Abstract

BACKGROUND: Data from 3 institutions revealed an abrupt increase in HeartMate II (Thoratec) pumpthrombosis starting in 2011, associated with 48% mortality at 6 months without transplantation or pumpexchange. We sought to discover if the increase occurred nationwide in Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) data, and if so (1) determine if acceleratedrisk continued, (2) identify predictors, (3) investigate institutional variability, and (4) assess mortalityafter pump thrombosis.
METHODS: From April 2008 to June 2014, 11,123 HeartMate II devices were implanted at 146institutions. Machine learning, non-parametric Random Forests for Survival was used to explore risk-adjusted thrombosis based on 87 pre-implant and implant variables, including implant date.
RESULTS: A total of 995 pumps thrombosed, with risk peaking within weeks of implant. The risk-adjusted increase in pump thrombosis began in 2010, reached a maximum in 2012, and then plateaued at a level that was 3.3-times higher than pre-2010. Pump exchange, younger age, and larger body massindex were important predictors, and institutional variability was largely explained by implant date,patient profile, and duration of support. The probability of death within 3 months after pump thrombosis was 24%.
CONCLUSIONS: Accelerated risk of HeartMate II thrombosis was confirmed by Interagency Registryfor Mechanically Assisted Circulatory Support data, with risk subsequently leveling at a risk-adjusted rate higher than observed pre-2010. This elevated thrombosis risk emphasizes the need for improved mechanical circulatory support systems and post-market surveillance of adverse events.Clinicians cognizant of these new data should incorporate them into their and their patients’expectations and understanding of risks relative to those of transplantation and continued medicaltherapy.

KEYWORDS: HeartMate II;pump thrombosis;INTERMACS;mortality;regulatory;left ventricular assist device (LVAD)

Abstract

Semantic Web technologies offer the potential to revolutionize management of health care data by increasing its interoperability and reusability while reducing the need for redundant data collection and storage. From 1998 through 2010, Cleveland Clinic sponsored a project designed to explore and develop this potential. The product of this effort, SemanticDB, is a suite of software tools and knowledge resources built to facilitate the collection, storage and use of the diverse data needed to conduct clinical research and health care quality reporting.

SemanticDB consists of three main components: 1) a content repository driven by a meta-model that facilitates collection and integration of data in an XML format and automatically converts the data to RDF; 2) an inference-mediated, natural language query interface designed to identify patients who meet complex inclusion and exclusion criteria; and 3) a data production pipeline that uses inference to generate customized views of the repository content for statistical analysis and reporting. Since 2008, this system has been used by the Cleveland Clinic's Heart and Vascular Institute to support numerous clinical investigations, and in 2009 Cleveland Clinic was certified to submit data produced in this manner to national quality monitoring databases sponsored by the Society of Thoracic Surgeons and the American College of Cardiology.

Keywords: clinical data, electronic medical records, inference, ontology, clinical research, quality reporting, RDF, Semantic Web

Abstract

Experience in applying semantic web technology to clinical health care research has revealed the need for a more consistent approach to RDF data production.

Data originating in diverse clinical systems, formats and data models typically must be transformed to hub ontologies, merged, inferenced, selected and finally transformed to consuming applications' ontologies and data formats for productive use. The process is quite repetitive: similar transformations are often performed each day, week or month with only slight variations to the data windows and queries. Ad hoc data production scripts can be created to help automate this process, but they tend to be fragile, inefficient, difficult to maintain and hard to visualize.

This presentation describes a simple framework and ontology for cached RDF data production pipelines, based on RESTful HTTP and SPARQL. These pipelines are easily visualized, and incremental data changes automatically propagate through to consuming applications.

