Characterization of high healthcare utilizer groups using administrative data from an electronic medical record database

Evaluation of a Structured Review Process for Emergency Department Return Visits with Admission

BACKGROUND

Review of emergency department (ED) revisits with admission allows the identification of improvement opportunities. Applying a health equity lens to revisits may highlight potential disparities in care transitions. Universal definitions or practicable frameworks for these assessments are lacking. The authors aimed to develop a structured methodology for this quality assurance (QA) process, with a layered equity analysis.

METHODS

The authors developed a classification instrument to identify potentially preventable 72-hour returns with admission (PPRA-72), accounting for directed, unrelated, unanticipated, or disease progression returns. A second review team assessed the instrument reliability. A self-reported race/ethnicity (R/E) and language algorithm was developed to minimize uncategorizable data. Disposition distribution, return rates, and PPRA-72 classifications were analyzed for disparities using Pearson chi-square and Fisher's exact tests.

RESULTS

The PPRA-72 rate was 4.8% for 2022 ED return visits requiring admission. Review teams achieved 93% agreement (κ = 0.51) for the binary determination of PPRA-72 vs. nonpreventable returns. There were significant differences between R/E and language in ED dispositions (p < 0.001), with more frequent admissions for the R/E White at the index visit and Other at the 72-hour return visit. Rates of return visits within 72 hours differed significantly by R/E (p < 0.001) but not by language (p = 0.156), with the R/E Black most frequent to have a 72-hour return. There were no differences between R/E (p = 0.446) or language (p = 0.248) in PPRA-72 rates. The initiative led to system improvements through informatics optimizations, triage protocols, provider feedback, and education.

CONCLUSION

The authors developed a review methodology for identifying improvement opportunities across ED 72-hour returns. This QA process enabled the identification of areas of disparity, with the continuous aim to develop next steps in ensuring health equity in care transitions.

Multimorbidity in middle-aged women and COVID-19: binary data clustering for unsupervised binning of rare multimorbidity features and predictive modeling

BACKGROUND

Multimorbidity is typically associated with deficient health-related quality of life in mid-life, and the likelihood of developing multimorbidity in women is elevated. We address the issue of data sparsity in non-prevalent features by clustering the binary data of various rare medical conditions in a cohort of middle-aged women. This study aims to enhance understanding of how multimorbidity affects COVID-19 severity by clustering rare medical conditions and combining them with prevalent features for predictive modeling. The insights gained can guide the development of targeted interventions and improved management strategies for individuals with multiple health conditions.

METHODS

The study focuses on a cohort of 4477 female patients, (aged 45-60) in Piedmont, Italy, and utilizes their multimorbidity data prior to the COVID-19 pandemic from their medical history from 2015 to 2019. The COVID-19 severity is determined by the hospitalization status of the patients from February to May 2020. Each patient profile in the dataset is depicted as a binary vector, where each feature denotes the presence or absence of a specific multimorbidity condition. By clustering the sparse medical data, newly engineered features are generated as a bin of features, and they are combined with the prevalent features for COVID-19 severity predictive modeling.

RESULTS

From sparse data consisting of 174 input features, we have created a low-dimensional feature matrix of 17 features. Machine Learning algorithms are applied to the reduced sparsity-free data to predict the Covid-19 hospital admission outcome. The performance obtained for the corresponding models are as follows: Logistic Regression (accuracy 0.72, AUC 0.77, F1-score 0.69), Linear Discriminant Analysis (accuracy 0.7, AUC 0.77, F1-score 0.67), and Ada Boost (accuracy 0.7, AUC 0.77, F1-score 0.68).

CONCLUSION

Mapping higher-dimensional data to a low-dimensional space can result in information loss, but reducing sparsity can be beneficial for Machine Learning modeling due to improved predictive ability. In this study, we addressed the issue of data sparsity in electronic health records and created a model that incorporates both prevalent and rare medical conditions, leading to more accurate and effective predictive modeling. The identification of complex associations between multimorbidity and the severity of COVID-19 highlights potential areas of focus for future research, including long COVID and intervention efforts.

Characterizing hospitalization trajectories in the high-need, high-cost population using electronic health record data

High utilization by a minority of patients accounts for a large share of health care costs, but the dynamics of this utilization remain poorly understood. We sought to characterize longitudinal trajectories of hospitalization among adult patients at an academic medical center from 2017 to 2023. Among 3404 patients meeting eligibility criteria, following an initial "rising-risk" period of 3 hospitalizations in 6 months, growth mixture modeling discerned 4 clusters of subsequent hospitalization trajectories: no further utilization, low chronic utilization, persistently high utilization with a slow rate of increase, and persistently high utilization with a fast rate of increase. Baseline factors associated with higher-order hospitalization trajectories included admission to a nonsurgical service, full code status, intensive care unit-level care, opioid administration, discharge home, and comorbid cardiovascular disease, end-stage kidney or liver disease, or cancer. Characterizing hospitalization trajectories and their correlates in this manner lays groundwork for early identification of those most likely to become high-need, high-cost patients.

Use of Care Guides to Reduce Emergency Department Visits by High-Frequency Utilizers

BACKGROUND

High-frequency utilizers are defined as patients who present 10 or more times to the emergency department in a rolling 12-month period. High-frequency utilizers contribute to emergency department overcrowding and misuse of resources, and reduce the efficiency of health care systems. Care guides have proven to be an effective tool in reducing high-frequency utilizers.

