Medicare meets the cloud: the development of a secure platform for the storage and analysis of claims data.

Introduction: Cloud-based solutions are a modern-day necessity for data intense computing. This case report describes in detail the development and implementation of Amazon Web Services (AWS) at Emory-a secure, reliable, and scalable platform to store and analyze identifiable research data from the Centers for Medicare and Medicaid Services (CMS). Materials and Methods: Interdisciplinary teams from CMS, MBL Technologies, and Emory University collaborated to ensure compliance with CMS policy that consolidates laws, regulations, and other drivers of information security and privacy. Results: A dedicated team of individuals ensured successful transition from a physical storage server to a cloud-based environment. This included implementing access controls, vulnerability scanning, and audit logs that are reviewed regularly with a remediation plan. User adaptation required specific training to overcome the challenges of cloud computing. Conclusion: Challenges created opportunities for lessons learned through the creation of an end-product accepted by CMS and shared across disciplines university-wide.

Medication reconciliation: comparing a customized medication history form to a standard medication form in a specialty clinic (CAMPII 2).

BACKGROUND: Medication history forms completed by patients are an essential part of the medication reconciliation process. OBJECTIVE: In a crossover prospective study, investigators compared the accuracy and acceptability of a "fill-in-the blank" medication history form (USUAL) to a customized form (CUSTOM) that contained a checklist of the 44 most frequently prescribed diabetes clinic medications. METHODS: The content of both forms was compared to a "gold-standard" medication list compiled by a clinical pharmacist who conducted a medication history and reviewed pharmacy profiles and medical chart. Subject preference and time to complete the forms were also determined. Accurate was defined as complete and correct (name, dose, and frequency) relative to the gold standard. RESULTS: A total of 77 subjects completed both forms. Complete list accuracy was poor; there was no difference in the accuracy between CUSTOM (6.5%) and USUAL (9.1%) (odds ratio [OR], 0.33; P = 0.62). Out of a total of 648 medications, subjects accurately listed 43.7% of medications on CUSTOM and 45.5% on USUAL (OR, 0.88; P = 0.41). The 44 medications on the checklist were more than twice as likely to be accurately reported using CUSTOM than with USUAL (OR, 2.1; P = 0.0002). More subjects preferred CUSTOM (65.7%) compared with USUAL (32.8%, P = 0.007). CONCLUSION: Medication self-report is very poor, and few subjects created an accurate list on either form. Subjects were more likely to report the drugs on the checklist using CUSTOM than when they used USUAL; however, there was no difference in the overall accuracy between CUSTOM and USUAL.

A Temporal Abstraction-based Extract, Transform and Load Process for Creating Registry Databases for Research.

In the CTSA era there is great interest in aggregating and comparing populations across institutions. These sites likely represent data differently in their clinical data warehouses and other databases. Clinical data warehouses frequently are structured in a generalized way that supports many constituencies. For research, there is a need to transform these heterogeneous data into a shared representation, and to perform categorization and interpretation to optimize the data representation for investigators. We are addressing this need by extending an existing temporal abstraction-based clinical database query system, PROTEMPA. The extended system allows specifying data types of interest in federated databases, extracting the data into a shared representation, transforming it through categorization and interpretation, and loading it into a registry database that can be refreshed. Such a registry's access control, data representation and query tools can be tailored to the needs of research while keeping local databases as the source of truth.

The "metabolic syndrome" is less useful than random plasma glucose to screen for glucose intolerance.

AIMS: To compare the utility of metabolic syndrome (MetS) to random plasma glucose (RPG) in identifying people with diabetes or prediabetes. METHODS: RPG was measured and an OGTT was performed in 1155 adults. Test performance was measured by area under the receiver-operating-characteristic curve (AROC). RESULTS: Diabetes was found in 5.1% and prediabetes in 20.0%. AROC for MetS with fasting plasma glucose (FPG) was 0.80 to detect diabetes, and 0.76 for diabetes or prediabetes--similar to RPG alone (0.82 and 0.72). However, the AROC for MetS excluding fasting plasma glucose was lower: 0.69 for diabetes (p<0.01 vs. both RPG and MetS with FPG), and 0.69 for diabetes or prediabetes. AROCs for MetS with FPG and RPG were comparable and higher for recognizing diabetes in blacks vs. whites, and females vs. males. MetS with FPG was superior to RPG for identifying diabetes only in subjects with age <40 or BMI <25. CONCLUSIONS: MetS features can be used to identify risk of diabetes, but predictive usefulness is driven largely by FPG. Overall, to identify diabetes or prediabetes in blacks and whites with varying age and BMI, MetS is no better than RPG--a more convenient and less expensive test.

