Considerations When Aggregating Data to Measure Performance Across Levels of the Health Care System.

BACKGROUND: Measuring quality at varying levels of the health care system requires attribution, a process of determining the patients and services for which each level is responsible. However, it is important to ensure that attribution approaches are equitable; otherwise, individuals may be assigned differentially based upon social determinants of health. METHODS: First, we used Medicaid claims (2010-2018) from Michigan to assess the proportion of children with sickle cell anemia who had less than 12 months enrollment within a single Medicaid health plan and could therefore not be attributed to a specific health plan. Second, we used the Medicaid Analytic eXtract data (2008-2009) from 26 states to simulate adapting the 30-Day Pediatric All-Condition Readmission measure to the Accountable Care Organization (ACO) level and examined the proportion of readmissions that could not be attributed. RESULTS: For the sickle cell measure, an average of 300 children with sickle cell anemia were enrolled in Michigan Medicaid each year. The proportion of children that could not be attributed to a Medicaid health plan ranged from 12.2% to 89.0% across years. For the readmissions measure, of the 1,051,365 index admissions, 22% were excluded in the ACO-level analysis because of being unable to attribute the patient to a health plan for the 30 days post discharge. CONCLUSIONS: When applying attribution models, it is essential to consider the potential to induce health disparities. Differential attribution may have unintentional consequences that deepen health disparities, particularly when considering incentive programs for health plans to improve the quality of care.

Diagnostic Algorithms to Study Post-Concussion Syndrome Using Electronic Health Records: Validating a Method to Capture an Important Patient Population.

Post-concussion syndrome (PCS) is characterized by persistent cognitive, somatic, and emotional symptoms after a mild traumatic brain injury (mTBI). Genetic and other biological variables may contribute to PCS etiology, and the emergence of biobanks linked to electronic health records (EHRs) offers new opportunities for research on PCS. We sought to validate the EHR data of PCS patients by comparing two diagnostic algorithms deployed in the Vanderbilt University Medical Center de-identified database of 2.8 million patient EHRs. The algorithms identified individuals with PCS by: 1) natural language processing (NLP) of narrative text in the EHR combined with structured demographic, diagnostic, and encounter data; or 2) coded billing and procedure data. The predictive value of each algorithm was assessed, and cases and controls identified by each approach were compared on demographic and medical characteristics. The NLP algorithm identified 507 cases and 10,857 controls. The negative predictive value in controls was 78% and the positive predictive value (PPV) in cases was 82%. Conversely, the coded algorithm identified 1142 patients with two or more PCS billing codes and had a PPV of 76%. Comparisons of PCS controls to both case groups recovered known epidemiology of PCS: cases were more likely than controls to be female and to have pre-morbid diagnoses of anxiety, migraine, and post-traumatic stress disorder. In contrast, controls and cases were equally likely to have attention deficit hyperactive disorder and learning disabilities, in accordance with the findings of recent systematic reviews of PCS risk factors. We conclude that EHRs are a valuable research tool for PCS. Ascertainment based on coded data alone had a predictive value comparable to an NLP algorithm, recovered known PCS risk factors, and maximized the number of included patients.

Imaging findings after acute sport-related concussion in American football players: A systematic review.

Sport-related concussion (SRC) has emerged as a major public health problem. The results of brain imaging studies following SRC have raised questions about long-term neurologic health, but the clinical implications of these findings remain unknown. A systematic review of brain imaging findings after SRC was performed utilizing the following inclusion criteria: football players, brain imaging within 6 months of SRC, and sample size >5. Studies were assessed for: 1) methodology, 2) imaging outcomes, and 3) number of positive statistical comparisons. Imaging was classified as immediate (≤1 week post-injury) or subacute (>1 week to 6 months post-injury). Eleven studies met inclusion criteria. Eight of the 11 studies conducted a total of 809 comparisons of brain function, of which 149 (18%) were statistically significant. Nine of the 11 studies (82%) reported positive immediate findings, but were more likely to be subject to recall bias (86% vs. 0%) and to lack baseline advanced brain imaging (78% vs. 50%) than negative studies. Only 3 of 9 studies that reported subacute findings (33%) reported positive results, and these positive studies were also more likely to be subject to recall bias (100% vs. 40%) and to lack baseline advanced brain imaging (100% vs. 67%) than negative studies. The results of the study demonstrate the transitory nature of positive imaging findings and methodological limitations that complicate study interpretation. Further research is required to correlate imaging findings with clinical outcomes.