OBJECTIVE

The objective of this quality improvement project was to determine if implementing a care guide for high-frequency utilizers to address the core needs of the patient and facilitate resources through case management consultation decreases the number of visits and the cost of unreimbursed care to the emergency department from high-frequency utilizers.

METHODS

We implemented care guides for high-frequency utilizers in September 2014. Prior to initiating the care guides, we educated the physicians, nurses, case managers, and social workers in the emergency department.

RESULTS

Following the implementation of the care guides, there was a steady decline in the number of high-frequency utilizers (338 in 2013-68 in 2021), the number of total emergency department visits by high-frequency utilizers (6025 in 2013-1033 in 2021), and unreimbursed care ($2,068,063 in 2013-$589,298 in 2021).

CONCLUSION

The use of care guides was a successful strategy in reducing emergency department visits and the cost of unreimbursed care by high-frequency utilizers by providing them with the education and resources they require to receive health care services in appropriate settings.

Development and validation of a tool to predict high-need, high-cost patients hospitalised with ischaemic heart disease

OBJECTIVE

To develop and validate a tool to predict patients with ischaemic heart disease (IHD) at risk of excessive healthcare resource utilisation.

DESIGN

A retrospective cohort study.

SETTING

We identified patients through the State of Florida Agency for Health Care Administration (N=586 518) inpatient dataset.

PARTICIPANTS

Adult patients (at least 40 years of age) admitted to the hospital with a diagnosis of IHD between 1 January 2007 and 31 December 2016.

PRIMARY OUTCOME MEASURES

We identified patients whose healthcare utilisation is higher than presumed (analysis of residuals) and used logistic regression (binary and multinomial) in estimating the predictive models to classify individual as high-need, high-care (HNHC) patients relative to inpatient visits (frequency of hospitalisation), cost and hospital length of stay. Discrimination power, prediction accuracy and model improvement for the binary logistic model were assessed using receiver operating characteristic statistic, the Brier score and the log-likelihood (LL)-based pseudo-R2, respectively. LL-based pseudo-R2 and Brier score were used for multinomial logistic models.

RESULTS

The binary logistic model had good discrimination power (c-statistic=0.6496), an accuracy of probabilistic predictions (Brier score) of 0.0621 and an LL-based pseudo-R2 of 0.0338 in the development cohort. The model performed similarly in the validation cohort (c-statistic=0.6480), an accuracy of probabilistic predictions (Brier score) of 0.0620 and an LL-based pseudo-R2 of 0.0380. A user-friendly Excel-based HNHC risk predictive tool was developed and readily available for clinicians and policy decision-makers.

CONCLUSIONS

The Excel-based HNHC risk predictive tool can accurately identify at-risk patients for HNHC based on three measures of healthcare expenditures.

Prediction and analysis of time series data based on granular computing

The advent of the Big Data era and the rapid development of the Internet of Things have led to a dramatic increase in the amount of data from various time series. How to classify, correlation rule mining and prediction of these large-sample time series data has a crucial role. However, due to the characteristics of high dimensionality, large data volume and transmission lag of sensor data, large sample time series data are affected by multiple factors and have complex characteristics such as multi-scale, non-linearity and burstiness. Traditional time series prediction methods are no longer applicable to the study of large sample time series data. Granular computing has unique advantages in dealing with continuous and complex data, and can compensate for the limitations of traditional support vector machines in dealing with large sample data. Therefore, this paper proposes to combine granular computing theory with support vector machines to achieve large-sample time series data prediction. Firstly, the definition of time series is analyzed, and the basic principles of traditional time series forecasting methods and granular computing are investigated. Secondly, in terms of predicting the trend of data changes, it is proposed to apply the fuzzy granulation algorithm to first convert the sample data into coarser granules. Then, it is combined with a support vector machine to predict the range of change of continuous time series data over a period of time. The results of the simulation experiments show that the proposed model is able to make accurate predictions of the range of data changes in future time periods. Compared with other prediction models, the proposed model reduces the complexity of the samples and improves the prediction accuracy.

Characteristics of High-Resource Health System Users in Rural and Remote Regions: A Scoping Review

A small proportion of health care users are recognized to use a significantly higher proportion of health system resources, largely due to systemic, inequitable access and disproportionate health burdens. These high-resource health system users are routinely characterized as older, with multiple comorbidities, and reduced access to adequate health care. Geographic trends also emerge, with more rural and isolated regions demonstrating higher rates of high-resource use than others. Despite known geographical discrepancies in health care access and outcomes, health policy and research initiatives remain focused on urban population centers. To alleviate mounting health system pressure from high-resource users, their characteristics must be better understood within the context in which i arises. To examine this, a scoping review was conducted to provide an overview of characteristics of high-resource users in rural and remote communities in Canada and Australia. In total, 21 papers were included in the review. Using qualitative thematic coding, primary findings characterized rural high-resource users as those of an older age; with increased comorbid conditions and condition severity; lower socioeconomic status; and elevated risk behaviors.

