1
|
Oldfield BJ, Pasha S, Mun S, Sedghi T, Zhu W, DeCew A, Flaherty-Hewitt M, Olson DP. Construction of a Pediatrics Risk Score to Predict High Health Care Costs Among a Community Health Center Cohort. Popul Health Manag 2020; 24:345-352. [PMID: 32639198 DOI: 10.1089/pop.2020.0035] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/13/2022] Open
Abstract
Risk-stratification strategies are needed for ambulatory pediatric populations. The authors sought to develop age-specific risk scores that predict high health care costs among an urban population. A retrospective cohort study was performed of children ages 1-18 years who received care at Fair Haven Community Health Care (FHCHC), a community health center in New Haven, Connecticut. Cost was estimated from charges in the electronic health record (EHR), which is shared with the only hospital system in the city. Using multivariable logistic regression models, independent predictors of being in the top decile of total charges during the 2017 calendar year were identified, drawing from covariates collected from the EHR prior to 2017. Random forest modeling was used to verify the feature importance of significant covariates and model performance from 2017 cost data were compared to those using 2018 cost data. Regression models were used to construct age-specific nomograms to predict cost. Among 8960 children who received care at FHCHC in the 18 months prior to 2017, covariate frequencies clustered in age groups 1-5 years, 6-11 years, and 12-18 years, so 3 age-specific models were constructed. Prior utilization variables predicted future costs, as did younger children who received specialty care and older children with behavioral health diagnoses. Final models for each age group had C statistics ≥0.68 using both 2017 and 2018 cost data. Prediction models can draw from elements accessible in the EHR to predict cost of ambulatory pediatric patients. Strategies to impact utilization among high-risk children are needed.
Collapse
Affiliation(s)
- Benjamin J Oldfield
- Fair Haven Community Health Care, New Haven, Connecticut, USA.,Department of Pediatrics, Yale School of Medicine, New Haven, Connecticut, USA.,Department of Medicine, Yale School of Medicine, New Haven, Connecticut, USA
| | - Saamir Pasha
- Department of Medicine, Yale School of Medicine, New Haven, Connecticut, USA
| | - Sophia Mun
- Department of Medicine, Yale School of Medicine, New Haven, Connecticut, USA
| | - Tannaz Sedghi
- University of Connecticut School of Medicine, Farmington, Connecticut, USA
| | - Weiwei Zhu
- Amazon.com, Inc., Seattle, Washington, USA
| | - Amanda DeCew
- Fair Haven Community Health Care, New Haven, Connecticut, USA
| | | | - Douglas P Olson
- Fair Haven Community Health Care, New Haven, Connecticut, USA
| |
Collapse
|
2
|
Lentz TA, Harman JS, Marlow NM, Beneciuk JM, Fillingim RB, George SZ. Factors associated with persistently high-cost health care utilization for musculoskeletal pain. PLoS One 2019; 14:e0225125. [PMID: 31710655 PMCID: PMC6844454 DOI: 10.1371/journal.pone.0225125] [Citation(s) in RCA: 20] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/27/2019] [Accepted: 10/29/2019] [Indexed: 02/07/2023] Open
Abstract
BACKGROUND Musculoskeletal pain conditions incur high costs and produce significant personal and public health consequences, including disability and opioid-related mortality. Persistence of high-cost health care utilization for musculoskeletal pain may help identify system inefficiencies that could limit value of care. The objective of this study was to identify factors associated with persistent high-cost utilization among individuals seeking health care for musculoskeletal pain. METHODS This was a retrospective cohort study of Medical Expenditure Panel Survey data (2008-2013) that included a non-institutionalized, population-based sample of individuals seeking health care for a musculoskeletal pain condition (n = 12,985). Expenditures associated with musculoskeletal pain conditions over two consecutive years were analyzed from prescribed medicine, office-based medical provider visits, outpatient department visits, emergency room visits, inpatient hospital stays, and home health visits. Persistent high-cost utilization was defined as being in the top 15th percentile for annual musculoskeletal pain-related expenditures over 2 consecutive years. We used multinomial regression to determine which modifiable and non-modifiable sociodemographic, health, and pain-related variables were associated with persistent high-cost utilization. RESULTS Approximately 35% of direct costs for musculoskeletal pain were concentrated among the 4% defined as persistent high-cost utilizers. Non-modifiable variables associated with expenditure group classification included age, race, poverty level, geographic region, insurance status, diagnosis type and total number of musculoskeletal pain diagnoses. Modifiable variables associated with increased risk of high expenditure classification were higher number of missed work days, greater pain interference, and higher use of prescription medication for pain, while higher self-reported physical and mental health were associated with lower risk of high expenditure classification. CONCLUSIONS Health care delivery models that prospectively identify these potentially modifiable factors may improve the costs and value of care for individuals with musculoskeletal pain prone to risk for high-cost care episodes.