Random plasma glucose in serendipitous screening for glucose intolerance: screening for impaired glucose tolerance study 2.

BACKGROUND: With positive results from diabetes prevention studies, there is interest in convenient ways to incorporate screening for glucose intolerance into routine care and to limit the need for fasting diagnostic tests. OBJECTIVE: The aim of this study is to determine whether random plasma glucose (RPG) could be used to screen for glucose intolerance. DESIGN: This is a cross-sectional study. PARTICIPANTS: The participants of this study include a voluntary sample of 990 adults not known to have diabetes. MEASUREMENTS: RPG was measured, and each subject had a 75-g oral glucose tolerance test several weeks later. Glucose intolerance targets included diabetes, impaired glucose tolerance (IGT), and impaired fasting glucose(110) (IFG(110); fasting glucose, 110-125 mg/dl, and 2 h glucose < 140 mg/dl). Screening performance was measured by area under receiver operating characteristic curves (AROC). RESULTS: Mean age was 48 years, and body mass index (BMI) was 30.4 kg/m(2); 66% were women, and 52% were black; 5.1% had previously unrecognized diabetes, and 24.0% had any "high-risk" glucose intolerance (diabetes or IGT or IFG(110)). The AROC was 0.80 (95% CI 0.74-0.86) for RPG to identify diabetes and 0.72 (0.68-0.75) to identify any glucose intolerance, both highly significant (p < 0.001). Screening performance was generally consistent at different times of the day, regardless of meal status, and across a range of risk factors such as age, BMI, high density lipoprotein cholesterol, triglycerides, and blood pressure. CONCLUSIONS: RPG values should be considered by health care providers to be an opportunistic initial screening test and used to prompt further evaluation of patients at risk of glucose intolerance. Such "serendipitous screening" could help to identify unrecognized diabetes and prediabetes.

Relationship of diabetes with cardiovascular disease-related hospitalization rates, length of stay, and charges: analysis by race/ethnicity, age, and sex.

OBJECTIVE: Determine relationship of diabetes with risk of cardiovascular disease hospitalizations and the effect on hospital length of stay and charges. DESIGN: A cross-sectional analysis of Georgia hospital discharge data for 1998 through 2001. PATIENTS: Patients hospitalized principally with one of six cardiovascular disease (CVD) conditions (myocardial infarction, ischemic heart disease, cardiac dysrhythmia, heart failure, cerebrovascular events, peripheral vascular disease) were identified in the hospital discharge data. MAIN OUTCOME MEASURES: Aggregated CVD-related hospitalization rates, length of stay, and charges were compared by presence of diabetes. Analyses were adjusted for age, sex, and race/ethnicity. RESULTS: A total of 3,900,337 discharges were recorded between 1998 to 2001. Of these, 468,957 discharges (12%) had one of the six selected CVD diagnoses (average age 67 years, average length of stay 4.7 days, average total charge $15,702, 48% women, 76% non-Hispanic Whites, 22% non-Hispanic Blacks, and 1% Hispanics). Diabetes was a concurrent diagnosis in 30% of these CVD-related discharges. CVD hospitalization rates were significantly higher and length of stay and total charges were significantly greater among non-Hispanic Whites and Blacks-but not in Hispanics-with diabetes compared to persons without diabetes. Diabetes had a similar effect on CVD hospitalizations among men and women, but the effect of diabetes was lessened with increasing age. CONCLUSION: These data suggests that aggressive outpatient modification of metabolic abnormalities in diabetes patients should be attempted to decrease risk of CVD-related hospitalization and lower the economic impact of these combined conditions.

Potentially modifiable metabolic factors and the risk of cardiovascular disease hospitalizations in urban African Americans with diabetes.