Abstract

Hospitals have come under increasing pressure to report a wide variety of health care process and outcomes measures to various governmental and non-governmental agencies. Typically, the data needed for these reports are captured in separate, unconnected databases designed to fulfill the specific requirements of individual agencies. Although initially expedient, this approach to quality reporting has several drawbacks: 1) it can result in costly redundant data collection if different reports have overlapping requirements; 2) similar data collected in separate databases can produce discrepancies in the same measures reported to different agencies; 3) data targeting specific quality reports often are difficult to use for other purposes, such as clinical research; and 4) the particular data collected are subject to the changing requirements of individual agencies, making multi-year comparisons difficult.

To overcome these problems, Cleveland Clinic's Heart and Vascular Institute has adopted a semantic approach to quality reporting, which uses ontology and rule-based inference to derive specific reported values from core clinical facts collected in a common RDF data repository. We use SemanticDB to collect core clinical data for individual patients through feeds from other clinical systems, and manual, forms-based data entry. These core patient medical data are stored in named RDF graphs, which can be expanded through a combination of forward and backward inference to include new classes useful for generating specific quality reports. These reports are generated with an application built on the Cyc ontology and reasoner that align logical and physical schemas, formulate appropriate SPARQL queries of the expanded graphs, and transform results to the required format. With this approach, core clinical data can be collected in a highly reusable form with little or no redundancy. In addition, differences or changes in reporting requirements can often be handled by modifications of the ontologies and rules guiding the inferential expansion of patient graphs and SPARQL queries rather than the actual data collected.

Abstract

Medical records come in many forms including enterprise-wide electronic health records mandated by the 2009 federal stimulus package, lab and clinic-based data management systems, administrative databases for scheduling, billing and inventory control, and secondary databases for clinical research and quality reporting. Making meaningful use of these disparate data sources to improve the cost and quality of patient care turns on our ability to accurately model their complex semantics and exploit these semantics through machine processing.

Since 2003, Cleveland Clinic has applied semantic technologies to making use of diverse data derived from the treatment of over 200,000 patients with cardiovascular disease over the past 30 years. This presentation outlines the practical issues and challenges faced, and describes how some of these issues have been overcome while others remain unsolved. Topics covered include:

• domain ontology development and utilization
• linking domain ontologies to specific instance data models
• data definition, validation and aggregation
• data utilization for health care quality analysis and reporting

Abstract

Medical records come in many forms including enterprise-wide electronic health records mandated by the 2009 federal stimulus package, lab and clinic-based data management systems, administrative databases for scheduling, billing and inventory control, and secondary databases for clinical research and quality reporting. Making meaningful use of these disparate data sources to improve the cost and quality of patient care turns on our ability to accurately model their complex semantics and exploit these semantics through machine processing.

Since 2003, Cleveland Clinic has applied semantic technologies to making use of diverse data derived from the treatment of over 200,000 patients with cardiovascular disease over the past 30 years. This presentation outlines the practical issues and challenges faced, and describes how some of these issues have been overcome while others remain unsolved. Topics covered include:

• domain ontology development and utilization
• linking domain ontologies to specific instance data models
• data definition, validation and aggregation
• data utilization for health care quality analysis and reporting

Abstract

The ability to identify patient populations for clinical research according to specific criteria is a critical component of clinical informatics. Traditionally, database systems that leverage relational databases are used to store and manage the data targeted for population identification. The process of starting with an abstract research question (posed in prose) and distilling it into the specific query plans and corresponding data structures necessary to answer such research questions is often significantly tedious and primarily driven by idiosyncrasies of how the data is stored rather than the underlying science that guides the intuition of the researcher. Contemporary clinical data management systems have yet to move beyond this approach.

Cleveland Clinic and Cycorp Inc. have developed a state-of-the-art query interface that builds query fragments through natural language-driven interactions with an investigator. This interface targets a clinical research repository comprised of patient record content expressed in an RDF dataset that conforms to a patient record OWL ontology and is queried over the SPARQL service protocol. It leverages the Cyc common sense ontology, Cyc's natural language processing capabilities, formalized mappings from the common sense ontology into the patient record ontology, and the availability of a SPARQL service for querying the patient population.