Differential Impact of a Plan-Led Standardized Complex Care Management Intervention on Subgroups of High-Cost High-Need Medicaid Patients

Interventions to better coordinate care for high-need high-cost (HNHC) Medicaid patients frequently fail to demonstrate changes in hospitalizations or emergency department (ED) use. Many of these interventions are modeled after practice-level complex care management (CCM) programs. The authors hypothesized that a national CCM program may be effective for some subgroups of HNHC patients, and the overall null effect may obfuscate subgroup-level impact. They used a previously published typology defining 6 subgroups of high-cost Medicaid patients and evaluated program impact by subgroup. The analysis used an individual-level interrupted time series with a comparison group. Intervention subjects were high-cost adult Medicaid patients who enrolled in 1 of 2 national CCM programs implemented by UnitedHealthcare (UHC) (n = 39,687). The comparators were patients who met CCM program criteria but were ineligible due to current enrollment in another UHC/Optum led program (N = 26,359). The intervention was a CCM program developed by UHC/Optum to provide "whole person care" delivering standardized interventions to address medical, behavioral, and social needs for HNHC Medicaid patients, and the outcome was probability of hospitalization or ED use in a given month, estimated at 12 months postenrollment. A reduction in risk of ED utilization for 4 of 6 subgroups was found. A reduction in risk of hospitalization for 1 of 6 subgroups was also found. The authors conclude that standardized health plan led CCM programs demonstrate effectiveness for certain subgroups of HNHC patients in Medicaid. This effectiveness is principally in reducing ED risk and may extend to the risk of hospitalization for a small number of patients.

Novel integration of governmental data sources using machine learning to identify super-utilization among U.S. counties

Background

Super-utilizers consume the greatest share of resource intensive healthcare (RIHC) and reducing their utilization remains a crucial challenge to healthcare systems in the United States (U.S.). The objective of this study was to predict RIHC among U.S. counties, using routinely collected data from the U.S. government, including information on consumer spending, offering an alternative method for identifying super-utilization among population units rather than individuals.

Methods

Cross-sectional data from 5 governmental sources in 2017 were used in a machine learning pipeline, where target-prediction features were selected and used in 4 distinct algorithms. Outcome metrics of RIHC utilization came from the American Hospital Association and included yearly: (1) emergency rooms visit, (2) inpatient days, and (3) hospital expenditures. Target-prediction features included: 149 demographic characteristics from the U.S. Census Bureau, 151 adult and child health characteristics from the Centers for Disease Control and Prevention, 151 community characteristics from the American Community Survey, and 571 consumer expenditures from the Bureau of Labor Statistics. SHAP analysis identified important target-prediction features for 3 RIHC outcome metrics.

Results

2475 counties with emergency rooms and 2491 counties with hospitals were included. The median yearly emergency room visits per capita was 0.450 [IQR:0.318, 0.618], the median inpatient days per capita was 0.368 [IQR: 0.176, 0.826], and the median hospital expenditures per capita was $2104 [IQR: $1299.93, 3362.97]. The coefficient of determination (R2), calculated on the test set, ranged between 0.267 and 0.447. Demographic and community characteristics were among the important predictors for all 3 RIHC outcome metrics.

Conclusions

Integrating diverse population characteristics from numerous governmental sources, we predicted 3-outcome metrics of RIHC among U.S. counties with good performance, offering a novel and actionable tool for identifying super-utilizer segments in the population. Wider integration of routinely collected data can be used to develop alternative methods for predicting RIHC among population units.

Association of Function, Symptoms, and Social Support Reported in Standardized Outpatient Clinic Questionnaires With Subsequent Hospital Discharge Disposition and 30-Day Readmissions

OBJECTIVE

To determine whether patient-reported information, routinely collected in an outpatient setting, is associated with readmission within 30 days of discharge and/or the need for post-acute care after a subsequent hospital admission.

DESIGN

Retrospective cohort study. Six domains of patient-reported information collected in the outpatient setting (psychological distress, respiratory symptoms, musculoskeletal pain, family support, mobility, and activities of daily living [ADLs]) were linked to electronic health record hospitalization data. Mixed effects logistic regression models with random intercepts were used to identify the association between the 6 domains and outcomes.

SETTING

Outpatient clinics and hospitals in a Midwestern health system.

PARTICIPANTS

7671 patients who were hospitalized 11,445 times between May 2004 and May 2014 (N=7671).

INTERVENTION

None.

MAIN OUTCOME MEASURES

30-day hospital readmission and discharge home vs facility.

RESULTS

Domains were significantly associated with 30-day readmission and placement in a facility. Specifically, mobility (odds ratio [OR]=1.30; 95% confidence interval [CI], 1.16, 1.46), ADLs (OR=1.27; 95% CI, 1.13, 1.42), respiratory symptoms (OR=1.26; 95% CI, 1.12, 1.41), and psychological distress (OR=1.20; 95% CI, 1.07, 1.35) had the strongest associations with 30-day readmission. The ADL (OR=2.52; 95% CI, 2.26, 2.81), mobility (OR=2.35; 95% CI, 2.10, 2.63), family support (OR=2.28; 95% CI, 1.98, 2.62), and psychological distress (OR=1.38; 95% CI, 1.25, 1.52) domains had the strongest associations with discharge to an institution.

CONCLUSIONS

Patient-reported function, symptoms, and social support routinely collected in outpatient clinics are associated with future 30-day readmission and discharge to an institutional setting. Whether these data can be leveraged to guide interventions to address patient needs and improve outcomes requires further research.

