Assessment of utilization efficiency using machine learning techniques: A study of heterogeneity in preoperative healthcare utilization among super-utilizers

Hyer, J. Madison; Paredes, Anghela Z.; White, Susan; Ejaz, Aslam; Pawlik, Timothy M.

doi:10.1016/j.amjsurg.2020.01.043

Cited by 17 publications

(32 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some interventions were assessed in relation to two separate and distinct HNHC patient cohorts who received the same treatment. [64][65][66][67][68] Other interventions looked at outcomes for HNHC patients, distinct from a more general patient population; some patients might have been included in multiple cohorts constructed for separate studies. 69,70,71 Still other interventions conducted multiple analyses on different subsets of their HNHC patient population to answer different research questions.…”

Section: Data Extraction and Risk-of-bias/risk-of-rigor Assessmentsmentioning

confidence: 99%

“…76,79,80,82,85,124,[136][137][138]159 However, while these data are generally considered necessary, they are not sufficient for identifying the population. The data on prior cost or use have been described as being limited to providing "broad brush information" because no algorithm or predictive model specification has been found that is able to reliably predict future high use across patients with chronic disease 66,133,137 (see also KQ 1 findings). One barrier to developing accurate projections is "regression to the mean" (a patient with high cost and use in a baseline period using closer to the average level of service and costs in the followup period).…”

Section: Determining Hnhc Patients Using Datamentioning

confidence: 99%

See 1 more Smart Citation

Management of High-Need, High-Cost Patients: A “Best Fit” Framework Synthesis, Realist Review, and Systematic Review

Berkman¹,

Chang²,

Seibert³

et al. 2021

View full text Add to dashboard Cite

Background. In the United States, patients referred to as high-need, high-cost (HNHC) constitute a very small percentage of the patient population but account for a disproportionally high level of healthcare use and cost. Payers, health systems, and providers would like to improve the quality of care and health outcomes for HNHC patients and reduce their costly use of potentially preventable or modifiable healthcare services, including emergency department (ED) and hospital visits. Methods. We assessed evidence of criteria that identify HNHC patients (best fit framework synthesis); developed program theories on the relationship among contexts, mechanisms, and outcomes of interventions intended to change HNHC patient behaviors (realist review); and assessed the effectiveness of interventions (systematic review). We searched databases, gray literature, and other sources for evidence available from January 1, 2000, to March 4, 2021. We included quantitative and qualitative studies of HNHC patients (high healthcare use or cost) age 18 and over who received intervention services in a variety of settings. Results. We included 110 studies (117 articles). Consistent with our best fit framework, characteristics associated with HNHC include patient chronic clinical conditions, behavioral health factors including depression and substance use disorder, and social risk factors including homelessness and poverty. We also identified prior healthcare use and race as important predictors. We found limited evidence of approaches for distinguishing potentially preventable or modifiable high use from all high use. To understand how and why interventions work, we developed three program theories in our realist review that explain (1) targeting HNHC patients, (2) engaging HNHC patients, and (3) engaging care providers in these interventions. Theories identify the need for individualizing and tailoring services for HNHC patients and the importance of building trusting relationships. For our systematic review, we categorized evidence based on primary setting. We found that ED-, primary care–, and home-based care models result in reduced use of healthcare services (moderate to low strength of evidence [SOE]); ED, ambulatory intensive caring unit, and primary care-based models result in reduced costs (low SOE); and system-level transformation and telephonic/mail models do not result in changes in use or costs (low SOE). Conclusions. Patient characteristics can be used to identify patients who are potentially HNHC. Evidence focusing specifically on potentially preventable or modifiable high use was limited. Based on our program theories, we conclude that individualized and tailored patient engagement and resources to support care providers are critical to the success of interventions. Although we found evidence of intervention effectiveness in relation to cost and use, the studies identified in this review reported little information for determining why individual programs work, for whom, and when.

show abstract

Section: Data Extraction and Risk-of-bias/risk-of-rigor Assessmentsmentioning

confidence: 99%

Section: Determining Hnhc Patients Using Datamentioning

confidence: 99%

Management of High-Need, High-Cost Patients: A “Best Fit” Framework Synthesis, Realist Review, and Systematic Review