OBJECTIVE: Diabetes and cardiovascular disease (CVD) are frequent causes of hospitalization in African Americans but have rarely been studied as coexisting diagnoses. We analyzed data from an urban African American diabetes patient population to identify variables associated with CVD hospitalizations. DESIGN: Demographic, disease, and metabolic characteristics of patients seen from 1991 to 1997 were extracted from an electronic patient tracking system. Data were linked to a statewide hospital discharge dataset to establish who was hospitalized between 1998 and 2001. Patients with a CVD hospitalization were compared to patients without a CVD hospitalization. RESULTS: 3397 diabetes patients (average age, 56 years; 65% women; 92% African American) were included in the analysis; 24% had hospitalizations primarily due to CVD. Persons with CVD hospitalizations were older and had diabetes longer, and fewer were women. Mean systolic blood pressure (SBP), low-density lipoprotein (LDL) cholesterol, triglyceride, and total cholesterol levels and urinary albumin/creatinine ratio were all higher among persons with CVD hospitalizations. In adjusted analyses, women had lower odds of experiencing a CVD hospitalization, but advancing age, diabetes duration, SBP, and LDL cholesterol were all associated with greater odds. CONCLUSIONS: In this predominantly African American patient sample with diabetes, specific factors (age, sex, diabetes duration, LDL cholesterol, SBP) were associated with CVD hospitalizations. Additional studies are needed to determine whether management of metabolic risk factors in outpatient settings will translate into lower hospitalization rates due to CVD in this population.

Poor glycemic control increases risk of hospitalization in urban African Americans with diabetes.

OBJECTIVE: Hospitalizations due to diabetes are more frequent among African Americans, but risk factors are not known. We analyzed data from an urban African American patient population to identify variables associated with hospitalizations attributable principally to diabetes. DESIGN: Demographic, disease, and metabolic characteristics on patients seen in an outpatient diabetes clinic during 1991 to 1997 were extracted from an electronic patient tracking system. Data were linked to a statewide hospital discharge dataset to capture all in-state hospitalizations from 1998 to 2001. Persons who required a hospitalization for diabetes were compared to the remainder of individuals in the database. RESULTS: A total of 3397 diabetes patients (average age 56 years; 65% women; 92% African American) were included in the analysis; 12% had a hospitalization primarily due to diabetes. Persons with a diabetes hospitalization were younger and had diabetes longer, and fewer were women. In addition, persons who had a diabetes-related hospitalization had evidence of poorer glycemic control with higher hemoglobin A1C (HbA1C) levels. Both the absolute change and rate of decline in HbA1C was less in persons who were hospitalized. In adjusted analyses, duration of diabetes and HbA1C remained significantly associated with risk of a diabetes hospitalization. CONCLUSIONS: In this predominantly African American patient sample with diabetes, poorer glycemic control increased the chances of hospitalization due to diabetes. Continued efforts to aggressively control hyperglycemia could decrease the need for a diabetes hospitalization in this population.

Common reasons for hospitalization among adult patients with diabetes.

OBJECTIVE: To determine reasons for hospitalization among adult patients with diabetes. METHODS: A cross-sectional analysis was conducted of hospital discharges in the state of Georgia for the years 1998 through 2001 that contained either a primary or a coexisting diagnosis of diabetes. With use of the Clinical Classification Software of the Agency for Healthcare Research and Quality, the principal diagnoses among diabetes-related hospital discharges were organized into diagnostic categories. RESULTS: Diabetes was listed as a diagnosis in 14% of all Georgia hospital discharges of adult patients during our study period (57% women; 62% non-Hispanic white; mean age, 64 years; mean length of stay, 5.7 days; and mean hospital charge, 13,540 dollars). Among patients with a diagnosis of diabetes, the 3 most common categories of discharges were "diseases of the circulatory system" (33%), "endocrine, nutritional, and metabolic; immunity disorders" (13%), and "diseases of the respiratory system: (11%). When infections were identified and aggregated, however, these conditions became the second most frequent discharge category (14% of all hospital discharges among patients with diabetes). "Congestive heart failure," "coronary atherosclerosis," and "acute myocardial infarction" were the first, second, and fifth most frequently found unique diagnoses, respectively, among patients with diabetes. CONCLUSION: In this study, diseases of the circulatory system were the most common diagnoses in hospital discharge data for adult patients with diabetes in Georgia. Hospitals should be cognizant of the increased burden placed on them by diabetes, and outpatient treatment of diabetes should focus on prevention of cardiovascular diseases to avoid hospitalizations.