High-cost users: drivers of inpatient healthcare expenditure concentration in urban China

Background

Total healthcare expenditures are concentrated among a small number of patients. To date, studies on the concentration of health care expenditures in developing countries are limited, mainly focusing on concentration measures and the demographic, clinical and socioeconomic characteristics of high-cost users (HCU). The drivers of the skewed overall distribution of health care expenditures are opaque. Using inpatient administrative claims data, this study provides new evidence on the concentration of healthcare expenditures in China; the demographic and clinical characteristics of high-cost users; and the drivers of the overall distribution of healthcare expenditures.

Methods

Utilizing administrative claims data for hospitalization in a prefecture-level city in China, we investigated the concentration of healthcare expenditure. We used recentered influence function (RIF) regression to examine the drivers of healthcare expenditure concentration, decomposing and estimating the effects of demographic and disease characteristics on the overall distribution of health care expenditures.

Results

Using a sample of 87,841 adults, we found extreme skewness in the distribution of inpatient medical expenditures in China, with approximately 49% of annual medical expenditures generated by the top 10% of inpatient groups. HCUs tend to be elderly and male, with high-frequency hospitalizations and long lengths of stay. In addition, healthcare expenditure concentration was related to diseases of the circulatory system, malignant neoplasms, diseases of the musculoskeletal system and connective tissue, diseases of the digestive system, injury and poisoning, and diseases of the respiratory system. Malignant and major diseases reinforced the concentration of healthcare spending, and a 10% increase in the prevalence of malignancy would result in a predicted Gini coefficient increase of 7.2%, heart disease of 0.92% and cerebrovascular disease of 1.5%. The above significant positive effects were not observed for hypertension and diabetes mellitus.

Conclusions

Our study provides new insights into the concentration of inpatient medical expenditures in China, including the precise picture of HCU expenditure concentration, the drivers of HCU expenditure concentration and the magnitude of their impact. With the aging of China's population and the profound shift in the disease spectrum, policymakers need to strengthen the early detection and intervention management of specific chronic diseases and high-risk populations, especially the early diagnosis and treatment of key cancers.

Identifying Characteristics Associated with the Concentration and Persistence of Medical Expenses among Middle-Aged and Elderly Adults: Findings from the China Health and Retirement Longitudinal Survey

Medical expenses, especially among middle-aged and elderly people, have increased in China over recent decades. However, few studies have analyzed the concentration or persistence of medical expenses among Chinese residents or vulnerable groups with longitudinal survey data. Based on the data of CHARLS (China Health and Retirement Longitudinal Study), this study sought to identify characteristics associated with the concentration and persistence of medical expenses among Chinese middle-aged and elderly adults and to help alleviate medical spending and the operational risk of social medical insurance. Concentration was measured using the cumulative percentages of ranked annual medical expenses and descriptive statistics were used to define the characteristics of individuals with high medical expenses. The persistence of medical expenses and associated factors were estimated using transfer rate calculations and Heckman selection modeling. The results show that total medical expenses were concentrated among a few adults and the concentration increased over time. People in the high medical expense group were more likely to be older, live in urban areas, be less wealthy, have chronic diseases, and attend higher-ranking medical institutions. Lagged medical expenses had a persistent positive effect on current medical expenses and the effect of a one-period lag was strongest. Individuals with chronic diseases during the lagged period had a higher likelihood of experiencing persistent medical expenses. Policy efforts should focus on preventive management, more efficient care systems, improvement of serious illness insurance level, and strengthening the persistent protection effect of social medical insurance to reduce the high medical financial risk and long-term financial healthcare burden in China.

Prevalence of medically unexplained symptoms in adults who are high users of healthcare services and magnitude of associated costs: a systematic review

INTRODUCTION

Medically unexplained symptoms (MUS) is a common clinical syndrome in primary and secondary healthcare service. Outcomes for patients with persistent MUS include increased disability, poorer quality of life and higher healthcare costs. The aim of this systematic review was to determine the prevalence of MUS in patients who are high users of healthcare or high-cost patients in comparison with routine users and the magnitude of associated costs.

DESIGN

A systematic review of the available literature.

DATA SOURCES AND ELIGIBILITY CRITERIA

The following electronic databases were systematically searched without language restriction from inception to June 2018 and updated on 22 October 2021: MEDLINE, PsycINFO, EMBASE, CINAHL and PROSPERO. Inclusion criteria included studies investigating adults aged ≥18 years, who were high healthcare users or accrued high healthcare costs, in which the prevalence and/or associated costs of MUS was quantified. Two reviewers independently extracted information on study characteristics, exposure and outcomes.

RESULTS

From 5622 identified publications, 25 studies from 9 countries involving 31 650 patients were selected for inclusion. Due to high risk of bias in many studies and heterogeneity between studies, results are described narratively. There were wide variations in prevalence estimates for MUS in high users of healthcare (2.9%-76%), but MUS was more prevalent in high use groups compared with low use groups in all but one of the 12 studies that included a comparator group. Only three studies investigated healthcare costs associated with MUS, and all three reported greater healthcare costs associated with MUS.

CONCLUSION

MUS has been found to be more prevalent in high use healthcare populations than comparator groups, but the magnitude of difference is difficult to estimate due to considerable heterogeneity between studies and potential for bias. Future studies should prioritise a standardised approach to this research area, with agreed definitions of MUS and high healthcare use.

PROSPERO REGISTRATION NUMBER

CRD42018100388.

