), but decreased naloxone ordering (RR=0.13 (95% CI, 0.02-0.90). Conclusions: OneSheet use was associated with several guideline-recommended chronic pain management actions, suggesting that they system may support care quality when clinicians use it. However, use was rare, indicating that increasing routine use through better workflow integration may be necessary to achieve population-level improvements in chronic pain care.

This study aimed to assess the effect of a passive, opt-in electronic health record (EHR)-based clinical decision support (CDS), the Chronic Pain OneSheet, on guideline-recommended chronic pain management in primary care. A pragmatic randomized controlled trial with a parallel group design was conducted between October 2020 and May 2022. Participants were 137 primary care clinicians (PCCs) treating qualifying patients with chronic pain at 25 primary care clinics within two academic health systems in the United States. PCCs were randomized in the EHR to have access to OneSheet or usual care. OneSheet aggregates guideline-relevant information in a single view and provides shortcuts to guideline-recommended actions (e.g., ordering urine drug screening [UDS] for patients prescribed opioids). We constructed five visit-level binary outcomes: (1) documenting pain-related goals; (2) documenting pain and function via Pain, Enjoyment of Life and General Activity (PEG) scale; (3) reviewing prescription drug monitoring programs (PDMPs); (4) ordering UDS; and (5) ordering naloxone. Analysis used generalized linear mixed models for each outcome. OneSheet access minimally increased rates of pain-related goal documentation (0.2 percentage point increase, p = 0.013), PEG scale documentation (0.7 percentage point increase, p < 0.001), and UDS orders (2.2 percentage point increase, p = 0.006). OneSheet access decreased the rate of ordering naloxone (0.5 percentage point decrease, p < 0.001). OneSheet access did not affect PDMP review rates (0.5 percentage point decrease, p = 0.382). OneSheet access did not result in clinically significant improvements in guideline-recommended management of chronic pain in primary care despite a robust user-centered design incorporating clinician input and EHR integration. Several factors likely limited OneSheet effectiveness, including limited ability to target certain patient visits, workflow limits on data collection and ordering, and evolving COVID-19 and opioid epidemic-related policies and procedures. These findings highlight specific limitations of OneSheet and the broader challenges of implementing effective EHR-based CDS in complex health care environments.

Objectives: To improve the identification of patients with health-related social needs (HRSNs) in the emergency department (ED), we developed and integrated a risk prediction score into an existing Fast Healthcare Interoperability Resources (FHIR)-based clinical decision support (CDS). Materials and Methods: We conducted 2 phases of individual semi-structured qualitative interviews with ED clinicians to identify HRSN risk score design preferences for CDS integration. Following this, we used patient HRSN screening survey, health information exchange (HIE), and clinical data to run logistic regressions, developing an HRSN risk score aligned with ED clinician preferences. Results: Emergency department clinicians preferred HRSN risk scores displayed via visual cues like color-coding with different ranges (low, medium, and high) with higher model sensitivity to avoid missing patients with HRSNs. The overall performance of the risk prediction model was modest. Risk scores for food insecurity, transportation barriers, and financial strain were more sensitive, aligning with users’ preference for inclusivity and accurately identifying patients likely to screen positive for these HRSNs. Discussion: The design and risk score model choices, such as visual displays with additional data, higher sensitivity thresholds, and use of different thresholds for fairness, may support effective CDS use by ED clinicians. Conclusion: Using HIE data and an external CDS is a feasible route for including patient HRSNs information in the ED. We relied on clinician preferences for incorporation into the existing CDS and were attentive to performance fairness. While the predictive performance of our risk score is modest, providing risk scores in this manner may potentially improve the identification of patients’ HRSNs in the ED.

