Assessment of Blood Biomarker Profile After Acute Concussion During Combative Training Among US Military Cadets: A Prospective Study From the NCAA and US Department of Defense CARE Consortium.

Importance: Validation of protein biomarkers for concussion diagnosis and management in military combative training is important, as these injuries occur outside of traditional health care settings and are generally difficult to diagnose. Objective: To investigate acute blood protein levels in military cadets after combative training-associated concussions. Design, Setting, and Participants: This multicenter prospective case-control study was part of a larger cohort study conducted by the National Collegiate Athletic Association and the US Department of Defense Concussion Assessment Research and Education (CARE) Consortium from February 20, 2015, to May 31, 2018. The study was performed among cadets from 2 CARE Consortium Advanced Research Core sites: the US Military Academy at West Point and the US Air Force Academy. Cadets who incurred concussions during combative training (concussion group) were compared with cadets who participated in the same combative training exercises but did not incur concussions (contact-control group). Clinical measures and blood sample collection occurred at baseline, the acute postinjury point (<6 hours), the 24- to 48-hour postinjury point, the asymptomatic postinjury point (defined as the point at which the cadet reported being asymptomatic and began the return-to-activity protocol), and 7 days after return to activity. Biomarker levels and estimated mean differences in biomarker levels were natural log (ln) transformed to decrease the skewness of their distributions. Data were collected from August 1, 2016, to May 31, 2018, and analyses were conducted from March 1, 2019, to January 14, 2020. Exposure: Concussion incurred during combative training. Main Outcomes and Measures: Proteins examined included glial fibrillary acidic protein, ubiquitin C-terminal hydrolase-L1, neurofilament light chain, and tau. Quantification was conducted using a multiplex assay (Simoa; Quanterix Corp). Clinical measures included the Sport Concussion Assessment Tool-Third Edition symptom severity evaluation, the Standardized Assessment of Concussion, the Balance Error Scoring System, and the 18-item Brief Symptom Inventory. Results: Among 103 military service academy cadets, 67 cadets incurred concussions during combative training, and 36 matched cadets who engaged in the same training exercises did not incur concussions. The mean (SD) age of cadets in the concussion group was 18.6 (1.3) years, and 40 cadets (59.7%) were male. The mean (SD) age of matched cadets in the contact-control group was 19.5 (1.3) years, and 25 cadets (69.4%) were male. Compared with cadets in the contact-control group, those in the concussion group had significant increases in glial fibrillary acidic protein (mean difference in ln values, 0.34; 95% CI, 0.18-0.50; P < .001) and ubiquitin C-terminal hydrolase-L1 (mean difference in ln values, 0.97; 95% CI, 0.44-1.50; P < .001) levels at the acute postinjury point. The glial fibrillary acidic protein level remained high in the concussion group compared with the contact-control group at the 24- to 48-hour postinjury point (mean difference in ln values, 0.22; 95% CI, 0.06-0.38; P = .007) and the asymptomatic postinjury point (mean difference in ln values, 0.21; 95% CI, 0.05-0.36; P = .01). The area under the curve for all biomarkers combined, which was used to differentiate cadets in the concussion and contact-control groups, was 0.80 (95% CI, 0.68-0.93; P < .001) at the acute postinjury point. Conclusions and Relevance: This study's findings indicate that blood biomarkers have potential for use as research tools to better understand the pathobiological changes associated with concussion and to assist with injury identification and recovery from combative training-associated concussions among military service academy cadets. These results extend the previous findings of studies of collegiate athletes with sport-associated concussions.

Differences in sport-related concussion for female and male athletes in comparable collegiate sports: a study from the NCAA-DoD Concussion Assessment, Research and Education (CARE) Consortium.