Interpretable Identification of Comorbidities Associated with Recurrent ED and Inpatient Visits

In the hospital setting, a small percentage of recurrent frequent patients contribute to a disproportional amount of healthcare resource utilization. Moreover, in many of these cases, patient outcomes can be greatly improved by reducing re-occurring visits, especially when they are associated with substance abuse, mental health, and medical factors that could be improved by social-behavioral interventions, outpatient or preventative care. Additionally, health care costs can be reduced significantly with fewer preventable recurrent visits. To address this, we developed a novel, interpretable framework that both identifies recurrent patients with high utilization and determines which comorbidities contribute most to their recurrent visits. Specifically, we present a novel algorithm, called the minimum similarity association rules (MSAR), which balances the confidence-support trade-off, to determine the conditions most associated with re-occurring Emergency department and inpatient visits. We validate MSAR on a large Electronic Health Record dataset, demonstrating the effectiveness and consistency in ability to find low-support comorbidities with high likelihood of being associated with recurrent visits, which is challenging for other algorithms such as XGBoost. Clinical relevance- In the era of value-based care and population health management, the proposal could be used for decision making to help reduce future recurrent admissions, improve patient outcomes and reduce the cost of healthcare.

Can diverse population characteristics be leveraged in a machine learning pipeline to predict resource intensive healthcare utilization among hospital service areas?

Background

Super-utilizers represent approximately 5% of the population in the United States (U.S.) and yet they are responsible for over 50% of healthcare expenditures. Using characteristics of hospital service areas (HSAs) to predict utilization of resource intensive healthcare (RIHC) may offer a novel and actionable tool for identifying super-utilizer segments in the population. Consumer expenditures may offer additional value in predicting RIHC beyond typical population characteristics alone.

Methods

Cross-sectional data from 2017 was extracted from 5 unique sources. The outcome was RIHC and included emergency room (ER) visits, inpatient days, and hospital expenditures, all expressed as log per capita. Candidate predictors from 4 broad groups were used, including demographics, adults and child health characteristics, community characteristics, and consumer expenditures. Candidate predictors were expressed as per capita or per capita percent and were aggregated from zip-codes to HSAs using weighed means. Machine learning approaches (Random Forrest, LASSO) selected important features from nearly 1,000 available candidate predictors and used them to generate 4 distinct models, including non-regularized and LASSO regression, random forest, and gradient boosting. Candidate predictors from the best performing models, for each outcome, were used as independent variables in multiple linear regression models. Relative contribution of variables from each candidate predictor group to regression model fit were calculated.

Results

The median ER visits per capita was 0.482 [IQR:0.351–0.646], the median inpatient days per capita was 0.395 [IQR:0.214–0.806], and the median hospital expenditures per capita was $2,302 [1$,544.70-$3,469.80]. Using 1,106 variables, the test-set coefficient of determination (R²) from the best performing models ranged between 0.184–0.782. The adjusted R² values from multiple linear regression models ranged from 0.311–0.8293. Relative contribution of consumer expenditures to model fit ranged from 23.4–33.6%.

Discussion

Machine learning models predicted RIHC among HSAs using diverse population data, including novel consumer expenditures and provides an innovative tool to predict population-based healthcare utilization and expenditures. Geographic variation in utilization and spending were identified.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-022-08154-4.

Identifying patterns of clinical conditions among high-cost older adult health care users using claims data: a latent class approach

Objectives

To identify patterns of clinical conditions among high-cost older adults health care users and explore the associations between characteristics of high-cost older adults and patterns of clinical conditions.

Methods

We analyzed data from the Shanghai Basic Social Medical Insurance Database, China. A total of 2927 older adults aged 60 years and over were included as the analysis sample. We used latent class analysis to identify patterns of clinical conditions among high-cost older adults health care users. Multinomial logistic regression models were also used to determine the associations between demographic characteristics, insurance types, and patterns of clinical conditions.

Results

Five clinically distinctive subgroups of high-cost older adults emerged. Classes included “cerebrovascular diseases” (10.6% of high-cost older adults), “malignant tumor” (9.1%), “arthrosis” (8.8%), “ischemic heart disease” (7.4%), and “other sporadic diseases” (64.1%). Age, sex, and type of medical insurance were predictors of high-cost older adult subgroups.

Conclusions

Profiling patterns of clinical conditions among high-cost older adults is potentially useful as a first step to inform the development of tailored management and intervention strategies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12939-022-01688-3.

The Impact of Multidisciplinary Conferences on Healthcare Utilization in Chronic Pain Patients

Approximately 100 million adults in the United States have chronic pain, though only a subset utilizes the vast majority of healthcare resources. Multidisciplinary care has been shown to improve outcomes in a variety of clinical conditions. There is concern that multidisciplinary care of chronic pain patients may overwhelm existing resources and increase healthcare utilization due to the volume of patients and the complexity of care. We report our findings on the use of multidisciplinary conferences (MDC) to facilitate care for the most complex patients seen at our tertiary center. Thirty-two of nearly 2,000 patients seen per year were discussed at the MDC, making up the top 2% of complex patients in our practice. We evaluated patients' numeric rating score (NRS) of pain, medication use, hospitalizations, emergency department visits, and visits to pain specialists prior to their enrollment in MDC and 1 year later. Matched samples were compared using Wilcoxon's signed rank test. Patients' NRS scores significantly decreased from 7.64 to 5.54 after inclusion in MDC (p < 0.001). A significant decrease in clinic visits (p < 0.001) and healthcare utilization (p < 0.05) was also observed. Opioid and non-opioid prescriptions did not change significantly (p = 0.43). 83% of providers agreed that MDC improved patient care. While previous studies have shown the effect of multi-disciplinary care, we show notable improvements with a team established around a once-a-month MDC.

