Search

Peer comparison feedback is a promising strategy for reducing opioid prescribing and opioid-related harms. Such comparisons may be particularly impactful among underestimating clinicians who do not perceive themselves as high prescribers relative to their peers. But peer comparisons could also unintentionally increase prescribing among overestimating clinicians who do not perceive themselves as lower prescribers than peers. The objective of this study was to assess if the impact of peer comparisons varied by clinicians' preexisting opioid prescribing self-perceptions. Subgroup analysis of a randomized trial of peer comparison interventions among emergency department and urgent care clinicians was used. Generalized mixed-effects models were used to assess whether the impact of peer comparisons, alone or combined with individual feedback, varied by underestimating or overestimating prescriber status. Underestimating and overestimating prescribers were defined as those who self-reported relative prescribing amounts that were lower and higher, respectively, than actual relative baseline amounts. The primary outcome was pills per opioid prescription. Among 438 clinicians, 54% (n = 236) provided baseline prescribing self-perceptions and were included in this analysis. Overall, 17% (n = 40) were underestimating prescribers whereas 5% (n = 11) were overestimating prescribers. Underestimating prescribers exhibited a differentially greater decrease in pills per prescription compared to nonunderestimating clinicians when receiving peer comparison feedback (1.7 pills, 95% CI, -3.2 to -0.2 pills) or combined peer and individual feedback (2.8 pills, 95% CI, -4.8 to -0.8 pills). In contrast, there were no differential changes in pills per prescription for overestimating versus nonoverestimating prescribers after receiving peer comparison (1.5 pills, 95% CI, -0.9 to 3.9 pills) or combined peer and individual feedback (3.0 pills, 95% CI, -0.3 to 6.2 pills). Peer comparisons were more impactful among clinicians who underestimated their prescribing compared to peers. By correcting inaccurate self-perceptions, peer comparison feedback can be an effective strategy for influencing opioid prescribing., (Copyright © 2023 the American College of Medical Quality.)

An initial opioid prescription with a greater number of pills is associated with a greater risk for future long-term opioid use, yet few interventions have reliably influenced individual clinicians' prescribing. Our objective was to evaluate the effect of feedback interventions for clinicians in reducing opioid prescribing. The interventions included feedback on a clinician's outlier prescribing (individual audit feedback), peer comparison, and both interventions combined. We conducted a four-arm factorial pragmatic cluster randomized trial at forty-eight emergency department (ED) and urgent care (UC) sites in the western US, including 263 ED and 175 UC clinicians with 294,962 patient encounters. Relative to usual care, there was a significant decrease in pills per prescription both for peer comparison feedback (-0.8) and for the combination of peer comparison and individual audit feedback (-1.2). This decrease was sustained during follow-up. There were no significant changes for individual audit feedback alone, and no interventions changed the proportion of encounters with an opioid prescription.

Objective: The purpose of this study was to determine whether patients who discuss bariatric surgery with their providers are more likely to undergo the procedure and to lose weight., Methods: A retrospective cohort study of adults with BMI ≥ 35 kg/m 2 treated between 2000 and 2015 was conducted to analyze the relationship between a discussion of bariatric surgery in the first year after study entry and weight changes (primary outcome) and receipt of bariatric surgery (secondary outcome) over 2 years after study entry. Natural language processing was used to identify the documentation of bariatric surgery discussion in electronic provider notes., Results: Out of 30,560 study patients, a total of 2,659 (8.7%) discussed bariatric surgery with their providers. The BMI of patients who discussed bariatric surgery decreased by 2.18 versus 0.21 for patients who did not (p < 0.001). In a multivariable analysis, patients who discussed bariatric surgery with their providers lost more weight (by 1.43 [change in BMI]; 95% CI: 1.29-1.57) and had greater odds (10.2; 95% CI: 9.0-11.6; p < 0.001) of undergoing bariatric surgery., Conclusions: Clinicians rarely discussed bariatric surgery with their patients. Patients who did have this discussion were more likely to lose weight and to undergo bariatric surgery., (© 2021 The Obesity Society.)