Collapse
Affiliation(s)
- Trevor A. Lentz
- Duke Clinical Research Institute and Department of Orthopaedic Surgery, Duke University, Durham, North Carolina, United States of America
- * E-mail:
| | - Jeffrey S. Harman
- Department of Behavioral Sciences and Social Medicine, Florida State University, Tallahassee, Florida, United States of America
| | - Nicole M. Marlow
- Department of Health Services Research, Management, and Policy, University of Florida, Gainesville, Florida, United States of America
| | - Jason M. Beneciuk
- Brooks Rehabilitation – College of Public Health & Health Professions Research Collaboration, Department of Physical Therapy, University of Florida, Gainesville, Florida, United States of America
| | - Roger B. Fillingim
- Pain Research & Intervention Center of Excellence, University of Florida, Gainesville, Florida, United States of America
| | - Steven Z. George
- Duke Clinical Research Institute and Department of Orthopaedic Surgery, Duke University, Durham, North Carolina, United States of America
| |
Collapse
|
3
|
Smooth Bayesian network model for the prediction of future high-cost patients with COPD. Int J Med Inform 2019; 126:147-155. [PMID: 31029256 DOI: 10.1016/j.ijmedinf.2019.03.017] [Citation(s) in RCA: 8] [Impact Index Per Article: 1.6] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/05/2018] [Revised: 02/28/2019] [Accepted: 03/26/2019] [Indexed: 02/05/2023]
Abstract
INTRODUCTION The clinical course of chronic obstructive pulmonary disease (COPD) is marked by acute exacerbation events that increase hospitalization rates and healthcare spending. The early identification of future high-cost patients with COPD may decrease healthcare spending by informing individualized interventions that prevent exacerbation events and decelerate disease progression. Existing studies of cost prediction of other chronic diseases have applied regression and machine-learning methods that cannot capture the complex causal relationships between COPD factors. Thus, the exploration of these factors through nonlinear, high-dimensional but explainable modeling is greatly needed. OBJECTIVES We aimed to develop a machine-learning model to identify future high-cost patients with COPD. Such a model should incorporate expert knowledge about causal relationships, and the method for estimating the model could provide more accurate predictions than other machine learning methods. METHODS We used the 2011-2013 medical insurance data of patients with COPD in a large city. The data set included demographic information and admission records. Leveraging on developments in graphical modeling methods, we proposed a smooth Bayesian network (SBN) model for the prediction of high-cost individuals using medical insurance data. The modeling method incorporated some expert knowledge about causal relationships (i.e., about the Bayesian network structure). We employed a smoothing kernel based on the weighted nearest neighborhood method in the SBN model to address overfitting, case-mix effect, and data sparsity (i.e., using data about "similar patients"). RESULTS The proposed SBN achieved the area under curve (AUC) of 0.80 and showed considerable improvement over the baseline machine-learning methods. Besides confirming the known factors from the literature, we found "region" (i.e., a suburban or urban area) to be a significant factor, and that in a 3-tier system with primary, secondary and tertiary hospitals, COPD patients who had been admitted to primary hospitals were more likely to develop into future high-cost patients than patients who had been admitted to tertiary hospitals. CONCLUSION The proposed SBN model not only obtained higher prediction accuracy and stronger generalizability than a number of benchmark machine-learning methods, but also used the Bayesian network to capture the complex causal relationships between different predictors by incorporating expert knowledge. Furthermore, a framework was developed to establish the relationships between exposure to historical trajectory and future outcome, which can also be applied to other temporal data to model different trajectory information and predict other outcomes.