An initial map of insertion and deletion (INDEL) variation in the human genome.

Although many studies have been conducted to identify single nucleotide polymorphisms (SNPs) in humans, few studies have been conducted to identify alternative forms of natural genetic variation, such as insertion and deletion (INDEL) polymorphisms. In this report, we describe an initial map of human INDEL variation that contains 415,436 unique INDEL polymorphisms. These INDELs were identified with a computational approach using DNA re-sequencing traces that originally were generated for SNP discovery projects. They range from 1 bp to 9989 bp in length and are split almost equally between insertions and deletions, relative to the chimpanzee genome sequence. Five major classes of INDELs were identified, including (1) insertions and deletions of single-base pairs, (2) monomeric base pair expansions, (3) multi-base pair expansions of 2-15 bp repeat units, (4) transposon insertions, and (5) INDELs containing random DNA sequences. Our INDELs are distributed throughout the human genome with an average density of one INDEL per 7.2 kb of DNA. Variation hotspots were identified with up to 48-fold regional increases in INDEL and/or SNP variation compared with the chromosomal averages for the same chromosomes. Over 148,000 INDELs (35.7%) were identified within known genes, and 5542 of these INDELs were located in the promoters and exons of genes, where gene function would be expected to be influenced the greatest. All INDELs in this study have been deposited into dbSNP and have been integrated into maps of human genetic variation that are available to the research community.

Disparities in diabetes-related hospitalizations: relationship of age, sex, and race/ethnicity with hospital discharges, lengths of stay, and direct inpatient charges.

OBJECTIVE: To identify any differences in hospitalization rates of diabetes patients by age, sex, or race/ethnicity. DESIGN: A cross-sectional study of Georgia hospital discharge data between 1998 and 2001. PATIENTS/PARTICIPANTS: Patients with a principal discharge diagnosis of diabetes. MAIN OUTCOME MEASURES: Adjusted hospitalization data (discharge rates, length of stay, direct charges) reported as standardized rates per 10,000 person-years, standardized rate differences, and standardized rate ratios, compared by age, sex, and race/ethnicity. RESULTS: Diabetes was the principal diagnosis in 50,301 discharges (average age, 51 years; length of stay, 5.1 days; median total charge, $5893). Persons > or = 60 years old had higher discharge rates, longer stays, and higher charges than persons 18-29 years old. Women had fewer hospitalizations, shorter stays, and lower charges than men. Non-Hispanic Blacks had more than three times as many hospitalizations, markedly longer stays, and higher charges than non-Hispanic Whites. Hispanics with diabetes had lower hospitalization rates, shorter stays, and lower charges than Whites. CONCLUSIONS: Differences by age, sex, and race/ethnicity in hospital discharge rates, lengths of stay, and charges exist for diabetes inpatients. Further study should examine potential causes (severity of disease, comorbidity, and differential access to preventive care) of these disparities.

Recently mobilized transposons in the human and chimpanzee genomes.

Transposable genetic elements are abundant in the genomes of most organisms, including humans. These endogenous mutagens can alter genes, promote genomic rearrangements, and may help to drive the speciation of organisms. In this study, we identified almost 11,000 transposon copies that are differentially present in the human and chimpanzee genomes. Most of these transposon copies were mobilized after the existence of a common ancestor of humans and chimpanzees, approximately 6 million years ago. Alu, L1, and SVA insertions accounted for >95% of the insertions in both species. Our data indicate that humans have supported higher levels of transposition than have chimpanzees during the past several million years and have amplified different transposon subfamilies. In both species, approximately 34% of the insertions were located within known genes. These insertions represent a form of species-specific genetic variation that may have contributed to the differential evolution of humans and chimpanzees. In addition to providing an initial overview of recently mobilized elements, our collections will be useful for assessing the impact of these insertions on their hosts and for studying the transposition mechanisms of these elements.