Improving the Prediction of Persistent High Healthcare Utilizers: Using an Ensemble Methodology

Background

A small proportion of high-need patients persistently use the bulk of health care services and incur disproportionate costs. Population health management (PHM) programs often refer to these patients as persistent high utilizers (PHUs). Accurate PHU prediction enables PHM programs to better align scarce health care resources with high-need PHUs while generally improving outcomes. While prior research in PHU prediction has shown promise, traditional regression methods used in these studies have yielded limited accuracy.

Objective

We are seeking to improve PHU predictions with an ensemble approach in a retrospective observational study design using insurance claim records.

Methods

We defined a PHU as a patient with health care costs in the top 20% of all patients for 4 consecutive 6-month periods. We used 2013 claims data to predict PHU status in next 24 months. Our study population included 165,595 patients in the Johns Hopkins Health Care plan, with 8359 (5.1%) patients identified as PHUs in 2014 and 2015. We assessed the performance of several standalone machine learning methods and then an ensemble approach combining multiple models.

Results

The candidate ensemble with complement naïve Bayes and random forest layers produced increased sensitivity and positive predictive value (PPV; 49.0% and 50.3%, respectively) compared to logistic regression (46.8% and 46.1%, respectively).

Conclusions

Our results suggest that ensemble machine learning can improve prediction of care management needs. Improved PPV implies reduced incorrect referral of low-risk patients. With the improved sensitivity/PPV balance of this approach, resources may be directed more efficiently to patients needing them most.

Assessing the Value of Unsupervised Clustering in Predicting Persistent High Health Care Utilizers: Retrospective Analysis of Insurance Claims Data

BACKGROUND

A high proportion of health care services are persistently utilized by a small subpopulation of patients. To improve clinical outcomes while reducing costs and utilization, population health management programs often provide targeted interventions to patients who may become persistent high users/utilizers (PHUs). Enhanced prediction and management of PHUs can improve health care system efficiencies and improve the overall quality of patient care.

OBJECTIVE

The aim of this study was to detect key classes of diseases and medications among the study population and to assess the predictive value of these classes in identifying PHUs.

METHODS

This study was a retrospective analysis of insurance claims data of patients from the Johns Hopkins Health Care system. We defined a PHU as a patient incurring health care costs in the top 20% of all patients' costs for 4 consecutive 6-month periods. We used 2013 claims data to predict PHU status in 2014-2015. We applied latent class analysis (LCA), an unsupervised clustering approach, to identify patient subgroups with similar diagnostic and medication patterns to differentiate variations in health care utilization across PHUs. Logistic regression models were then built to predict PHUs in the full population and in select subpopulations. Predictors included LCA membership probabilities, demographic covariates, and health utilization covariates. Predictive powers of the regression models were assessed and compared using standard metrics.

RESULTS

We identified 164,221 patients with continuous enrollment between 2013 and 2015. The mean study population age was 19.7 years, 55.9% were women, 3.3% had ≥1 hospitalization, and 19.1% had 10+ outpatient visits in 2013. A total of 8359 (5.09%) patients were identified as PHUs in both 2014 and 2015. The LCA performed optimally when assigning patients to four probability disease/medication classes. Given the feedback provided by clinical experts, we further divided the population into four diagnostic groups for sensitivity analysis: acute upper respiratory infection (URI) (n=53,232; 4.6% PHUs), mental health (n=34,456; 12.8% PHUs), otitis media (n=24,992; 4.5% PHUs), and musculoskeletal (n=24,799; 15.5% PHUs). For the regression models predicting PHUs in the full population, the F1-score classification metric was lower using a parsimonious model that included LCA categories (F1=38.62%) compared to that of a complex risk stratification model with a full set of predictors (F1=48.20%). However, the LCA-enabled simple models were comparable to the complex model when predicting PHUs in the mental health and musculoskeletal subpopulations (F1-scores of 48.69% and 48.15%, respectively). F1-scores were lower than that of the complex model when the LCA-enabled models were limited to the otitis media and acute URI subpopulations (45.77% and 43.05%, respectively).

CONCLUSIONS

Our study illustrates the value of LCA in identifying subgroups of patients with similar patterns of diagnoses and medications. Our results show that LCA-derived classes can simplify predictive models of PHUs without compromising predictive accuracy. Future studies should investigate the value of LCA-derived classes for predicting PHUs in other health care settings.

Super fragmented: a nationally representative cross-sectional study exploring the fragmentation of inpatient care among super-utilizers

Background

Super-utilizers with 4 or more admissions per year frequently receive low-quality care and disproportionately contribute to healthcare costs. Inpatient care fragmentation (admission to multiple different hospitals) in this population has not been well described.

Objective

To determine the prevalence of super-utilizers who receive fragmented care across different hospitals and to describe associated risks, costs, and health outcomes.

