Reducing Health Care Costs Through Patient Targeting: Risk Adjustment Modeling to Predict Patients Remaining High Cost

Validation of the Risk Instrument for Screening in the Community (RISC) among Older Adults in the Emergency Department

Although several short-risk-prediction instruments are used in the emergency department (ED), there remains insufficient evidence to guide healthcare professionals on their use. The Risk Instrument for Screening in the Community (RISC) is an established screen comprising three Likert scales examining the risk of three adverse outcomes among community-dwelling older adults at one-year: institutionalisation, hospitalisation, and death, which are scored from one (rare/minimal) to five (certain/extreme) and combined into an Overall RISC score. In the present study, the RISC was externally validated by comparing it with different frailty screens to predict risk of hospitalisation (30-day readmission), prolonged length of stay (LOS), one-year mortality, and institutionalisation among 193 consecutive patients aged ≥70 attending a large university hospital ED in Western Ireland, assessed for frailty, determined by comprehensive geriatric assessment. The median LOS was 8 ± 9 days; 20% were re-admitted <30 days; 13.5% were institutionalised; 17% had died; and 60% (116/193) were frail. Based on the area under the ROC curve scores (AUC), the Overall RISC score had the greatest diagnostic accuracy for predicting one-year mortality and institutionalisation: AUC 0.77 (95% CI: 0.68-0.87) and 0.73 (95% CI: 0.64-0.82), respectively. None of the instruments were accurate in predicting 30-day readmission (AUC all <0.70). The Overall RISC score had good accuracy for identifying frailty (AUC 0.84). These results indicate that the RISC is an accurate risk-prediction instrument and frailty measure in the ED.

Neutrophil-to-Lymphocyte Ratios and Infections after Traumatic Brain Injury: Associations with Hospital Resource Utilization and Long-Term Outcome

Traumatic brain injury (TBI) induces immune dysfunction that can be captured clinically by an increase in the neutrophil-to-lymphocyte ratio (NLR). However, few studies have characterized the temporal dynamics of NLR post-TBI and its relationship with hospital-acquired infections (HAI), resource utilization, or outcome. We assessed NLR and HAI over the first 21 days post-injury in adults with moderate-to-severe TBI (n = 196) using group-based trajectory (TRAJ), changepoint, and mixed-effects multivariable regression analysis to characterize temporal dynamics. We identified two groups with unique NLR profiles: a high (n = 67) versus a low (n = 129) TRAJ group. High NLR TRAJ had higher rates (76.12% vs. 55.04%, p = 0.004) and earlier time to infection (p = 0.003). In changepoint-derived day 0–5 and 6–20 epochs, low lymphocyte TRAJ, early in recovery, resulted in more frequent HAIs (p = 0.042), subsequently increasing later NLR levels (p ≤ 0.0001). Both high NLR TRAJ and HAIs increased hospital length of stay (LOS) and days on ventilation (p ≤ 0.05 all), while only high NLR TRAJ significantly increased odds of unfavorable six-month outcome as measured by the Glasgow Outcome Scale (GOS) (p = 0.046) in multivariable regression. These findings provide insight into the temporal dynamics and interrelatedness of immune factors which collectively impact susceptibility to infection and greater hospital resource utilization, as well as influence recovery.

Utility of catastrophizing, body symptom diagram score and history of opioid use to predict future health care utilization after a primary care visit for musculoskeletal pain

BACKGROUND

Self-report information about pain and pain beliefs are often collected during initial consultation for musculoskeletal pain. These data may provide utility beyond the initial encounter, helping provide further insight into prognosis and long-term interactions of the patient with the health system.

OBJECTIVE

The aim of this study was to determine if pain catastrophizing and pain-related body symptoms can predict future health care utilization.

METHODS

This was a longitudinal cohort study. Baseline data were collected after receiving initial care for a musculoskeletal disorder in a multidisciplinary clinic within a large military hospital. Subjects completed the Pain Catastrophizing Scale, a region-specific disability measure, numeric pain rating scale and a body symptom diagram. Health care utilization data for 1 year prior and after the visit were extracted from the Military Health System Data Repository. Multivariable regression models appropriate for skewed and count data were developed to predict (i) musculoskeletal-specific medical visits, (ii) 12-month opioid use, (iii) musculoskeletal-specific medical costs and (iv) total medical costs. We investigated whether a pain catastrophizing × body symptom diagram interaction improved prediction, and developed separate models for opioid-naïve individuals and those with a history of opioid use in an exploratory analysis.

RESULTS

Pain catastrophizing but not body symptom diagram was a significant predictor of musculoskeletal visits, musculoskeletal costs and total medical costs. Exploratory analyses suggest these relationships are most robust for patients with a history of opioid use.

CONCLUSIONS

Pain catastrophizing can identify risk of high health care utilization and costs, even after controlling for common clinical variables. Addressing pain catastrophizing in the primary care setting may help to mitigate future health care utilization and costs, while improving clinical outcomes. These results provide direction for future validation studies in larger and more traditional primary care settings.