Increasing patient empowerment and improving diabetes care by utilizing a computer-based patient "roadmap".

The 20.8 million Americans with diabetes are at risk of amputation, kidney failure, blindness, and death which could be decreased if glucose control were better. Patients need motivation and empowerment to perform the daily management of diabetes. We are using informatics to help them organize their questions for providers and to generate "road maps" of their progress and future directions of care.

An informatics model for clinical research--screening for impaired glucose tolerance study.

A major challenge for large scale, multi-site research studies is the time-sensitive tracking of numerous subjects, samples and results through multiple encounters and processes, by many team members in several different locations. This may be compounded by collaborations with other studies to share samples, data collection and subjects. The Screening for Impaired Glucose Tolerance (SIGT) study has met this challenge by developing a secure, informatics system with web and client server interfaces.

An informatics-supported intervention improves diabetes control in a primary care setting.

Although research has shown that proper management of diabetes can improve outcomes, glucose control is worsening. This partly reflects the failure of providers to intensify diabetes therapy when indicated, termed clinical inertia. Our intervention used (a) decision support reminders which provided patient specific recommendations for management at each visit, and (b) computer generated provider specific feedback on performance. This intervention improved the frequency with which providers intensified the therapy and improved glycemic control.

Common reasons for hospitalization in urban diabetes patients.

OBJECTIVES: Determine principal reasons for hospitalization in a predominantly urban, African American diabetes patient population. DESIGN: Data for outpatients with a diagnosis of diabetes were abstracted from electronic records. The number of hospitalizations from 1998 through 2001 was determined after linking our dataset with a statewide discharge dataset. Principal diagnoses were grouped into 18 multilevel diagnostic classes using the Agency for Healthcare Research and Quality's Clinical Classifications Software. PATIENTS: A total of 6505 unique patients had 20,344 discharges from 1998 through 2001; 92% were listed as African Americans and 61% as women. MAIN OUTCOME MEASURES: Frequency of each multilevel diagnostic class and the most commonly occurring diagnoses. RESULTS: The most common multilevel diagnostic classes were "diseases of the circulatory system" (29.0% of all discharges) and "endocrine, nutritional, and metabolic; immunity disorders" (17.1%). The five most commonly occurring unique diagnoses were "congestive heart failure," "diabetes with ketoacidosis or uncontrolled diabetes," "coronary atherosclerosis," "diabetes with other manifestations," and "pneumonia, organism unspecified." Nearly 16% of all discharged patients had diagnoses related to infection. The five most frequent diagnoses related to infection were "pneumonia, organism unspecified," "urinary tract infection, site not specified," "infection and inflammation, internal prosthetic device," "cellulitis and abscess of leg," and "postoperative infection." CONCLUSIONS: In this predominantly urban, African American diabetes patient population, potentially preventable hospitalizations involving diseases such as congestive heart failure and diabetes occur with high frequency. Better understanding of the risk factors underlying these hospitalizations--particularly those involving modifiable metabolic variables--requires further investigation.

Natural genetic variation caused by transposable elements in humans.

Transposons and transposon-like repetitive elements collectively occupy 44% of the human genome sequence. In an effort to measure the levels of genetic variation that are caused by human transposons, we have developed a new method to broadly detect transposon insertion polymorphisms of all kinds in humans. We began by identifying 606,093 insertion and deletion (indel) polymorphisms in the genomes of diverse humans. We then screened these polymorphisms to detect indels that were caused by de novo transposon insertions. Our method was highly efficient and led to the identification of 605 nonredundant transposon insertion polymorphisms in 36 diverse humans. We estimate that this represents 25-35% of approximately 2075 common transposon polymorphisms in human populations. Because we identified all transposon insertion polymorphisms with a single method, we could evaluate the relative levels of variation that were caused by each transposon class. The average human in our study was estimated to harbor 1283 Alu insertion polymorphisms, 180 L1 polymorphisms, 56 SVA polymorphisms, and 17 polymorphisms related to other forms of mobilized DNA. Overall, our study provides significant steps toward (i) measuring the genetic variation that is caused by transposon insertions in humans and (ii) identifying the transposon copies that produce this variation.