Research design

We analyzed inpatient data from the Health Care Utilization Project’s State Inpatient and Emergency Department database from 6 states from 2013. After identifying hospital super-utilizers, we stratified by the number of different hospitals visited in a 1-year period. We determined how patient demographics, costs, and outcomes varied by degree of fragmentation. We then examined how fragmentation would influence a hospital’s ability to identify super-utilizers.

Subjects

Adult patients with 4 or more inpatient stays in 1 year.

Measures

Patient demographics, cost, 1-year hospital reported mortality, and probability that a single hospital could correctly identify a patient as a super-utilizer.

Results

Of the 167,515 hospital super-utilizers, 97,404 (58.1%) visited more than 1 hospital in a 1-year period. Fragmentation was more likely among younger, non-white, low-income, under-insured patients, in population-dense areas. Patients with fragmentation were more likely to be admitted for chronic disease management, psychiatric illness, and substance abuse. Inpatient fragmentation was associated with higher yearly costs and lower likelihood of being identified as a super-utilizer.

Conclusions

Inpatient care fragmentation is common among super-utilizers, disproportionately affects vulnerable populations. It is associated with high yearly costs and a decreased probability of correctly identifying super-utilizers.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-021-06323-5.

Social media language of healthcare super-utilizers

An understanding of healthcare super-utilizers' online behaviors could better identify experiences to inform interventions. In this retrospective case-control study, we analyzed patients' social media posts to better understand their day-to-day behaviors and emotions expressed online. Patients included those receiving care in an urban academic emergency department who consented to share access to their historical Facebook posts and electronic health records. Super-utilizers were defined as patients with more than six visits to the Emergency Department (ED) in a year. We compared posts by super-utilizers with a matched group using propensity scoring based on age, gender and Charlson comorbidity index. Super-utilizers were more likely to post about confusion and negativity (D = .65, 95% CI-[.38, .95]), self-reflection (D = .63 [.35, .91]), avoidance (D = .62 [.34, .90]), swearing (D = .52 [.24, .79]), sleep (D = .60 [.32, .88]), seeking help and attention (D = .61 [.33, .89]), psychosomatic symptoms, (D = .49 [.22, .77]), self-agency (D = .56 [.29, .85]), anger (D = .51, [.24, .79]), stress (D = .46, [.19, .73]), and lonely expressions (D = .44, [.17, .71]). Insights from this study can potentially supplement offline community care services with online social support interventions considering the high engagement of super-utilizers on social media.

[Predictive validity of a functional classification method in older adultsValidação preditiva de método de classificação funcional em idosos]

Objetivo.

Evaluar la validez predictiva de un método de clasificación funcional (CF) sobre el uso de los servicios de urgencias y hospitalización, mortalidad y costos de la atención en salud en adultos mayores.

Métodos.

Estudio de cohorte retrospectivo que incluyó 2 168 adultos mayores en un programa de atención de las enfermedades crónicas no transmisibles (ECNT) en Medellín (Colombia). Los pacientes fueron estratificados según un método de CF con base en el estado funcional, presencia de factores de riesgo y control de la comorbilidad. Durante un año de seguimiento, se evaluó la validez predictiva de la CF sobre los desenlaces estudiados; se midieron la discriminación y la calibración con el estadístico-C y de Hosmer-Lemeshow (H-L), respectivamente.

Resultados.

El promedio de edad fue 74,6 ± 7,9 años; el 40,8% (n = 884) fueron hombres y 7,7% (n = 168) murieron. El riesgo de muerte (razón de posibilidades [OR, por su sigla en inglés]: 1,767; 3,411; 8,525), hospitalización (OR: 1,397; 2,172; 3,540) y un costo elevado de la atención en salud (OR: 1,703; 2,369; 5,073) aumentaron en la medida que hubo un deterioro en la CF, clases 2B, 3 y 4, respectivamente. El modelo predictivo para el desenlace muerte mostró una buena capacidad de discriminación (estadístico-C = 0,721) y calibración (estadístico de H-L = 10,200; P = 0,251).

Conclusión.

Existe una relación de dosis y respuesta entre el deterioro de la CF y un riesgo más elevado de muerte, hospitalización y costo elevado. La CF tiene validez predictiva para la tasa de mortalidad y podría utilizarse para la estratificación de adultos mayores en programas de atención de las ECNT con miras a dirigir las acciones de intervención.

Boston Violence Intervention Advocacy Program: Challenges and Opportunities for Client Engagement and Goal Achievement

OBJECTIVES

A better understanding of the factors affecting client engagement in hospital-based violence intervention programs (HVIPs), and which types of client needs prove most challenging to achieve, may be of key importance in developing novel, targeted strategies to violence intervention. In this study, we examined the demographics and injury characteristics of violently injured patients by their level of engagement with the Boston Violence Intervention Advocacy Program (VIAP) and determined the degree of client goal achievement through VIAP client services.

METHODS

This retrospective study was performed using a cohort of patients presenting to the Boston Medical Center emergency department for a violent penetrating injury due to community violence between 2013 and 2018. Data on client demographics, injury characteristics, and client needs were collected from the VIAP data repository. Cox proportional hazard regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals to assess the difference in hazards of client goal achievement by need type.