An 11-year analysis of peripheral nerve injuries in high school sports.

INTRODUCTION: Sports surveillance databases provide valuable information regarding common ailments, yet fewer studies have focused on more rare peripheral nerve injuries. Our objective was to characterize peripheral nerve injuries in high school athletics with respect to incidence, time loss, mechanism, and diagnoses. METHODS: Sport-related nerve injury data on high school athletes were collected during the 2005/2006 through 2015/2016 academic years via the High School Reporting Information Online (RIO) database. All injuries were reported by certified athletic trainers (ATs). Descriptive statistics were performed. RESULTS: A total of 588 peripheral nerve injuries were recorded during the 2005/06-2015/16 academic years, with an overall incidence of 1.46/100,000 athlete-exposures (AE; 95%CI: 1.34, 1.58). Boys' football had the majority of injuries (71.3%) and the highest injury rate (5.46/100,000AE; 95%CI: 4.93, 5.98), followed by boys' wrestling (7.1%) and boys' baseball (3.4%). Over half (50.3%) of peripheral nerve injuries resulted in time loss < 1 week, while 9.4% resulting in the athletes prematurely ending their seasons. The most common mechanisms were player contact (67.3%), overuse (10.0%), and surface contact (9.7%). A specific diagnosis was available for 40 (6.8%) injuries, including upper extremity stinger (n = 26), spinal cord neurapraxia (n = 3), subacromial nerve impingement (n = 2) neuroma (n = 2), axillary nerve palsy (n = 1), sciatic nerve impingement (n = 1), femoral nerve impingement (n = 1), tarsal tunnel syndrome (n = 1), peroneal neuropathy (n = 1), thoracic outlet syndrome (n = 1), and ulnar nerve subluxation (n = 1). DISCUSSION: Recognized peripheral nerve injuries are rare among high school athletes, occurring most commonly in boys' football. While most are minor, approximately 1:10 were season-ending. Specific diagnoses were available for 7% of injuries, with upper extremity stingers being the most commonly reported diagnosis. Working with ATs to identify and implement methods to obtain more specific diagnostic information via surveillance will help researchers better understand the epidemiology of peripheral nerve injuries.

Association between Practice Participation in a Pediatric-focused Medical Home Learning Collaborative and Reduction of Preventable Emergency Department Visits by Publicly-insured Children in Massachusetts.

INTRODUCTION: This study evaluates the impact of practice participation in a pediatric patient-centered medical home learning collaborative on preventable emergency department (ED) visits among children in MassHealth (Massachusetts Medicaid/Children's Health Insurance Program). METHODS: Claims and enrollment data were extracted for child MassHealth members (aged 3-18) comprising 2 groups: members enrolled in a group of 13 child-serving practices that participated in an intensive, 29-month long patient-centered medical home learning collaborative (intervention group), and members enrolled in a group of 12 comparison practices with roughly similar panel size, type, and geographic location (comparison group). Preventable ED visits were identified using a modified version of the New York University ED algorithm. Two analyses were then conducted: (1) a repeat cross-sectional analysis among children enrolled in intervention or comparison group practices during baseline (first half of 2011) and follow-up (second half of 2013) periods; and (2) a longitudinal analysis among a subset of children enrolled for the full study period (2011-2013). Both analyses tested whether the effect of the intervention differed for children with versus without chronic conditions (effect modification). RESULTS: Preventable ED visits declined from baseline to follow-up among children in both intervention and comparison practices. In the cross-sectional analysis, the decrease was the same in both practice groups, and for children with versus without chronic conditions. The longitudinal analysis shows a statistically significantly greater decrease among children with chronic conditions enrolled in the intervention practices (P = 0.02). CONCLUSION: Children with chronic conditions might receive the greatest benefit from receiving care in a medical home setting.