Collapse
|
4
|
Matiz LA, Robbins-Milne L, Rausch JA. EMR Adaptations to Support the Identification and Risk Stratification of Children with Special Health Care Needs in the Medical Home. Matern Child Health J 2019; 23:919-924. [PMID: 30617441 DOI: 10.1007/s10995-018-02718-9] [Citation(s) in RCA: 5] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/29/2022]
Abstract
Introduction Children with special health care needs (CSHCN) are a high risk population with complex medical issues and needs. It is challenging to care for them in a busy, pediatric practice without understanding how many exist and how best to allocate resources. EMRs can be adapted to develop registries and stratify patients to promote population health management. Methods Adaptations were made to the EMR in September 2013 to capture CSHCN and the associated risk level during well-child visits prospectively. All physicians were trained on the definition of CSHCN and on risk stratification levels 1, 2, 3A and 3B. An analysis using one-way ANOVA for children ages 0-21, seen between September 1, 2011 and August 31, 2015, who were identified and stratified after September 2013, was conducted to determine utilization patterns on hospital admissions, emergency department (ED), subspecialty, and primary care visits. Results A total of 4687 CSHCN were identified during the study period. Of the CSHCN, 45% were Level 1, 41% Level 2, 7% 3A and 7% 3B. There were significant differences in utilization across the tiers of CSHCN with the highest level of stratification (3B) demonstrating the most hospital admissions and primary care visits. Level 3B and level 3A (unstable) had significantly more ED visits. Additionally, as tiers increased from level 1 to 3B there was an increase in subspecialty provider utilization (p < 0.0001). Discussion The EMR adaptations developed for CSHCN identified the expected number of CSHCN and predicted utilization patterns across primary, subspecialty, ED and in-patient care.
Collapse
Affiliation(s)
- L Adriana Matiz
- Department of Pediatrics, Columbia University Medical Center, 622 West 168th Street, VC 417, New York, NY, 10032, USA. .,NewYork Presbyterian Hospital-Ambulatory Care Network, 622 West 168th Street, VC-417, New York, NY, USA.
| | - Laura Robbins-Milne
- Department of Pediatrics, Columbia University Medical Center, 622 West 168th Street, VC 417, New York, NY, 10032, USA.,NewYork Presbyterian Hospital-Ambulatory Care Network, 622 West 168th Street, VC-417, New York, NY, USA
| | - John A Rausch
- Department of Pediatrics, Columbia University Medical Center, 622 West 168th Street, VC 417, New York, NY, 10032, USA.,NewYork Presbyterian Hospital-Ambulatory Care Network, 622 West 168th Street, VC-417, New York, NY, USA
| |
Collapse
|
5
|
Mental Health Conditions and Health Care Payments for Children with Chronic Medical Conditions. Acad Pediatr 2019; 19:44-50. [PMID: 30315948 DOI: 10.1016/j.acap.2018.10.001] [Citation(s) in RCA: 21] [Impact Index Per Article: 4.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 06/06/2018] [Revised: 10/04/2018] [Accepted: 10/10/2018] [Indexed: 12/18/2022]
Abstract
OBJECTIVE To estimate additional payments associated with co-existing mental health or substance use disorders (MH/SUDs) among commercially insured children and youth with chronic medical conditions (CMCs) and to determine whether children's MH/SUDs have similar associations with parental health care payments. METHODS Cross-sectional analysis of a national database of paid commercial insurance claims for 2012-2013. Participants were children and youth ages 0 to 26 years covered as dependents on parents' health insurance and categorized by the presence or absence of any of 11 chronic medical conditions and MH/SUDs. We determined the numbers of children and youth with CMCs and paid health care claims categorized as hospital, professional, and pharmacy services and as medical or behavioral. We compared paid claims for children and youth with CMCs with and without co-occurring MH/SUDs and for their parents. RESULTS The sample included almost 6.6 million children and youth and 5.8 million parents. Compared to children without CMCs, children with CMCs had higher costs, even higher for children with CMCs who also had MH/SUDs. Children with CMCs and co-occurring MH/SUDs had 2.4 times the annual payments of those with chronic conditions alone, especially for medical expenses. Estimated additional annual payments associated with MH/SUDs in children with CMCs were $8.8 billion. Parents of children with CMCs and associated MH/SUDs had payments 59% higher than those for parents of children with CMCs alone. CONCLUSIONS MH/SUDs in children and youth with CMCs are associated with higher total health care payments for both patients and their parents, suggesting potential benefits from preventing or reducing the impact of MH/SUDs among children and youth with CMCs.