Management of insulin therapy in urban diabetes patients is facilitated by use of an intelligent dosing system.

The Intelligent Dosing System (IDS, Dimensional Dosing Systems, Inc., Wexford, PA) is a software suite that incorporates patient-specific, dose-response data in a mathematical model, and then calculates the new dose of agent needed to achieve the next desired therapeutic goal. We evaluated use of the IDS for titrating insulin therapy. The IDS was placed on handheld platforms and provided to practitioners to use in adjusting total daily insulin dose. Fasting glucose, random glucose, and hemoglobin A1c were used as markers against which insulin could be adjusted. Values of markers expected at the next follow-up visit, as predicted by the model, were compared with levels actually observed. For 264 patients, 334 paired visits were analyzed. Average age was 54 years, diabetes' duration was 10 years, and body mass index was 33.2 kg/m(2); 57% were female, 88% were African American, and 92% had type 2 diabetes. The correlation between IDS suggested and actual prescribed total daily dose was high (r = 0.99), suggesting good acceptability of the IDS by practitioners. Significant decreases in fasting glucose, random glucose, and hemoglobin A1c levels were seen (all P < 0.0001). No significant difference between average expected and observed follow-up fasting glucose values was found (145 vs. 149 mg/dL, P = 0.42), and correlation was high (r = 0.79). Mean observed random glucose value at follow-up was comparable to the IDS predicted level (167 vs. 168 mg/dL, P = 0.97), and correlation was high (r = 0.73). Observed follow-up hemoglobin A1c was higher than the value expected (7.9% vs. 7.4%, P < 0.0055), but correlation was good (r = 0.70). These analyses suggest the IDS is a useful adjunct for decisions regarding insulin therapy even when using a variety of markers of glucose control, and can be used by practitioners to assist in attainment of glycemic goals.

Single nucleotide polymorphisms (SNPs) that map to gaps in the human SNP map.

An international effort is underway to generate a comprehensive haplotype map (HapMap) of the human genome represented by an estimated 300,000 to 1 million 'tag' single nucleotide polymorphisms (SNPs). Our analysis indicates that the current human SNP map is not sufficiently dense to support the HapMap project. For example, 24.6% of the genome currently lacks SNPs at the minimal density and spacing that would be required to construct even a conservative tag SNP map containing 300,000 SNPs. In an effort to improve the human SNP map, we identified 140,696 additional SNP candidates using a new bioinformatics pipeline. Over 51,000 of these SNPs mapped to the largest gaps in the human SNP map, leading to significant improvements in these regions. Our SNPs will be immediately useful for the HapMap project, and will allow for the inclusion of many additional genomic intervals in the final HapMap. Nevertheless, our results also indicate that additional SNP discovery projects will be required both to define the haplotype architecture of the human genome and to construct comprehensive tag SNP maps that will be useful for genetic linkage studies in humans.

Functional genomics reveals relationships between the retrovirus-like Ty1 element and its host Saccharomyces cerevisiae.

Retroviruses and their relatives, the long terminal repeat (LTR) retrotransposons, carry out complex life cycles within the cells of their hosts. We have exploited a collection of gene deletion mutants developed by the Saccharomyces Genome Deletion Project to perform a functional genomics screen for host factors that influence the retrovirus-like Ty1 element in yeast. A total of 101 genes that presumably influence many different aspects of the Ty1 retrotransposition cycle were identified from our analysis of 4483 homozygous diploid deletion strains. Of the 101 identified mutants, 46 had significantly altered levels of Ty1 cDNA, whereas the remaining 55 mutants had normal levels of Ty1 cDNA. Thus, approximately half of the mutants apparently affected the early stages of retrotransposition leading up to the assembly of virus-like particles and cDNA replication, whereas the remaining half affected steps that occur after cDNA replication. Although most of the mutants retained the ability to target Ty1 integration to tRNA genes, 2 mutants had reduced levels of tRNA gene targeting. Over 25% of the gene products identified in this study were conserved in other organisms, suggesting that this collection of host factors can serve as a starting point for identifying host factors that influence LTR retroelements and retroviruses in other organisms. Overall, our data indicate that Ty1 requires a large number of cellular host factors to complete its retrotransposition cycle efficiently.