Background/objective: Risk-stratification tools for cardiac complications after noncardiac surgery based on preoperative risk factors are used to inform postoperative management. However, there is limited evidence on whether risk stratification can be improved by incorporating data collected intraoperatively, particularly for low-risk patients., Methods: We conducted a retrospective cohort study of adults who underwent noncardiac surgery between 2014 and 2018 at four hospitals in the United States. Logistic regression with elastic net selection was used to classify in-hospital major adverse cardiovascular events (MACE) using preoperative and intraoperative data ("perioperative model"). We compared model performance to standard risk stratification tools and professional society guidelines that do not use intraoperative data., Results: Of 72,909 patients, 558 (0.77%) experienced MACE. Those with MACE were older and less likely to be female. The perioperative model demonstrated an area under the receiver operating characteristic curve (AUC) of 0.88 (95% CI, 0.85-0.92). This was higher than the Lee Revised Cardiac Risk Index (RCRI) AUC of 0.79 (95% CI, 0.74-0.84; P < .001 for AUC comparison). There were more MACE complications in the top decile (n = 1,465) of the perioperative model's predicted risk compared with that of the RCRI model (n = 58 vs 43). Additionally, the perioperative model identified 2,341 of 7,597 (31%) patients as low risk who did not experience MACE but were recommended to receive postoperative biomarker testing by a risk factor-based guideline algorithm., Conclusions: Addition of intraoperative data to preoperative data improved prediction of cardiovascular complication outcomes after noncardiac surgery and could potentially help reduce unnecessary postoperative testing.

Importance: Acute kidney injury (AKI) is one of the most common complications after noncardiac surgery. Yet current postoperative AKI risk stratification models have substantial limitations, such as limited use of perioperative data., Objective: To examine whether adding preoperative and intraoperative data is associated with improved prediction of noncardiac postoperative AKI., Design, Setting, and Participants: A prognostic study using logistic regression with elastic net selection, gradient boosting machine (GBM), and random forest approaches was conducted at 4 tertiary academic hospitals in the United States. A total of 42 615 hospitalized adults with serum creatinine measurements who underwent major noncardiac surgery between January 1, 2014, and April 30, 2018, were included in the study. Serum creatinine measurements from 365 days before and 7 days after surgery were used in this study., Main Outcomes and Measures: Postoperative AKI (defined by the Kidney Disease Improving Global Outcomes within 7 days after surgery) was the primary outcome. The area under the receiver operating characteristic curve (AUC) was used to assess discrimination., Results: Among 42 615 patients who underwent noncardiac surgery, the mean (SD) age was 57.9 (15.7) years, 23 943 (56.2%) were women, 27 857 (65.4%) were white, and the most frequent surgery types were orthopedic (15 718 [36.9%]), general (8808 [20.7%]), and neurologic (6564 [15.4%]). The rate of postoperative AKI was 10.1% (n = 4318). The progressive addition of clinical data improved model performance across all modeling approaches, with GBM providing the highest discrimination by AUC. In GBM models, the AUC increased from 0.712 (95% CI, 0.694-0.731) using prehospitalization variables to 0.804 (95% CI, 0.788-0.819) using preoperative variables (inclusive of prehospitalization variables) (P < .001 for AUC comparison). The AUC further increased to 0.817 (95% CI, 0.802-0.832) when adding intraoperative variables (P < .001 for comparison vs model using preoperative variables). However, the statistically significant improvements in discrimination did not appear to be clinically significant. In particular, the AKI rate among patients classified as high risk improved from 29.1% to 30.0%, a net of 15 patients were appropriately reclassified as high risk, and an additional 15 patients were appropriately reclassified as low risk., Conclusions and Relevance: The findings of the study suggest that electronic health record data may be used to accurately stratify patients at risk of perioperative AKI, but the modest improvements from adding intraoperative data should be weighed against challenges in using intraoperative data.

We tested the value of adding data from the operating room to models predicting in-hospital death. We assessed model performance using two metrics, the area under the receiver operating characteristic curve (AUROC) and the area under the precision-recall curve (AUPRC), to illustrate the differences in information they convey in the setting of class imbalance. Data was collected on 74,147 patients who underwent major noncardiac surgery and 112 unique features were extracted from electronic health records. Sets of features were incrementally added to models using logistic regression, naïve Bayes, random forest, and gradient boosted machine methods. AUROC increased as more features were added, but changes were small for some modeling approaches. In contrast, AUPRC, which reflects positive predicted value, exhibited improvements across all models. Using AUPRC highlighted the added value of intraoperative data, not seen consistently with AUROC, and that with class imbalance AUPRC may serve as the more clinically relevant criterion.