Health-related social needs (HRSNs) encompass various non-medical risks from a patient’s life circumstances. The emergency department (ED) is a crucial yet challenging setting for addressing patient HRSNs, a clinical decision support (CDS) intervention could assist in identifying patients at high risk of having HRSNs. This project aims to implement and evaluate a CDS intervention that offers ED clinicians risk prediction scores to determine which patients will likely screen positive for one or more HRSNs. The FHIR-based CDS intervention, implemented in the ED setting of a health system in Indianapolis, Indiana, will use health information exchange data to generate logit-derived probability scores that estimate an adult patient’s likelihood of screening positive for each of the following HRSNs: housing instability, food insecurity, transportation barriers, financial strain, and history of legal involvement. For each HRSN, ED clinicians will have access to the patient’s likelihood of screening positive categorized as “high,” “medium,” or “low” based on tertiles in the distribution of each likelihood score. Clinician participation in the CDS will be voluntary. The intervention’s effects will be assessed using a difference-in-difference approach with a pre-post design and a propensity-matched comparison group of ED patients from the same metropolitan area. Outcomes of interest include whether a formal HRSN screening was conducted, whether a referral was made to an HRSN service provider (e.g., social worker), and whether a repeat ED revisit (at 3, 7, and 30 days) or primary care follow-up (within 7 days) occurred. Efficiently and accurately identifying patients with HRSNs could help link them to needed services, improving outcomes and reducing healthcare costs. This protocol will contribute to a growing body of research on the role of CDS interventions in facilitating improved screenings and referrals for HRSNs. Clincialtrials.gov NCT06655974

Background: Health-related social needs (HRSNs), such as housing instability, food insecurity, and financial strain, are increasingly prevalent among patients. Healthcare organizations must first correctly identify patients with HRSNs to refer them to appropriate services or offer resources to address their HRSNs. Yet, current identification methods are suboptimal, inconsistently applied, and cost prohibitive. Machine learning (ML) predictive modeling applied to existing data sources may be a solution to systematically and effectively identify patients with HRSNs. The performance of ML predictive models using data from electronic health records (EHRs) and other sources has not been compared to other methods of identifying patients needing HRSN services. Methods: A screening questionnaire that included housing instability, food insecurity, transportation barriers, legal issues, and financial strain was administered to adult ED patients at a large safety-net hospital in the mid-Western United States (n = 1,101). We identified those patients likely in need of HRSN-related services within the next 30 days using positive indications from referrals, encounters, scheduling data, orders, or clinical notes. We built an XGBoost classification algorithm using responses from the screening questionnaire to predict HRSN needs (screening questionnaire model). Additionally, we extracted features from the past 12 months of existing EHR, administrative, and health information exchange data for the survey respondents. We built ML predictive models with these EHR data using XGBoost (ML EHR model). Out of concerns of potential bias, we built both the screening question model and the ML EHR model with and without demographic features. Models were assessed on the validation set using sensitivity, specificity, and Area Under the Curve (AUC) values. Models were compared using the Delong test. Results: Almost half (41%) of the patients had a positive indicator for a likely HRSN service need within the next 30 days, as identified through referrals, encounters, scheduling data, orders, or clinical notes. The screening question model had suboptimal performance, with an AUC = 0.580 (95%CI = 0.546, 0.611). Including gender and age resulted in higher performance in the screening question model (AUC = 0.640; 95%CI = 0.609, 0.672). The ML EHR models had higher performance. Without including age and gender, the ML EHR model had an AUC = 0.765 (95%CI = 0.737, 0.792). Adding age and gender did not improve the model (AUC = 0.722; 95%CI = 0.744, 0.800). The screening questionnaire models indicated bias with the highest performance for White non-Hispanic patients. The performance of the ML EHR-based model also differed by race and ethnicity. Conclusion: ML predictive models leveraging several robust EHR data sources outperformed models using screening questions only. Nevertheless, all models indicated biases. Additional work is needed to design predictive models for effectively identifying all patients with HRSNs.

Background Health-related social needs (HRSNs) are the unmet social and economic needs (e.g., housing instability) that affect individuals' health and well-being. HRSNs are associated with more emergency department (ED) visits, longer stays, and worse health outcomes. More than a third of ED patients have at least one HRSN, yet patients are rarely screened for HRSNs in the ED. A clinical decision support (CDS) system with predictive modeling offers a promising approach to identifying patients systematically and efficiently with HRSNs in the ED. Objective This study aimed to identify ED clinician and staff preferences for designing and implementing an HRSN-related CDS system. Methods A multistep, user-centered design study involving qualitative semistructured interviews, observations of ED workflows, and a multidisciplinary design workshop. Results We conducted 16 semistructured interviews with ED clinicians and staff. Following the interviews, three research team members observed ED workflows, focusing on patient entry and clinician and staff usage of the electronic health record (EHR) system. Finally, we conducted a 3-hour multidisciplinary design workshop. An HRSN-related CDS system should be visually appealing, color-coordinated, and easily accessible in the EHR. An HRSN-related CDS system should target a select group of ED patients (to be discharged from the ED) and highlight a select set of critical HRSN issues early in the workflow to adjust clinical care adequately. An HRSN-related CDS system should provide a list of actions and the ability to notify the clinical team if the patient's HRSNs were addressed. Conclusion The user-centered design identified a set of specific preferences for an HRSN-related CDS system to be implemented in the ED. Future work will focus on implementing and refining the CDS system and assessing the rates of changes in clinical care (e.g., rates of referrals) to address patient HRSNs in the ED.