In this session, we describe the specifics of this framework, enumerate its strengths and weaknesses, and derive insights into the opportunities and challenges of adopting Semantic Web Technologies for the purpose of managing and identifying patient populations.

Abstract

Hospitals have come under increasing pressure to report a wide variety of health care process and outcomes measures to various governmental and non-governmental agencies. Typically, the data needed for these reports are captured in separate, unconnected databases designed to fulfill the specific requirements of individual agencies. Although initially expedient, this approach to quality reporting has several drawbacks: 1) it can result in costly redundant data collection if different reports have overlapping requirements; 2) similar data collected in separate databases can produce discrepancies in the same measures reported to different agencies; 3) data targeting specific quality reports often are difficult to use for other purposes, such as clinical research; and 4) the particular data collected are subject to the changing requirements of individual agencies, making multi-year comparisons difficult.

To overcome these problems, Cleveland Clinic's Heart and Vascular Institute has adopted a semantic approach to quality reporting, which uses ontology and rule-based inference to derive specific reported values from core clinical facts collected in a common RDF data repository. We use SemanticDB to collect core clinical data for individual patients through feeds from other clinical systems, and manual, forms-based data entry. These core patient medical data are stored in named RDF graphs, which can be expanded through a combination of forward and backward inference to include new classes useful for generating specific quality reports. These reports are generated with an application built on the Cyc ontology and reasoner that align logical and physical schemas, formulate appropriate SPARQL queries of the expanded graphs, and transform results to the required format. With this approach, core clinical data can be collected in a highly reusable form with little or no redundancy. In addition, differences or changes in reporting requirements can often be handled by modifications of the ontologies and rules guiding the inferential expansion of patient graphs and SPARQL queries rather than the actual data collected.

Abstract

To understand the effects of European contact on the organization, size, and mobility of Pueblo populations in the Southwest requires detailed knowledge of the occupational histories of the aggregated settlements that typify the late prehistoric and early historic record. Unfortunately, such understanding is generally lacking because the methods used to document occupational histories of settlements tend to either obscure fine- grained temporal distinctions or necessitate costly, and politically objectionable, large- scale excavations. To overcome these difficulties, we use surface expressions to analyze the occupational and population history of San Marcos Pueblo (LA98), an aggregated, late prehistoric site in the Galisteo Basin of New Mexico that persisted to the Pueblo Revolt of 1680. Field methods include detailed mapping of the settlement and systematic surface collections of middens. Frequency seriation, correspondence analysis, and mean ceramic dates of decorated ceramic rims comprise our principal analytic methods and demonstrate that the pueblo was abandoned four times before 1680. Causes of abandonment are discussed. Relative scale measures of population show demographic fluctuations with maximum aggregation during the fifteenth century. Despite demographic pulses, the pueblo remained vital until the terminal abandonment.

Abstract

OBJECTIVE: To (1) characterize atrial fibrillation complicating lung cancer resection, (2) evaluate its temporal relationship to other postoperative complications, and (3) assess its economics.

METHODS: From January 1998 to August 2002, 604 patients underwent anatomic lung cancer resection. Atrial fibrillation prevalence, onset, and temporal associations with other postoperative complications were determined. Propensity matching was used to assess economics.

RESULTS: Atrial fibrillation occurred in 113 patients (19%), peaking on postoperative day 2. Older age, male gender, heart failure, clamshell incision, and right pneumonectomy were risk factors (P < .01). Although atrial fibrillation was solitary in 75 patients (66%), other postoperative complications occurred in 38. Respiratory and infectious complications were temporally linked with atrial fibrillation onset. In 91 propensity-matched pairs, patients developing atrial fibrillation had more other postoperative complications (30% vs. 9%, P < .0004), had longer postoperative stays (median 8 vs 5 days, P < .0001), incurred higher costs (cost ratio 1.8, 68% confidence limits 1.6-2.1), and had higher in-hospital mortality (8% vs 0%, P = .01). Even when atrial fibrillation was a solitary complication, hospital stay was longer (median 7 vs 5 days, P < .0001), and cost was higher (cost ratio 1.5, 68% confidence limits 1.2-1.6).