Berkman¹,

Chang²,

Seibert³

et al. 2021

View full text Add to dashboard Cite

show abstract

“…A total of 31 reviewed studies (63.3%) evaluated the use of AI/ML applications in optimizing preoperative patient selection or projecting surgical costs, through prediction of hospital LOS, discharges, readmissions, and other cost-contributing factors ( Table 1 , Table 2 ). Sixteen studies (32.7%) evaluated AI/ML applications to accurately predict patient reoperations, operating time, hospital LOS, discharges, readmissions, or surgical and inpatient costs [ [14] , [15] , [16] , [17] , [18] , [19] , [20] , [21] , [22] , [23] , [24] , [25] , [26] , [27] , [28] , [29] ]. In addition, 16 studies (32.7%) used patients’ preoperative risk factors and other patient-specific variables to optimize the patient selection and surgical planning process through the use of AI/ML-based predictions of surgical outcomes and postoperative complications [ [30] , [31] , [32] , [33] , [34] , [35] , [36] , [37] , [38] , [39] , [40] , [41] , [42] , [43] , [44] ].…”

Section: Resultsmentioning

confidence: 99%

“…Several of the reviewed studies used AI/ML models to accurately predict the risk of a range of postoperative complications and adverse events [ 19 , 29 , [47] , [50] , [51] , [60] ]. TKA and total hip arthroplasty revisions and reoperations are also modeled with AI/ML algorithms in some studies, [ 15 , 16 , 21 , 64 ] as well as hospital readmissions [ 20 , 21 , 26 , 27 ]. In the postoperative period, AI/ML tools offer surgeons the ability to predict patients’ outcomes after surgery, including functional outcomes and PRO scores [ 14 , 32 , 33 , 43 , 45 , [48] , [53] , [54] , [57] , [58] , [59] , [61] ].…”

Section: Discussionmentioning

confidence: 99%

“…The majority of the decision support studies evaluated AI/ML model performance using receiver operating characteristic/AUC, accuracy, sensitivity, and specificity. Two studies did not test model performance, but instead used cluster analysis to classify patients based on preoperative risk factors and other variables to predict their association with inpatient costs and functional outcomes [20,45]. AI/ML applications of cost prediction were used in 23 models across 11 studies, which reported an average AUC of 0.77 (SD 0.08) (Table 3).…”

Section: Administrative and Clinical Decision Support Applicationsmentioning

confidence: 99%

See 1 more Smart Citation

Artificial Learning and Machine Learning Decision Guidance Applications in Total Hip and Knee Arthroplasty: A Systematic Review

et al. 2021

View full text Add to dashboard Cite

Background: Artificial intelligence (AI) and machine learning (ML) modeling in hip and knee arthroplasty (total joint arthroplasty [TJA]) is becoming more commonplace. This systematic review aims to quantify the accuracy of current AI-and ML-based application for cognitive support and decision-making in TJA. Methods: A comprehensive search of publications was conducted through the EMBASE, Medline, and PubMed databases using relevant keywords to maximize the sensitivity of the search. No limits were placed on level of evidence or timing of the study. Findings were reported according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Analysis of variance testing with post-hoc Tukey test was applied to compare the area under the curve (AUC) of the models. Results: After application of inclusion and exclusion criteria, 49 studies were included in this review. The application of AI/ML-based models and average AUC is as follows: cost prediction-0.77, LOS and discharges-0.78, readmissions and reoperations-0.66, preoperative patient selection/planning-0.79, adverse events and other postoperative complications-0.84, postoperative pain-0.83, postoperative patient-reported outcomes measures and functional outcome-0.81. Significant variability in model AUC across the different decision support applications was found (P < .001) with the AUC for readmission and reoperation models being significantly lower than that of the other decision support categories. Conclusions: AI/ML-based applications in TJA continue to expand and have the potential to optimize patient selection and accurately predict postoperative outcomes, complications, and associated costs. On average, the AI/ML models performed best in predicting postoperative complications, pain, and patientreported outcomes and were less accurate in predicting hospital readmissions and reoperations.

show abstract

Emergency Department Utilization Following Hepatopancreatic Surgery Among Medicare Beneficiaries

Paro

Dalmacy

Hyer

et al. 2021

Journal of Gastrointestinal Surgery

View full text Add to dashboard Cite

Assessment of utilization efficiency using machine learning techniques: A study of heterogeneity in preoperative healthcare utilization among super-utilizers

Cited by 17 publications

References 27 publications

Management of High-Need, High-Cost Patients: A “Best Fit” Framework Synthesis, Realist Review, and Systematic Review

Management of High-Need, High-Cost Patients: A “Best Fit” Framework Synthesis, Realist Review, and Systematic Review

Artificial Learning and Machine Learning Decision Guidance Applications in Total Hip and Knee Arthroplasty: A Systematic Review

Emergency Department Utilization Following Hepatopancreatic Surgery Among Medicare Beneficiaries

Contact Info

Product

Resources

About