Defining 'actionable' high- costhealth care use: results using the Canadian Institute for Health Information population grouping methodology

Background

A small proportion of the population consumes the majority of health care resources. High-cost health care users are a heterogeneous group. We aim to segment a provincial population into relevant homogenous sub-groups to provide actionable information on risk factors associated with high-cost health care use within sub-populations.

Methods

The Canadian Institute for Health Information (CIHI) Population Grouping methodology was used to define mutually exclusive and clinically relevant health profile sub-groups. High-cost users (> = 90th percentile of health care spending) were defined within each sub-group. Univariate analyses explored demographic, socio-economic status, health status and health care utilization variables associated with high-cost use. Multivariable logistic regression models were constructed for the costliest health profile groups.

Results

From 2015 to 2017, 1,175,147 individuals were identified for study. High-cost users consumed 41% of total health care resources. Average annual health care spending for individuals not high-cost were $642; high-cost users were $16,316. The costliest health profile groups were ‘long-term care’, ‘palliative’, ‘major acute’, ‘major chronic’, ‘major cancer’, ‘major newborn’, ‘major mental health’ and ‘moderate chronic’. Both ‘major acute’ and ‘major cancer’ health profile groups were largely explained by measures of health care utilization and multi-morbidity. In the remaining costliest health profile groups modelled, ‘major chronic’, ‘moderate chronic’, ‘major newborn’ and ‘other mental health’, a measure of socio-economic status, low neighbourhood income, was statistically significantly associated with high-cost use.

Interpretation

Model results point to specific, actionable information within clinically meaningful subgroups to reduce high-cost health care use. Health equity, specifically low socio-economic status, was statistically significantly associated with high-cost use in the majority of health profile sub-groups. Population segmentation methods, and more specifically, the CIHI Population Grouping Methodology, provide specificity to high-cost health care use; informing interventions aimed at reducing health care costs and improving population health.

Forecasting Single Disease Cost of Cataract Based on Multivariable Regression Analysis and Backpropagation Neural Network

In medical services, charge according to the disease is an important way to promote the reform of pricing mechanism, control the unreasonable growth of medical expenses, as well as reduce the burden on patients. Single disease cost forecasting that both identify potential influencing or driving factors and enable better proactive estimation of costs can guide the management and control of medical costs. This study aimed to identify the factors that affect the medical costs of single disease cataract and compare 2 regression models for anticipating acceptable medical cost forecasts. For this purpose, 483 patients with cataract surgery completed in West China Hospital from May 1, 2015, to October 1, 2015, were selected from hospital information system. For cost forecasting, multivariable regression analysis (MRA) and backpropagation neural network (BPNN) were used. Analysis of data was performed with SPSS21.0 and MATLAB2014a software. Total medical costs of patients with cataract (n = 483) ranged from 2015.00 to 13 359.00 CNY, and the mean ± standard deviation is 6292.29 ± 2639.43 CNY. Factors influencing costs of cataract in the MRA include, in importance order, intraocular lens (IOL) implantation (|r|: 0.805, P < .01), doctor level (|r|: 0.644, P < .01), payment source (|r|: 0.554, P < .01), admission status (|r|: 0.326, P < .01), additional diagnosis (|r|: 0.260, P < .01), type of surgery (|r|: 0.127, P < .05), and type of anesthesia (|r|: 0.126, P < .05). In terms of forecasting performance, BPNN (average error: 2.81%) outperforms, yet is less interpretable than MRA (average error: 5.79%). Both MRA and BPNN are technically and economically feasible in generating medical costs of cataract. And some insights on using results of the forecasting model in controlling and reducing disease costs are obtained.

Concordium 2016: Data and Knowledge Transforming Health

Introduction and Context:

Concordium 2016 celebrated the potential for data and knowledge to transform health. Through a series of plenaries, presentations, workshops and demonstrations, the conference highlighted projects among four themes: effectiveness and outcomes research, health care analytics and operations, public and population health, and quality improvement.

Papers in the Special Issue:

The eight papers that comprise this special issue of eGEMs provide exemplars of solutions to the Big Data problems faced in today’s healthcare environment.

Cross-Cutting Elements and Overlapping Themes:

Several of the papers contain elements of multiple overlapping themes. We integrate these into five overlapping themes: telehealth, user-centered design/usability, clinic workflow, patient-centered care, and population health management through prediction modeling and risk adjustment.

Conclusion and Future Directions:

The effort to leverage all types of Big Data to improve health and healthcare is a monumental effort that will require the work of numerous stakeholders, and one that will unfold incrementally over time. This collection of eight papers reflects the current state of the art. Concordium 2017 will take a different form, inviting a small set of leaders in the field to focus on the next round of exciting and provocative research currently underway to improve the nation’s health.