Collapse
|
6
|
Sterling S, Chi F, Weisner C, Grant R, Pruzansky A, Bui S, Madvig P, Pearl R. Association of behavioral health factors and social determinants of health with high and persistently high healthcare costs. Prev Med Rep 2018; 11:154-159. [PMID: 30003015 PMCID: PMC6039851 DOI: 10.1016/j.pmedr.2018.06.017] [Citation(s) in RCA: 19] [Impact Index Per Article: 3.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/09/2018] [Revised: 05/23/2018] [Accepted: 06/25/2018] [Indexed: 01/27/2023] Open
Abstract
A high proportion of U.S. health care costs are attributable to a relatively small proportion of patients. Understanding behavioral and social factors that predict initial and persistent high costs for these "high utilizers" is critical for health policy-makers. This prospective observational study was conducted at Kaiser Permanente Northern California (KPNC), an integrated healthcare delivery system with 4.1 million members. A stratified random sample of high-cost vs. non-high-cost adult KPNC members matched by age, gender, race/ethnicity, type of health insurance, and medical severity (N = 378) was interviewed between 3/14/2013 and 3/20/2014. Data on health care costs and clinical diagnoses between 1/1/2008 and 12/31/2012 were derived from the electronic health record (EHR). Social-economic status, depression symptoms, adverse childhood experiences (ACEs), interpersonal violence, financial stressors, neighborhood environment, transportation access, and patient activation and engagement were obtained through telephone interviews. Initial and subsequent high-cost status were defined as being classified in top 20% cost levels over 1/1/2009-12/31/2011 and 1/1/2012-12/31/2012, respectively. Psychiatric diagnosis (OR 2.55, 95% CI 1.52-4.29, p < 0.001), financial stressors (OR 1.97, 95% CI 1.19-3.26, p = 0.009), and ACEs (OR 1.10, 95% CI 1.00-1.20, p = 0.051) predicted initial high-cost status. ACEs alone predicted persistent high-cost status in the subsequent year (OR 1.12, 95% CI 1.00-1.25, p = 0.050). Non-medical factors such as psychiatric problems, financial stressors and adverse childhood experiences contribute significantly to the likelihood of high medical utilization and cost. Efforts to predict and reduce high utilization must include measuring and potentially addressing these factors.
Collapse
Affiliation(s)
- Stacy Sterling
- Division of Research, Kaiser Permanente Northern California, United States
| | - Felicia Chi
- Division of Research, Kaiser Permanente Northern California, United States
| | - Constance Weisner
- Division of Research, Kaiser Permanente Northern California, United States
- University of California, San Francisco, United States
| | - Richard Grant
- Division of Research, Kaiser Permanente Northern California, United States
| | | | - Sandy Bui
- The Permanente Medical Group, United States
| | | | | |
Collapse
|
7
|
Louis DZ, Callahan CA, Robeson M, Liu M, McRae J, Gonnella JS, Lombardi M, Maio V. Predicting risk of hospitalisation: a retrospective population-based analysis in a paediatric population in Emilia-Romagna, Italy. BMJ Open 2018; 8:e019454. [PMID: 29730620 PMCID: PMC5942467 DOI: 10.1136/bmjopen-2017-019454] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 01/16/2023] Open
Abstract
OBJECTIVES Develop predictive models for a paediatric population that provide information for paediatricians and health authorities to identify children at risk of hospitalisation for conditions that may be impacted through improved patient care. DESIGN Retrospective healthcare utilisation analysis with multivariable logistic regression models. DATA Demographic information linked with utilisation of health services in the years 2006-2014 was used to predict risk of hospitalisation or death in 2015 using a longitudinal administrative database of 527 458 children aged 1-13 years residing in the Regione Emilia-Romagna (RER), Italy, in 2014. OUTCOME MEASURES Models designed to predict risk of hospitalisation or death in 2015 for problems that are potentially avoidable were developed and evaluated using the C-statistic, for calibration to assess performance across levels of predicted risk, and in terms of their sensitivity, specificity and positive predictive value. RESULTS Of the 527 458 children residing in RER in 2014, 6391 children (1.21%) were hospitalised for selected conditions or died in 2015. 49 486 children (9.4%) of the population were classified in the 'At Higher Risk' group using a threshold of predicted risk >2.5%. The observed risk of hospitalisation (5%) for the 'At Higher Risk' group was more than four times higher than the overall population. We observed a C-statistic of 0.78 indicating good model performance. The model was well calibrated across categories of predicted risk. CONCLUSIONS It is feasible to develop a population-based model using a longitudinal administrative database that identifies the risk of hospitalisation for a paediatric population. The results of this model, along with profiles of children identified as high risk, are being provided to the paediatricians and other healthcare professionals providing care to this population to aid in planning for care management and interventions that may reduce their patients' likelihood of a preventable, high-cost hospitalisation.