Long-Term Neuroimaging Findings in American Football Players: Systematic Review.

BACKGROUND: Lasting neuroimaging changes after participation in American football are an increasing public health concern. The clinical relevance of imaging findings remains unknown. METHODS: A systematic review was performed with the following inclusion criteria: football players, brain imaging ≥2 years from previous concussion or retirement, and sample size ≥5. Studies were assessed for 1) methodology (control selection, type I error, and recall bias), 2) imaging outcomes, and 3) number of significant statistical comparisons. RESULTS: Sixteen studies (all cross-sectional studies) met the inclusion criteria. Highest level of play included high school (n = 1), college (n = 3), and professional (n = 12). Thirteen of the 16 studies made a total 456 comparisons of brain activity, of which 171 were statistically significant (38%). Nine of 16 studies (56%) had appropriate controls, and 5 of 16 studies (31%) appropriately accounted for type I error. To obtain player concussion history, all studies (16/16) had recall bias or unclear methodology. Imaging outcome measures included structural magnetic resonance imaging (MRI) (29.2%), diffusion tensor imaging (25%), radioactive tracer uptake on positron emission tomography (16.7%), patterns of connectivity on functional MRI (fMRI) (12.5%), transcranial magnetic stimulation (8.3%), arterial spin labeling MRI (4.2%), and metabolic changes on 1H-magnetic resonance spectroscopy (4.2%). CONCLUSIONS: Long-term neuroimaging findings in American football players are heterogeneous in both methodology and findings. Understanding the clinical importance of statistically significant findings is complicated by methodological limitations and study design. Further research is required to correlate imaging findings with clinical outcomes.

A football helmet prototype that reduces linear and rotational acceleration with the addition of an outer shell.

OBJECTIVEAmid the public health controversy surrounding American football, a helmet that can reduce linear and rotational acceleration has the potential to decrease forces transmitted to the brain. The authors hypothesized that a football helmet with an outer shell would reduce both linear and rotational acceleration. The authors' objectives were to 1) determine an optimal material for a shock-absorbing outer shell and 2) examine the ability of an outer shell to reduce linear and/or rotational acceleration.METHODSA laboratory-based investigation was undertaken using an extra-large Riddell Revolution football helmet. Two materials (Dow Corning Dilatant Compound and Sorbothane) were selected for their non-Newtonian properties (changes in viscosity with shear stress) to develop an outer shell. External pads were attached securely to the helmet at 3 locations: the front boss, the side, and the back. The helmet was impacted 5 times per location at 6 m/sec with pneumatic ram testing. Two-sample t-tests were used to evaluate linear/rotational acceleration differences between a helmet with and a helmet without the outer shell.RESULTSSorbothane was superior to the Dow Corning compound in force reduction and recovered from impact without permanent deformation. Of 5 different grades, 70-duro (a unit of hardness measured with a durometer) Sorbothane was found to have the greatest energy dissipation and stiffness, and it was chosen as the optimal outer-shell material. The helmet prototype with the outer shell reduced linear acceleration by 5.8% (from 75.4g to 71.1g; p < 0.001) and 10.8% (from 89.5g to 79.8g; p = 0.033) at the side and front boss locations, respectively, and reduced rotational acceleration by 49.8% (from 9312.8 rad/sec2 to 4671.7 rad/sed2; p < 0.001) at the front boss location.CONCLUSIONSSorbothane (70 duro) was chosen as the optimal outer-shell material. In the outer-shell prototype helmet, the results demonstrated a 5%-10% reduction in linear acceleration at the side and front boss locations, and a 50% reduction in rotational acceleration at the front boss location. Given the paucity of publicly reported helmet-design literature and the importance of rotational acceleration in head injuries, the substantial reduction seen in rotational acceleration with this outer-shell prototype holds the potential for future helmet-design improvements.

Genotoxicity assessment of ethylenediamine dinitrate (EDDN) and diethylenetriamine trinitrate (DETN).