OBJECTIVES: To examine sex differences in sport-related concussion (SRC) across comparable sports. METHODS: Prospective cohort of collegiate athletes enrolled between 2014 and 2017 in the Concussion Assessment, Research and Education Consortium study. RESULTS: Among 1071 concussions (females=615; 57.4%), there was no difference in recovery (median days to full return to play) (females=13.5 (IQR 9.0, 23.1) vs males=11.8 (IQR 8.1, 19.0), p=0.96). In subgroup analyses, female recovery was longer in contact (females=12.7 days (IQR 8.8, 21.4) vs males=11.0 days (IQR 7.9, 16.2), p=0.0021), while male recovery was longer in limited contact sports (males=16.9 days (IQR 9.7, 101.7) vs females=13.8 days (IQR 9.1, 22.0), p<0.0001). There was no overall difference in recovery among Division I schools (females=13.7 (IQR 9.0, 23.1) vs males=12.2 (IQR 8.2 19.7), p=0.5), but females had longer recovery at the Division II/III levels (females=13.0 (IQR 9.2, 22.7) vs males=10.6 (IQR 8.1, 13.9), p=0.0048). CONCLUSION: Overall, no difference in recovery between sexes across comparable women's and men's sports in this collegiate cohort was found. However, females in contact and males in limited contact sports experienced longer recovery times, while females had longer recovery times at the Division II/III level. These disparate outcomes indicate that, while intrinsic biological sex differences in concussion recovery may exist, important, modifiable extrinsic factors may play a role in concussion outcomes.

Acute Sport Concussion Assessment Optimization: A Prospective Assessment from the CARE Consortium.

BACKGROUND: Numerous medical organizations recommend a multifaceted approach to the assessment of concussion occurring during sporting events. A number of tools are available to clinicians, with a wide breadth of sensitivity and specificity; however, little work has been done to evaluate the combined efficiency of these tools in concussed male and female athletes from a broad array of collegiate sports and with variable time from the pre-season baseline evaluation. OBJECTIVE: The aim of this study was to optimize the concussion assessment battery for application within the first 72 h of injury, and to identify the necessary baseline retesting frequency. METHODS: Between 2014 and 2017, a total of 1458 National Collegiate Athletic Association (NCAA) athletes sustaining 1640 diagnosed concussions completed a baseline assessment each year of the investigation and were evaluated up to three times within the first 72 h of injury using a standardized assessment protocol. Classification and regression tree analyses were implemented to identify the most efficient multifaceted assessment pathway to quantify concussion-related outcomes. Results were optimized for assessments occurring within 1 h post-injury, 1-24 h post-injury, and 24-72 h post-injury when using the raw post-injury assessment performance, difference scores from baseline evaluations occurring in the same year, and difference scores from baseline evaluations occurring the year prior. RESULTS: At each of the assessment time points, the analyses indicated that alone or in combination, a symptom evaluation, Balance Error Scoring System (BESS) scores collected on the firm surface, and Standardized Assessment of Concussion (SAC) total score offered the best overall performance when compared with pre-morbid performance captured in the same season. Optimized sensitivity of the multifaceted approach was 61% within 1 h of injury, 67% at the 1-24 h interval, and 55% at the 24-72 h interval when difference scores from the same-season baseline were available. CONCLUSIONS: This investigation identified key concussion assessments in quantifying post-concussion performance among student athletes, that were maximized when same-season pre-morbid evaluations were available. Consistent with clinical recommendations, medical professionals should continue to focus on symptom reporting, postural control, and neurocognitive screening to support the clinical examination when making a concussion diagnosis.

Correction to: Test-Retest Reliability and Interpretation of Common Concussion Assessment Tools: Findings from the NCAA-DoD CARE Consortium.

The article Test-Retest Reliability and Interpretation of Common Concussion Assessment Tools.

Baseline Performance of NCAA Athletes on a Concussion Assessment Battery: A Report from the CARE Consortium.

BACKGROUND: Sport-related concussion and repetitive head impact exposure in contact sports continue to receive increased attention in public and medical spheres. The Concussion Assessment, Research and Education (CARE) Consortium, a multicenter cooperative, was established to study the natural history of concussion in National Collegiate Athletic Association (NCAA) collegiate student-athletes across 29 colleges and universities in the United States. The purpose of this investigation is to provide normative data from the CARE Consortium and evaluate for differences between sport categories. METHODS: NCAA student-athletes were evaluated annually for general demographics and sport-specific characteristics before the start of the competitive season. We collected demographic and medical history information and evaluated each student-athlete's neurocognitive function, neurological status, postural stability, and self-reported symptoms. Sports were categorized by the amount of contact typically associated with the sport (i.e., contact, limited contact, non-contact). Comparisons between the three sport categories for the evaluated variables were made using linear or zero inflated negative binomial regression models adjusted for gender, concussion history, and household income. RESULTS: Over a 2-year period (August 2014-July 2016), 15,681 NCAA athletes completed preseason evaluations. Overall, 53% of the athletes were in the contact sport group, 31% were in the limited contact group and 17% were in the non-contact group. After adjusting for covariates, there were statistically significant differences found between athlete groups, although the differences and effect sizes were small and not clinically significant. The contact sport group had better scores on Immediate Post-Concussion Assessment Testing (ImPACT®) visual and verbal memory, Sport Concussion Assessment Tool (SCAT) symptom checklist, and Brief Symptom Inventory-18 (BSI-18), but slower ImPACT reaction time and worse scores on Standardized Assessment of Concussion (SAC). Further, the data indicate that some ImPACT score distributions were noticeably different from those presented in the technical manual. CONCLUSIONS: In this large, racially and socio-economically diverse cohort of male and female college athletes, we found no evidence that student-athletes participating in contact sports have clinically meaningful deficits in pre-season cognitive and balance testing. They also did not report significantly more symptoms of psychological distress when compared with student-athletes in non-contact or limited contact sports. In addition, the data suggest potential limitations when using published ImPACT norms when evaluating injured athletes.