Collapse
Affiliation(s)
- Daniel Z Louis
- Center for Medical Research in Medical Education and Health Care, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, Pennsylvania, USA
| | - Clara A Callahan
- Center for Medical Research in Medical Education and Health Care, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, Pennsylvania, USA
| | - Mary Robeson
- Center for Medical Research in Medical Education and Health Care, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, Pennsylvania, USA
| | - Mengdan Liu
- Center for Medical Research in Medical Education and Health Care, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, Pennsylvania, USA
| | - Jacquelyn McRae
- Jefferson College of Population Health, Thomas Jefferson University, Philadelphia, Pennsylvania, USA
| | - Joseph S Gonnella
- Center for Medical Research in Medical Education and Health Care, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, Pennsylvania, USA
| | - Marco Lombardi
- Risk Management and Clinical Governance, Parma Local Health Authority, Parma, Italy
| | - Vittorio Maio
- Center for Medical Research in Medical Education and Health Care, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, Pennsylvania, USA
- Jefferson College of Population Health, Thomas Jefferson University, Philadelphia, Pennsylvania, USA
| |
Collapse
|
8
|
Pediatric Specialty Care Model for Management of Chronic Respiratory Failure: Cost and Savings Implications and Misalignment With Payment Models. Pediatr Crit Care Med 2018; 19:412-420. [PMID: 29406371 DOI: 10.1097/pcc.0000000000001472] [Citation(s) in RCA: 19] [Impact Index Per Article: 3.2] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 11/26/2022]
Abstract
OBJECTIVE To describe program design, costs, and savings implications of a critical care-based care coordination model for medically complex children with chronic respiratory failure. DESIGN All program activities and resultant clinical outcomes were tracked over 4 years using an adapted version of the Care Coordination Measurement Tool. Patient characteristics, program activity, and acute care resource utilization were prospectively documented in the adapted version of the Care Coordination Measurement Tool and retrospectively cross-validated with hospital billing data. Impact on total costs of care was then estimated based on program outcomes and nationally representative administrative data. SETTING Tertiary children's hospital. SUBJECTS Critical Care, Anesthesia, Perioperative Extension and Home Ventilation Program enrollees. INTERVENTIONS None. MEASUREMENTS AND MAIN RESULTS The program provided care for 346 patients and families over the study period. Median age at enrollment was 6 years with more than half deriving secondary respiratory failure from a primary neuromuscular disease. There were 11,960 encounters over the study period, including 1,202 home visits, 673 clinic visits, and 4,970 telephone or telemedicine encounters. Half (n = 5,853) of all encounters involved a physician and 45% included at least one care coordination activity. Overall, we estimated that program interventions were responsible for averting 556 emergency department visits and 107 hospitalizations. Conservative monetization of these alone accounted for annual savings of $1.2-2 million or $407/pt/mo net of program costs. CONCLUSIONS Innovative models, such as extension of critical care services, for high-risk, high-cost patients can result in immediate cost savings. Evaluation of financial implications of comprehensive care for high-risk patients is necessary to complement clinical and patient-centered outcomes for alternative care models. When year-to-year cost variability is high and cost persistence is low, these savings can be estimated from documentation within care coordination management tools. Means of financial sustainability, scalability, and equal access of such care models need to be established.
Collapse
|
9
|
Phillips CD. The Pediatric Home Care/Expenditure Classification Model (P/ECM): A Home Care Case-Mix Model for Children Facing Special Health Care Challenges. Health Serv Insights 2016; 8:35-43. [PMID: 26740744 PMCID: PMC4694607 DOI: 10.4137/hsi.s35366] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/22/2015] [Revised: 11/17/2015] [Accepted: 11/19/2015] [Indexed: 11/21/2022] Open
Abstract
Case-mix classification and payment systems help assure that persons with similar needs receive similar amounts of care resources, which is a major equity concern for consumers, providers, and programs. Although health service programs for adults regularly use case-mix payment systems, programs providing health services to children and youth rarely use such models. This research utilized Medicaid home care expenditures and assessment data on 2,578 children receiving home care in one large state in the USA. Using classification and regression tree analyses, a case-mix model for long-term pediatric home care was developed. The Pediatric Home Care/Expenditure Classification Model (P/ECM) grouped children and youth in the study sample into 24 groups, explaining 41% of the variance in annual home care expenditures. The P/ECM creates the possibility of a more equitable, and potentially more effective, allocation of home care resources among children and youth facing serious health care challenges.
Collapse
Affiliation(s)
- Charles D Phillips
- Department of Health Policy and Management, School of Public Health, Health Science Center, Texas A&M University, College Station, TX, USA
| |
Collapse
|