Ethylenediamine dinitrate (EDDN) and diethylenetriamine trinitrate (DETN) are relatively insensitive explosive compounds that are being explored as safe alternatives to other more sensitive compounds. When used in combination with other high explosives they are an improvement and may provide additional safety during storage and use. The genetic toxicity of these compounds was evaluated to predict the potential adverse human health effects from exposure by using a standard genetic toxicity test battery which included: a gene mutation test in bacteria (Ames), an in vitro Chinese Hamster Ovary (CHO) cell chromosome aberration test and an in vivo mouse micronucleus test. The results of the Ames test showed that EDDN increased the mean number of revertants per plate with strain TA100, without activation, at 5000µg/plate compared to the solvent control, which indicated a positive result. No positive results were observed with the other tester strains with or without activation in Salmonella typhimurium strains TA98, TA1535, TA1537, and Escherichia coli strain WP2 uvrA. DETN was negative for all Salmonella tester strains and E. coli up to 5000µg/plate both with and without metabolic activation. The CHO cell chromosome aberration assay was performed using EDDN and DETN at concentrations up to 5000µg/mL. The results indicate that these compounds did not induce structural chromosomal aberrations at all tested concentrations in CHO cells, with or without metabolic activation. EDDN and DETN, when tested in vivo in the CD-1 mouse at doses up to 2000mg/kg, did not induce any significant increase in the number of micronuclei in bone marrow erythrocytes. These studies demonstrate that EDDN is mutagenic in one strain of Salmonella (TA100) but was negative in other strains, for in vitro induction of chromosomal aberrations in CHO cells, and for micronuclei in the in vivo mouse micronucleus assay. DETN was not genotoxic in all in vitro and in vivo tests. These results show the in vitro and in vivo genotoxicity potential of these chemicals.

Genotoxicity assessment of an energetic propellant compound, 3-nitro-1,2,4-triazol-5-one (NTO).

3-Nitro-1,2,4-triazol-5-one (NTO) is an energetic explosive proposed for use in weapon systems, to reduce the sensitivity of warheads. In order to develop toxicity data for safety assessment, we investigated the genotoxicity of NTO, using a battery of genotoxicity tests, which included the Ames test, Chinese Hamster Ovary (CHO) cell chromosome aberration test, L5178Y TK(+/-) mouse lymphoma mutagenesis test and rat micronucleus test. NTO was not mutagenic in the Ames test or in Escherichia coli (WP2uvrA). NTO did not induce chromosomal aberrations in CHO cells, with or without metabolic activation. In the L5178Y TK(+/-) mouse lymphoma mutagenesis test, all of the NTO-treated cultures had mutant frequencies that were similar to the average frequencies of solvent control-treated cultures, indicating a negative result. Confirmatory tests for the three in vitro tests also produced negative results. The potential in vivo clastogenicity and aneugenicity of NTO was evaluated using the rat peripheral blood micronucleus test. NTO was administered by oral gavage to male and female Sprague-Dawley rats for 14 days at doses up to 2g/kg/day. Flow cytometric analysis of peripheral blood demonstrated no significant induction of micronucleated reticulocytes relative to the vehicle control (PEG-200). These studies reveal that NTO was not genotoxic in either in vitro or in vivo tests and suggest a low risk of genetic hazards associated with exposure.

Genotoxicity of 10 cigarette smoke condensates in four test systems: comparisons between assays and condensates.