Test-Retest Reliability and Interpretation of Common Concussion Assessment Tools: Findings from the NCAA-DoD CARE Consortium.

BACKGROUND: Concussion diagnosis is typically made through clinical examination and supported by performance on clinical assessment tools. Performance on commonly implemented and emerging assessment tools is known to vary between administrations, in the absence of concussion. OBJECTIVE: To evaluate the test-retest reliability of commonly implemented and emerging concussion assessment tools across a large nationally representative sample of student-athletes. METHODS: Participants (n = 4874) from the Concussion Assessment, Research, and Education Consortium completed annual baseline assessments on two or three occasions. Each assessment included measures of self-reported concussion symptoms, motor control, brief and extended neurocognitive function, reaction time, oculomotor/oculovestibular function, and quality of life. Consistency between years 1 and 2 and 1 and 3 were estimated using intraclass correlation coefficients or Kappa and effect sizes (Cohen's d). Clinical interpretation guidelines were also generated using confidence intervals to account for non-normally distributed data. RESULTS: Reliability for the self-reported concussion symptoms, motor control, and brief and extended neurocognitive assessments from year 1 to 2 ranged from 0.30 to 0.72 while effect sizes ranged from 0.01 to 0.28 (i.e., small). The reliability for these same measures ranged from 0.34 to 0.66 for the year 1-3 interval with effect sizes ranging from 0.05 to 0.42 (i.e., small to less than medium). The year 1-2 reliability for the reaction time, oculomotor/oculovestibular function, and quality-of-life measures ranged from 0.28 to 0.74 with effect sizes from 0.01 to 0.38 (i.e., small to less than medium effects). CONCLUSIONS: This investigation noted less than optimal reliability for most common and emerging concussion assessment tools. Despite this finding, their use is still necessitated by the absence of a gold standard diagnostic measure, with the ultimate goal of developing more refined and sound tools for clinical use. Clinical interpretation guidelines are provided for the clinician to apply with a degree of certainty in application.

A National Study on the Effects of Concussion in Collegiate Athletes and US Military Service Academy Members: The NCAA-DoD Concussion Assessment, Research and Education (CARE) Consortium Structure and Methods.

BACKGROUND: The natural history of mild traumatic brain injury (TBI) or concussion remains poorly defined and no objective biomarker of physiological recovery exists for clinical use. The National Collegiate Athletic Association (NCAA) and the US Department of Defense (DoD) established the Concussion Assessment, Research and Education (CARE) Consortium to study the natural history of clinical and neurobiological recovery after concussion in the service of improved injury prevention, safety and medical care for student-athletes and military personnel. OBJECTIVES: The objectives of this paper were to (i) describe the background and driving rationale for the CARE Consortium; (ii) outline the infrastructure of the Consortium policies, procedures, and governance; (iii) describe the longitudinal 6-month clinical and neurobiological study methodology; and (iv) characterize special considerations in the design and implementation of a multicenter trial. METHODS: Beginning Fall 2014, CARE Consortium institutions have recruited and enrolled 23,533 student-athletes and military service academy students (approximately 90% of eligible student-athletes and cadets; 64.6% male, 35.4% female). A total of 1174 concussions have been diagnosed in participating subjects, with both concussion and baseline cases deposited in the Federal Interagency Traumatic Brain Injury Research (FITBIR) database. CONCLUSIONS: Challenges have included coordinating regulatory issues across civilian and military institutions, operationalizing study procedures, neuroimaging protocol harmonization across sites and platforms, construction and maintenance of a relational database, and data quality and integrity monitoring. The NCAA-DoD CARE Consortium represents a comprehensive investigation of concussion in student-athletes and military service academy students. The richly characterized study sample and multidimensional approach provide an opportunity to advance the field of concussion science, not only among student athletes but in all populations at risk for mild TBI.