Objective: To evaluate the prevalence of seven social factors using physician notes as compared to claims and structured electronic health records (EHRs) data and the resulting association with 30-day readmissions., Study Setting: A multihospital academic health system in southeastern Massachusetts., Study Design: An observational study of 49,319 patients with cardiovascular disease admitted from January 1, 2011, to December 31, 2013, using multivariable logistic regression to adjust for patient characteristics., Data Collection/extraction Methods: All-payer claims, EHR data, and physician notes extracted from a centralized clinical registry., Principal Findings: All seven social characteristics were identified at the highest rates in physician notes. For example, we identified 14,872 patient admissions with poor social support in physician notes, increasing the prevalence from 0.4 percent using ICD-9 codes and structured EHR data to 16.0 percent. Compared to an 18.6 percent baseline readmission rate, risk-adjusted analysis showed higher readmission risk for patients with housing instability (readmission rate 24.5 percent; p < .001), depression (20.6 percent; p < .001), drug abuse (20.2 percent; p = .01), and poor social support (20.0 percent; p = .01)., Conclusions: The seven social risk factors studied are substantially more prevalent than represented in administrative data. Automated methods for analyzing physician notes may enable better identification of patients with social needs., (© Health Research and Educational Trust.)

Aims: To design and validate a natural language processing algorithm to identify insulin therapy decline from the text of physician notes, and to determine the prevalence of insulin therapy decline and its impact on insulin initiation., Methods: We designed the algorithm using the publicly available natural language processing platform Canary. We evaluated the accuracy of the algorithm on 1501 randomly selected primary care physicians' notes from the electronic medical record system of a large academic medical centre. Using the validated language model, we then studied the prevalence of insulin therapy decline between 2000 and 2014., Results: The algorithm identified documentation of insulin therapy decline with a sensitivity of 100% (95% CI 82.4-100), a positive predictive value of 95% (95% CI 74.4-99.9), and a specificity of 99.9% (95% CI 99.6-100.0). We identified 3295 insulin-naïve adults with Type 2 diabetes who were recommended insulin therapy; 984 of them (29.9%) initially declined insulin. People with HbA 1c ≥ 75 mmol/mol (9.0%) were more likely [766/2239 (34.2%)] to have declined insulin than people with HbA 1c 53-63 mmol/mol (7.0-7.9%) and 64-74 mmol/mol (8.0-8.9%; P < 0.0001). Among the people who initially declined but ultimately started insulin [374/984 (38.0%)], mean time to insulin initiation was 790 days., Conclusions: Insulin therapy decline is common, potentially leading to progression of hyperglycaemia and a delay in achievement of glycaemic control. Further investigation is needed to determine the reasons, risk factors and long-term outcomes of this important clinical phenomenon., (© 2017 Diabetes UK.)

Background: Electronic health records are being increasingly used by nurses with up to 80% of the health data recorded as free text. However, only a few studies have developed nursing-relevant tools that help busy clinicians to identify information they need at the point of care., Objective: This study developed and validated one of the first automated natural language processing applications to extract wound information (wound type, pressure ulcer stage, wound size, anatomic location, and wound treatment) from free text clinical notes., Methods and Design: First, two human annotators manually reviewed a purposeful training sample (n=360) and random test sample (n=1100) of clinical notes (including 50% discharge summaries and 50% outpatient notes), identified wound cases, and created a gold standard dataset. We then trained and tested our natural language processing system (known as MTERMS) to process the wound information. Finally, we assessed our automated approach by comparing system-generated findings against the gold standard. We also compared the prevalence of wound cases identified from free-text data with coded diagnoses in the structured data., Results: The testing dataset included 101 notes (9.2%) with wound information. The overall system performance was good (F-measure is a compiled measure of system's accuracy=92.7%), with best results for wound treatment (F-measure=95.7%) and poorest results for wound size (F-measure=81.9%). Only 46.5% of wound notes had a structured code for a wound diagnosis., Conclusions: The natural language processing system achieved good performance on a subset of randomly selected discharge summaries and outpatient notes. In more than half of the wound notes, there were no coded wound diagnoses, which highlight the significance of using natural language processing to enrich clinical decision making. Our future steps will include expansion of the application's information coverage to other relevant wound factors and validation of the model with external data., (Copyright © 2016 Elsevier Ltd. All rights reserved.)

About 1 in 10 adults are reported to exhibit clinical depression and the associated personal, societal, and economic costs are significant. In this study, we applied the MTERMS NLP system and machine learning classification algorithms to identify patients with depression using discharge summaries. Domain experts reviewed both the training and test cases, and classified these cases as depression with a high, intermediate, and low confidence. For depression cases with high confidence, all of the algorithms we tested performed similarly, with MTERMS' knowledge-based decision tree slightly better than the machine learning classifiers, achieving an F-measure of 89.6%. MTERMS also achieved the highest F-measure (70.6%) on intermediate confidence cases. The RIPPER rule learner was the best performing machine learning method, with an F-measure of 70.0%, and a higher precision but lower recall than MTERMS. The proposed NLP-based approach was able to identify a significant portion of the depression cases (about 20%) that were not on the coded diagnosis list.