The particulate fraction of cigarette smoke, cigarette smoke condensate (CSC), is genotoxic in many short-term in vitro tests and is carcinogenic in rodents. However, no study has evaluated a series of CSCs prepared from a diverse set of cigarettes and produced with different smoking machine regimens in several short-term genotoxicity tests. Here we report on the genotoxicity of 10 CSCs prepared from commercial cigarettes that ranged from ultra-low tar per cigarette (< or =6.5 mg) to full flavor (>14.5 mg) as determined by the Federal Trade Commission (FTC) smoking regimen, a reference cigarette blended to be representative of a U.S. FTC-regimen low-tar cigarette, and experimental cigarettes constructed of single tobacco types. CSCs were tested in the presence of rat liver S9 in the Salmonella plate-incorporation assay using frameshift strains TA98 and YG1041; in micronucleus and comet assays in L5178Y/Tk(+/-) 7.3.2C mouse lymphoma cells, and in CHO-K(1) cells for chromosome aberrations. All 10 CSCs were mutagenic in both strains of Salmonella, and the rank order of their mutagenic potencies was similar. Their mutagenic potencies in Salmonella spanned 7-fold when expressed as rev/mug CSC but 158-fold when expressed as rev/mg nicotine; the range of genotoxic potencies of the CSCs in the other assays was similar regardless of how the data were expressed. All 10 CSCs induced micronuclei with a 3-fold range in their potency. All but one CSC induced DNA damage over a 20-fold range, and all but one CSC induced chromosome aberrations over a 4-fold range. There was no relation among the genotoxic potencies of the CSCs across the assays, and a qualitative advantage of the addition of the other assays to the Salmonella assay was not supported by our findings. Although consideration of nicotine levels may improve the relevance of the quantitative data obtained in the Salmonella and possibly comet assays, compensatory smoking habits and other factors may make the data from the assays used here have qualitative but not quantitative value in assessing risk of cigarette types and cigarette smoking to human health.

Changes in service availability in california hospitals, 1995 to 2002.

Hospitals face serious financial challenges in the current healthcare marketplace. In response to these challenges, they may alter their service offerings, eliminating services that are perceived as money-losing or adding new services in areas where profitability is expected to be greater. Although research has examined hospital closures, the more subtle phenomenon of hospital service changes has not been systematically studied. This issue is important because different types of hospital service changes could have different effects on hospital financial viability: extensive service closures could contribute to a downward spiral leading to hospital closure, whereas adding new services might help improve a hospital's finances. This article' examines changes in hospital service availability in California general acute care hospitals between 1995 and 2002. Our major findings indicate that many California hospitals made changes in their service offerings during the study period, although few made extensive changes. Altogether, about half of the hospitals in our study population either closed or opened at least one service. Nearly one-fourth of the hospitals in our study population closed one or more services, whereas just under one-third opened one or more new services. However, the vast majority of the hospitals that closed or added a service made only one or two such changes. In addition, few hospitals both closed and opened services. The service closed most frequently was normal newborn labor and delivery (obstetrics), whereas inpatient rehabilitation was the most frequently opened service. Hospitals that made the most service changes tended to be small, rural, and financially troubled at the start of the study period. Among this group of hospitals, service closures were associated with continued financial deterioration, whereas new service openings were associated with improvements in key financial ratios.

GIMS: an integrated data storage and analysis environment for genomic and functional data.

Effective analyses in functional genomics require access to many kinds of biological data. For example, the analysis of upregulated genes in a microarray experiment might be aided by information concerning protein interactions or proteins' cellular locations. However, such information is often stored in different formats at different sites, in ways that may not be amenable to integrated analysis. The Genome Information Management System (GIMS) is an object database that integrates genomic data with data on the transcriptome, protein-protein interactions, metabolic pathways and annotations, such as gene ontology terms and identifiers. The resulting system supports the running of analyses over this integrated data resource, and provides comprehensive facilities for handling and interrelating the results of these analyses. GIMS has been used to store Saccharomyces cerevisiae data, and we demonstrate how the integrated storage of diverse types of data can be beneficial for analysis, using combinations of complex queries. As an example, we describe how GIMS has been used to analyse a collection of aryl alcohol dehydrogenase gene deletion mutants. The GIMS database can be accessed remotely using a Java application that can be downloaded from http://img.cs.man.ac.uk/gims.

Mouse lymphoma thymidine kinase gene mutation assay: International Workshop on Genotoxicity Tests Workgroup report--Plymouth, UK 2002.