Can lay health workers promote better medical self-management by persons living with HIV? An evaluation of the Positive Choices program.

OBJECTIVE: To evaluate Positive Choices (PC), a program that employed lay health workers to motivate antiretroviral adherence among persons living with HIV with coverage from Indiana's high-risk insurance pool. METHODS: Four hundred and forty nine participants living in the greater Indianapolis area were randomly allocated to treatment (n = 91) or control (n = 358) groups and followed for one year. RESULTS: Compared to control subjects, PC subjects were more likely to adhere to HIV medications (medication possession ratio adherence ≥ 0.95, OR = 1.83, p = 0.046), and to achieve undetectable viral load (<50 copies/mL, OR = 2.01, p = 0.011) in the 12 months following introduction of PC. There were no significant differences observed between groups in any of self-reported health status indicators. CONCLUSION: Estimates suggest that PC clients were 16% more likely to have undetectable viral loads than clients in standard care. The incremental program cost was approximately $10,000 for each additional person who achieved an undetectable viral load. PRACTICE IMPLICATIONS: As persons living with HIV experience greater longevity and healthcare reform expands coverage to these high-risk populations, greater demands will be placed on the HIV-care workforce. Results suggest lay health workers may serve as effective adjuncts to professional care providers.

Chlamydia screening among young women: individual- and provider-level differences in testing.

OBJECTIVE: We assessed differences in chlamydia screening rates according to race/ethnicity, insurance status, age, and previous sexually transmitted infection (STI) or pregnancy. METHODS: A retrospective cohort study was performed using electronic medical record and billing data for women 14 to 25 years of age in 2002-2007, assessing differences in the odds of a chlamydia test being performed at that visit. RESULTS: Adjusted odds of a chlamydia test being performed were lower among women 14 to 15 years of age (odds ratio: 0.83 [95% confidence interval: 0.70-1.00]) and 20 to 25 years of age (20-21 years, odds ratio: 0.78 [95% confidence interval: 0.70-0.89]; 22-23 years, odds ratio: 0.76 [95% confidence interval: 0.67-0.87]; 24-25 years, odds ratio: 0.64 [95% confidence interval: 0.57-0.73]), compared with women 18 to 19 years of age. Black women had 3 times increased odds (odds ratio: 2.96 [95% confidence interval: 2.66-3.28]) and Hispanic women nearly 13 times increased odds (odds ratio: 12.89 [95% confidence interval: 10.85-15.30]) of testing, compared with white women. Women with public (odds ratio: 1.74 [95% confidence interval: 1.58-1.91]) and public pending (odds ratio: 6.85 [95% confidence interval: 5.13-9.15]) insurance had increased odds of testing, compared with women with private insurance. After first STI diagnosis, differences according to race/ethnicity persisted but were smaller; after first pregnancy, differences persisted. CONCLUSIONS: Despite recommendations to screen all sexually active young women for chlamydia, providers screened women differently according to age, race/ethnicity, and insurance status, although differences were reduced after first STI or pregnancy.

Case finding with incomplete administrative data: observations on playing with less than a full deck.

Capacity constraints and efficiency considerations require that disease management programs identify patients most likely to benefit from intervention. Predictive modeling with available administrative data has been used as a strategy to match patients with appropriate interventions. Administrative data, however, can be plagued by problems of incompleteness and delays in processing. In this article, we examine the effects of these problems on the effectiveness of using administrative data to identify suitable candidates for disease management, and we evaluate various proposed solutions. We build prospective models using regression analysis and evaluate the resulting stratification algorithms using R² statistics, areas under receiver operator characteristic curves, and cost concentration ratios. We find delays in receipt of data reduce the effectiveness of the stratification algorithm, but the degree of compromise depends on what proportion of the population is targeted for intervention. Surprisingly, we find that supplementing partial data with a longer panel of more outdated data produces algorithms that are inferior to algorithms based on a shorter window of more recent data. Demographic data add little to algorithms that include prior claims data, and are an inadequate substitute when claims data are unavailable. Supplementing demographic data with additional information on self-reported health status improves the stratification performance only slightly and only when disease management is targeted to the highest risk patients. We conclude that the extra costs associated with surveying patients for health status information or retrieving older claims data cannot be justified given the lack of evidence that either improves the effectiveness of the stratification algorithm.