CONCLUSION: Atrial fibrillation occurs in 1 in 5 patients after lung cancer resection, with peak onset on postoperative day 2. Risk factors are both patient and procedure related, and atrial fibrillation may herald other serious complications. Although often solitary, atrial fibrillation is associated with longer hospital stay and higher cost. It therefore requires prompt treatment and should stimulate investigation for other problems.

Abstract

Ceramic cooking pots throughout the world vary in exterior surface treatment from smooth to roughly textured. An intriguing example of this variation occurred in the Puebloan region of the southwestern United States where cooking pots changed from scraped plain to highly textured, corrugated vessels between the seventh and eleventh centuries AD, and then reverted back to plain-surfaced by the fifteenth century. To investigate potential cost and performance differences between plain and corrugated cooking pots, a set of controlled experiments were performed, which document manufacturing costs, cooking effectiveness, and vessel durability. These experiments indicate that while corrugation may have increased manufacturing costs, neck corrugations improved vessel handling, upper body corrugations yielded greater control over cooking, and basal corrugations extended vessel use-life. Discerning the explanatory significance of these results for cooking pot change in the Southwest and elsewhere requires additional data on the contexts in which these pots were made and used.

Keywords

reverse engineering; experimental archaeology; cooking pots; Puebloan southwest

Abstract

Archaeological research has documented the broad outlines of the development of corrugated utility pottery from plain and neck-banded antecedents. However, a reliance on typological descriptions has obscured the technological details of this development. An attribute analysis of six well-dated utility-ware assemblages from southwestern Colorado indicates that corrugation appeared first in this region during the eighth century A. D. as wide, non-overlapping coils left unsmoothed around jar necks. During the ninth and tenth centuries, the variety and frequency of neck-banding increased with the introduction of narrower coils, overlapping of adjacent coils, and incising and indenting of coil surfaces. By the early eleventh century, one recent neck-banding variant, narrow, substantially overlapped and indented coils, replaced almost all others, and was extended over the entire exterior surface of jars for the first time. Performance benefits of this full-body, indented corrugation may explain its rapid adoption in southwestern Colorado.

Abstract

OBJECTIVES: Surgical staging and resection of lung cancer may be done as 1 operation (combined) or 2 (staged). This study evaluates the safety and efficiency of these treatment strategies.

METHODS: From 1998 to July 2001, 343 patients underwent bronchoscopy, mediastinoscopy, and thoracotomy without induction chemoradiotherapy by 3 surgeons. Fifty-seven patients were staged and 286 combined. Staged patients had higher clinical stage (P < .001). Propensity-matched groups were compared to adjust for this and other differences. Factors associated with safety and efficiency were identified by propensity-adjusted multivariable analysis.

RESULTS: Mortality and morbidity were similar for both strategies. Efficiency, measured by shorter operative time (1.2 hours) and lower cost (25%), was better for combined strategy (P < .001). Hospital stay was similar, but revenue was 12% higher for the staged strategy (P < .001). In propensity-matched comparisons excluding surgeon, results were similar to the above. Comparisons including surgeon demonstrated similar cost and revenue for both strategies. Increased mortality and morbidity were associated only with patient and tumor characteristics: male gender, worsening Eastern Cooperative Oncology Group performance status, and increasing pathological node classification. All measures of efficiency worsened with increasing pathological classifications. Staged strategy was associated with increased operative time and revenue, while one surgeon and patient smoking history were associated with increased hospital stay and costs.

CONCLUSIONS: The combined strategy provides efficient, safe health care for clinically operable lung cancer patients, but it may not be as financially rewarding as the staged strategy. Treatment strategy is only 1 of many determinants of efficiency.