The Mouse Lymphoma Assay (MLA) Workgroup of the International Workshop on Genotoxicity Tests (IWGT) met on June 28th and 29th, 2002, in Plymouth, England. This meeting of the MLA group was devoted to discussing the criteria for assay acceptance and appropriate approaches to data evaluation. Prior to the meeting, the group conducted an extensive analysis of data from both the microwell and soft agar versions of the assay. For the establishment of criteria for assay acceptance, 10 laboratories (6 using the microwell method and 4 using soft agar) provided data on their background mutant frequencies, plating efficiencies of the negative/vehicle control, cell suspension growth, and positive control mutant frequencies. Using the distribution curves generated from this data, the Workgroup reached consensus on the range of values that should be used to determine whether an individual experiment is acceptable. In order to establish appropriate approaches for data evaluation, the group used a number of statistical methods to evaluate approximately 400 experimental data sets from 10 laboratories entered into a database created for the earlier MLA Workshop held in New Orleans [Environ. Mol. Mutagen. 40 (2002) 292]. While the Workgroup could not, during this meeting, make a final recommendation for the evaluation of data, a general strategy was developed and the Workgroup members agreed to evaluate this new proposed approach using their own laboratory data. This evaluation should lead to a consensus global approach for data evaluation in the near future.

The occupational transformation of the mental health system.

The mental health workforce has changed dramatically since the mid-1970s. Nonphysician providers, particularly psychologists and clinical social workers, have become a much larger share of the workforce. While the supply of psychiatrists has been relatively stable, there has been a dramatic increase in the supply of psychologists and social workers. Changes in clinical practice, combined with the continued expansion of managed care into mental health, will largely determine the future composition and supply of the mental health workforce.

A systematic approach to modeling, capturing, and disseminating proteomics experimental data.

Both the generation and the analysis of proteome data are becoming increasingly widespread, and the field of proteomics is moving incrementally toward high-throughput approaches. Techniques are also increasing in complexity as the relevant technologies evolve. A standard representation of both the methods used and the data generated in proteomics experiments, analogous to that of the MIAME (minimum information about a microarray experiment) guidelines for transcriptomics, and the associated MAGE (microarray gene expression) object model and XML (extensible markup language) implementation, has yet to emerge. This hinders the handling, exchange, and dissemination of proteomics data. Here, we present a UML (unified modeling language) approach to proteomics experimental data, describe XML and SQL (structured query language) implementations of that model, and discuss capture, storage, and dissemination strategies. These make explicit what data might be most usefully captured about proteomics experiments and provide complementary routes toward the implementation of a proteome repository.

Mouse lymphoma thymidine kinase gene mutation assay: follow-up International Workshop on Genotoxicity Test Procedures, New Orleans, Louisiana, April 2000.

The Mouse Lymphoma Assay (MLA) Workgroup of the International Workshop on Genotoxicity Test Procedures held a second harmonization meeting just prior to the U.S. Environmental Mutagen Society Meeting in New Orleans, LA, in April 2000. The discussion focused on several important aspects of the MLA, including: 1) cytotoxicity measures and their determination, 2) use of a 24-hr treatment, 3) the ability of the assay to detect aneugens, and 4) concentration selection. Prior to the meeting the group developed Microsoft Excel Workbooks for data entry. Ten laboratories entered their data into the workbooks (primarily as coded chemicals). The Excel Workbooks were used to facilitate data analysis by generating an extensive set of graphs that were evaluated by the meeting participants. Based on the Workgroup's previous agreement that a single cytotoxicity measure should be established for both the microwell and soft agar versions of the assay, the Workgroup analyzed the submitted data and unanimously agreed that the relative total growth (RTG) should be used as the cytotoxicity measure for concentration selection and data evaluation. The Workgroup also agreed that the various cytotoxicity measures should be calculated using the same methods regardless of whether the soft agar or microwell version of the assay was used. In the absence of sufficient data to make a definitive determination, the Workgroup continued to endorse the International Committee on Harmonization recommendation for the use of 24-hr treatment and made some specific 24-hr treatment protocol recommendations. The Workgroup recognized the ability of the MLA to detect at least some aneugens and also developed general guidance and requirements for appropriate concentration selection.