The structure and content of telephonic scripts found useful in a Medicaid Chronic Disease Management Program.

In 2003, the Indiana Office of Medicaid Policy and Planning launched the Indiana Chronic Disease Management Program (ICDMP), a programme intended to improve the health and healthcare utilization of 15,000 Aged, Blind and Disabled Medicaid members living with diabetes and/or congestive heart failure in Indiana. Within ICDMP, programme components derived from the Chronic Care Model and education based on an integrated theoretical framework were utilized to create a telephonic care management intervention that was delivered by trained, non-clinical Care Managers (CMs) working under the supervision of a Registered Nurse. CMs utilized computer-assisted health education scripts to address clinically important topics, including medication adherence, diet, exercise and prevention of disease-specific complications. Employing reflective listening techniques, barriers to optimal self-management were assessed and members were encouraged to engage in health-improving actions. ICDMP evaluation results suggest that this low-intensity telephonic intervention shifted utilization and lowered costs. We discuss this patient-centred method for motivating behaviour change, the theoretical constructs underlying the scripts and the branched-logic format that makes them suitable to use as a computer-based application. Our aim is to share these public-domain materials with other programmes.

The Indiana Chronic Disease Management Program's impact on medicaid claims: a longitudinal, statewide evaluation.

BACKGROUND: : Disease management programs have grown in popularity over the past decade as a strategy to curb escalating healthcare costs for persons with chronic diseases. OBJECTIVES: : To evaluate the effect of the Indiana Chronic Disease Management Program (ICDMP) on the longitudinal changes in Medicaid claims statewide. RESEARCH DESIGN: : Phased implementation of a chronic disease management program in 3 regions of the state. Fourteen repeated cohorts of Medicaid members were drawn over a period of 3.5 years and the trends in claims were evaluated using a repeated measures model. SUBJECTS: : A total of 44,218 Medicaid members with diabetes and/or congestive heart failure in 3 geographic regions in Indiana. RESULTS: : Across all 3 regions and both disease classes, we found a flattening of cost trends between the pre- and post-ICDMP-initiation periods. This change in the slopes was significant for all of the models except for congestive heart failure in southern Indiana. Thus, the average per member claims paid was increasing at a faster rate before ICDMP but slowed once the program was initiated. To distinguish shorter and longer-term effects related to ICDMP, we estimated annual slopes within the pre- and post-ICDMP- time periods. A similar pattern was found in all regions: claims were increasing before ICDMP, flattened in the years around program initiation, and remained flat in the final year of follow-up. CONCLUSIONS: : This analysis shows that the trend in average total claims changed significantly after the implementation of ICDMP, with a decline in the rate of increase in claims paid observed for targeted Medicaid program populations across the state of Indiana.

An evaluation of educational outreach to improve evidence-based prescribing in Medicaid: a cautionary tale.

RATIONALE, AIMS AND OBJECTIVES: Evidence suggests that educational outreach ('academic detailing') improves evidence-based prescribing. We evaluated the impact of an academic detailing programme intended to increase new statin prescriptions. METHODS: In a 2 x 2 factorial design we evaluated the effect of an academic detailing programme with/without telephonic care management for patients. Eligible patients were continuously enrolled Medicaid members at high risk for cardiovascular disease utilization who were not receiving statin medication in the 18 months prior to the intervention. All primary care prescribers assigned to these patients were randomized by clinic to academic detailing. Two trained nurses provided the detailing to prescribers, including specific discussion about the use of statins in this high-risk patient population. Nurses left the prescribers with a summary of clinical practice guidelines, a one-page detailing sheet and a list of patients under the care of the prescriber who were candidates for statins. The primary outcome was the incidence of a new statin prescription claim during the 6-month intervention period and the subsequent 6 months. Logistic regression models were used to estimate main effects of the interventions and to adjust for potential confounding variables in the study. RESULTS: Forty-eight clinics were randomized, effectively randomizing a total of 284 patients and 128 prescribers. Among the 284 patients, 46 (16%) received a new statin claim during the evaluation period. Controlling for significant bivariate associations, the academic detailing intervention had no significant effect on new statin prescriptions compared with the control group (odds ratio = 0.8, 95% confidence interval: 0.4-1.6, P = 0.5). CONCLUSION: Among this Medicaid population at high risk for cardiovascular events, an academic detailing programme to increase statin prescriptions was not effective. To assist others to learn from our failed effort, we identify and discuss critical elements in the design and implementation of the programme that could account for these results.