RESULTS

Of the 2,243 victims of violent injury, 839 (37.4%) patients engaged with VIAP. Significant predictors of client engagement include younger age, Black race, permanent home, existing mental health diagnosis, gunshot wound, and more severe injuries. Conversely, older age, homelessness, substance use, stab wound, and less severe injuries predicted refusal of VIAP services. For clients who chose to engage with VIAP, needs related to education (HR = 0.47, 95% CI = 0.38 to 0.58), employment (HR = 0.66, 95% CI = 0.57 to 0.77), and housing (HR = 0.76, 95% CI = 0.68 to 0.86) were significantly less likely to be achieved compared to basic needs.

CONCLUSIONS

This study demonstrates that VIAP is effectively engaging the client population that HVIPs have been designed to support. HVIPs should consider novel strategies to engage vulnerable populations not typically targeted by intervention programs. These results speak to the difficulties of program attrition and the complexities of altering the life course for victims of violence.

Characterizing health care utilization following hospitalization for a traumatic brain injury: a retrospective cohort study

Objective: The purpose of this study was to characterize health services utilization among individuals hospitalized with a traumatic brain injury (TBI) 1-year post-injury.Methods: Using a retrospective cohort design, adult patients (n = 32, 042) hospitalized with a traumatic brain injury between 2005 and 2014 were selected from a statewide traumatic brain injury registry. Data on health services utilization for 1-year post-injury were extracted from electronic medical and administrative records. Descriptive statistics and logistic regression were used to characterize the cohort and a subgroup of superutilizers of health services.Results: One year after traumatic brain injury, 56% of participants used emergency department services, 80% received inpatient services, and 93% utilized outpatient health services. Superutilizers had ≥3 emergency department visits, ≥3 inpatient admissions, or ≥26 outpatient visits 1-year post-injury. Twenty-six percent of participants were superutilizers of emergency department services, 30% of inpatient services, and 26% of outpatient services. Superutilizers contributed to 81% of emergency department visits, 70% of inpatient visits, and 60% of outpatient visits. Factors associated with being a superutilizer included sex, race, residence, and insurance type.Conclusions: Several patient characteristics and demographic factors influenced patients' healthcare utilization post-TBI. Findings provide opportunities for developing targeted interventions to improve patients' health and traumatic brain injury-related healthcare delivery.

The Association between Emergency Department Super-Utilizer Status and Willingness to Participate in Research

BACKGROUND

Research based on emergency departments (EDs) primarily focuses on medical conditions. There is limited research that investigates patients who willingly participate in research. This current study explored ED super-utilizers' (SUs') and nonsuper-utilizers' (NSUs') attitudes toward research.

OBJECTIVE

The study assesses the willingness of SUs to participate in research. We hypothesize that the SU population will be as interested as nonutilizers in participating in medical research.

METHODS

This prospective observational study stratified participants into SU and NSU cohorts based on their self-reported number of ED visits within 6 months. Surveys were captured in a secured database and analyzed using SAS 9.4.

RESULTS

7,481 completed questionnaires. SUs were more interested in participating in all types of research compared to NSUs. Both groups were most willing to participate in surveys. Neither group was particularly interested in studies that required medications. SUs were not more willing to participate in studies without payment than NSUs. Both groups trusted researchers at the same rates.

CONCLUSION

Although rarely included in medical research, SUs were more willing to participate in nearly all types of research and expressed a similar trust in medical research when compared to nonsuper-utilizers.

Right-Site Care Programme with a community-based family medicine clinic in Singapore: secondary data analysis of its impact on mortality and healthcare utilisation

OBJECTIVE

Stable patients with chronic conditions could be appropriately cared for at family medicine clinics (FMC) and discharged from hospital specialist outpatient clinics (SOCs). The Right-Site Care Programme with Frontier FMC emphasised care organised around patients in community rather than hospital-based providers, with one identifiable primary provider. This study evaluated impact of this programme on mortality and healthcare utilisation.

DESIGN

A retrospective study without randomisation using secondary data analysis of patients enrolled in the intervention matched 1:1 with unenrolled patients as controls.

SETTING

Programme was supported by the Ministry of Health in Singapore, a city-state nation in Southeast Asia with 5.6 million population.

PARTICIPANTS

Intervention group comprises patients enrolled from January to December 2014 (n=684) and control patients (n=684) with at least one SOC and no FMC attendance during same period.

INTERVENTIONS

Family physician in Frontier FMC managed patients in consultation with relevant specialist physicians or fully managed patients independently. Care teams in SOCs and FMC used a common electronic medical records system to facilitate care coordination and conducted regular multidisciplinary case conferences.

PRIMARY OUTCOME MEASURES

Deidentified linked healthcare administrative data for time period of January 2011 to December 2017 were extracted. Three-year postenrolment mortality rates and utilisation frequencies and charges for SOC, public primary care centres (polyclinic), emergency department attendances and emergency, non-day surgery inpatient and all-cause admissions were compared.

RESULTS

Intervention patients had lower mortality rate (HR=0.37, p<0.01). Among those with potential of postenrolment polyclinic attendance, intervention patients had lower frequencies (incidence rate ratio (IRR)=0.60, p<0.01) and charges (mean ratio (MR)=0.51, p<0.01). Among those with potential of postenrolment SOC attendance, intervention patients had higher frequencies (IRR=2.06, p<0.01) and charges (MR=1.86, p<0.01).

CONCLUSIONS

Intervention patients had better survival, probably because their chronic conditions were better managed with close monitoring, contributing to higher total outpatient attendance frequencies and charges.