Objective: To determine whether availability of behavioral health crisis care services is associated with changes in emergency department (ED) utilization.Data Sources and Study Setting: We used longitudinal panel data (2016–2021) on ED utilization from the Healthcare Cost and Utilization Project's State ED Databases and a novel dataset on crisis care services compiled using information from the Substance Abuse and Mental Health Services Administration's National Directories of Mental Health Treatment Facilities. A total of 1002 unique zip codes from Arizona, Florida, Kentucky, Maryland, and Wisconsin were included in our analyses.Study Design: To estimate the effect of crisis care availability on ED utilization, we used a linear regression model with zip code and year fixed effects and standard errors accounting for clustering at the zip code‐level. ED utilization related to mental, behavioral, and neurodevelopmental (MBD) disorders served as our primary outcome. We also examined pregnancy‐related ED utilization as a nonequivalent dependent variable to assess residual bias in effect estimates.Data Collection/Extraction Methods: We extracted data on crisis care services offered by mental health treatment facilities (n = 14,726 facility‐years) from the National Directories. MBD‐related ED utilization was assessed by applying the Clinical Classification Software Refined from the Healthcare Cost and Utilization Project to the primary ICD‐10‐CM diagnosis code on each ED encounter (n = 101,360,483). All data were aggregated to the zip code‐level (n = 6012 zip‐years).Principal Findings: The overall rate of MBD‐related ED visits between 2016 and 2021 was 1610 annual visits per 100,000 population. Walk‐in crisis stabilization services were associated with reduced MBD‐related ED utilization (coefficient = −0.028, p = 0.009), but were not significantly associated with changes in pregnancy‐related ED utilization.Conclusions: Walk‐in crisis stabilization services were associated with reductions in MBD‐related ED utilization. Decision‐makers looking to reduce MBD‐related ED utilization should consider increasing access to this promising alternative model.

Rationale: Chronic non‐cancer pain (CNCP) is a leading driver of disability. Primary care clinicians treat most patients with CNCP. Yet, they are often unable to identify appropriate pain treatments, mainly due to concerns about the safety and effectiveness of available medications. Clinical practice guidelines (CPGs) can be useful tools to guide primary care clinicians in selecting pain treatments based on the best available evidence.Objectives: To undertake a systematic review of CPGs that address the management of adults with CNCP, regardless of underlying condition type, in primary care.Method: We systematically reviewed and synthesised current CPGs for managing adults with CNCP in primary care (2013–2023). We followed a stepwise systematic process to synthesise key CPG recommendations: extracted and analysed each recommendation, synthesised by compiling similar recommendations using a thematic analysis approach, and assessed the strength of CPG recommendations to create a final, unified set of recommendations. We focused on identifying CPGs containing recommendations on the following topics: (a) opioid pain management, (b) non‐opioid pharmacological pain management, (c) non‐pharmacological pain management, and (d) patient‐centred communication around pain management, prevention, and organisation of care.Results: We included 13 CPGs, 8 of which focused solely on use of opioids, emphasising the lack of long‐term effectiveness and safety concerns, being mainly based on the expert consensus. As an exception, high‐quality evidence recommended referring patients with suspected opioid use disorder to specialist addiction services for medication‐assisted treatment. Recommendations for non‐opioid pain management were often contradictory and based on the expert consensus. Patient‐centred pain management combined with exercise‐based interventions and psychological therapies are appropriate strategies for managing patients with CNCP.Conclusion: Most CPGs focused on opioid management, with contradictory recommendations for non‐opioid management based on low‐quality evidence. Additional research is needed to strengthen the evidence for using non‐opioid and non‐pharmacological interventions to manage patients with CNCP.