The net fiscal impact of a chronic disease management program: Indiana Medicaid.

In 2003 the Indiana Office of Medicaid Policy and Planning implemented the Indiana Chronic Disease Management Program (ICDMP). This paper reports on the fiscal impact of the ICDMP from the state's perspective, as estimated from the outcomes of a randomized trial. Medicaid members with congestive heart failure (CHF) or diabetes, or both, were randomly assigned by practice site to chronic disease management services or standard care. The effect of the ICDMP varied by disease group and risk class: while cost savings were achieved in the CHF subgroup, disease management targeted to patients with only diabetes resulted in no significant fiscal impact.

The Indiana Chronic Disease Management Program.

The Indiana Chronic Disease Management Program (ICDMP) is intended to improve the quality and cost-effectiveness of care for Medicaid members with congestive heart failure (chronic heart failure), diabetes, asthma, and other conditions. The ICDMP is being assembled by Indiana Medicaid primarily from state and local resources and has seven components: (1) identification of eligible participants to create regional registries, (2) risk stratification of eligible participants, (3) nurse care management for high-risk participants, (4) telephonic intervention for all participants, (5) an Internet-based information system, (6) quality improvement collaboratives for primary care practices, and (7) program evaluation. The evaluation involves a randomized controlled trial in two inner-city group practices, as well as a statewide observational design. This article describes the ICDMP, highlights challenges, and discusses approaches to its evaluation.

Telephonic case-finding of major depression in a Medicaid chronic disease management program for diabetes and heart failure.

OBJECTIVE: Major depression is common in low-income and chronically ill persons and is a barrier for effective chronic disease care. We evaluated a Medicaid-sponsored strategy for detecting depressive symptoms in adults with diabetes or congestive heart failure. METHODS: Using a two-item screening tool, 890 adults enrolled in the Indiana Chronic Disease Management Program were assessed by telephone for depressive symptoms between December 2003 and March 2004. A subset of 386 participants also completed the eight-item Patient Health Questionnaire (PHQ-8) depression measure. Antidepressant use was examined using pharmacy claims. RESULTS: Depressed mood or anhedonia was reported by 51% of participants. About one in four participants had a PHQ-8 score indicating a high risk for major depression (score >or=10). The two-item screen was 96% sensitive [95% confidence interval (CI), 89-99%] and 60% specific (95% CI, 54-65%) for identifying members at high risk for depression by the full PHQ-8 instrument. Only half of participants with high-risk PHQ-8 scores had a pharmacy claim indicating that an antidepressant medication was filled within 120 days of the depression screening. CONCLUSIONS: A two-stage, telephonic approach involving the PHQ-8 instrument for Medicaid members with either depressed mood or anhedonia could identify two clinically depressed persons for every nine members screened.

Indiana chronic disease management program risk stratification analysis.

OBJECTIVE: The objective of this study was to compare the ability of risk stratification models derived from administrative data to classify groups of patients for enrollment in a tailored chronic disease management program. SUBJECTS: This study included 19,548 Medicaid patients with chronic heart failure or diabetes in the Indiana Medicaid data warehouse during 2001 and 2002. MEASURES: To predict costs (total claims paid) in FY 2002, we considered candidate predictor variables available in FY 2001, including patient characteristics, the number and type of prescription medications, laboratory tests, pharmacy charges, and utilization of primary, specialty, inpatient, emergency department, nursing home, and home health care. METHODS: We built prospective models to identify patients with different levels of expenditure. Model fit was assessed using R statistics, whereas discrimination was assessed using the weighted kappa statistic, predictive ratios, and the area under the receiver operating characteristic curve. RESULTS: We found a simple least-squares regression model in which logged total charges in FY 2002 were regressed on the log of total charges in FY 2001, the number of prescriptions filled in FY 2001, and the FY 2001 eligibility category, performed as well as more complex models. This simple 3-parameter model had an R of 0.30 and, in terms in classification efficiency, had a sensitivity of 0.57, a specificity of 0.90, an area under the receiver operator curve of 0.80, and a weighted kappa statistic of 0.51. CONCLUSION: This simple model based on readily available administrative data stratified Medicaid members according to predicted future utilization as well as more complicated models.