Abstract

The scale, intensity, and character of interaction among Pueblo people during the 13th century A.D. likely played an important role in the processes and events leading to the abandonment of the Northern San Juan region in the 1280s. Characteristics of pottery production and distribution in the Sand Canyon locality provide one means of investigating these interactions. Variation among Pueblo III settlements in the use of temper and available raw clay sources, and the distribution of pottery production tools demonstrate the existence of at least two production areas within the locality. The nature of the boundary between these production areas indicates a complex pattern of settlement and community interaction that challenges models based on settlement proximity. Further, an almost complete lack of extra-regional pottery at Pueblo III settlements suggests that the Northern San Juan region may have been economically isolated from other regions inhabited by Pueblo people.

Abstract

Corrugated pottery is a unique utility ware made by leaving construction coils unobliterated, and manipulating these exposed coils to produce a rough exterior surface. Ancestral Puebloan populations in the American Southwest made this pottery in various forms between AD 650 and 1450. Although archaeologists have tried to explain corrugation for over 100 years, none of these explanations has proven satisfactory. I employ an evolutionary approach in an attempt to explain the rise and fall of corrugated pottery.

Using analyses of ancient pottery, experiments with modern replicas, and syntheses of published data, I document that corrugation developed gradually from plain ware through a series of innovations in the size of coils and how they were applied and manipulated. Initially, corrugation appeared in the southern Mogollon region as wide, filleted, and unindented bands on jar necks, and spread north into the Anasazi areas of the Colorado Plateau. The development of overlapped and indented neck bands during the ninth and tenth centuries ultimately led to the use of these techniques over the entire exterior surface during the eleventh century producing full-body, indented corrugation. Full-body corrugation spread rapidly across a large area of the Colorado Plateau, and remained dominant until the fifteenth century when a return to plain pottery occurred. The development of corrugation coincided with an increase in the use of these vessels to cook food.

Although most of the innovations appeared first as decorative elaborations, some also affected the cost and performance of the vessels. The use of narrower and overlapped coils increased the time required to form a vessel over the earlier plain vessels. However, the extension of the textured surface to the upper body and basal parts of vessels significantly improved control over cooking and the use-life of vessels respectively.

Explanations of corrugation emphasize how the environment with which corrugated pottery interacted changed through time, and how these changes affected innovation rates, and the selection or drift of particular corrugation variants. In formulating explanations, I address four problems: the advent of neck banding, experimentation with neck banding, the rapid adoption of full-body corrugation, and the return to plain-surfaced pottery.

Abstract

Documenting variation and change in the use of particular plants and plant parts for fuel in ancient households contributes to an understanding of settlement location, local and regional abandonments and resource depletion. Chemical analysis of ash surviving in hearths and other thermal features to determine the kinds and relative amounts of fuels consumed may be less biased by formation processes than the macrofossil record currently employed to document ancient fuel use. Our studies indicate that it is possible to distinguish, chemically, ash of common fuel taxa and tissue types (bark, wood, etc.) in both modern and ancient samples of fuels found in the northern Colorado Plateau region of the American Southwest. However, the chemical signatures of the ancient and modern material of the same taxon differ, indicating possible alteration by post-depositional processes. Although multiple regression performs well in determining the relative contributions of different fuels to modern ash mixtures, possible post-depositional alterations and incomplete characterization of the range of within-taxon variation, currently limit our confidence in applications of this approach to ancient ash from Pueblo III settlements in south-west Colorado.

Abstract

The construction of burrows and movement of sediment by pocket gophers alter archaeological deposits by causing vertical size-sorting of artifacts, destruction of fragile artifacts, disruption of sedimentary structures, and organic enrichment of the subsurface. To evaluate the long-term effects of exposure to burrowing, a simulation was developed based on quantitative information on pocket gopher burrows and rates of sediment movement. Simulation results indicate the development of a distinct stone zone composed predominantly of particles greater than 6 cm after 4000-5000 years, and a logarithmic pattern to the rate of strata disruption. The patterns produced by the simulation compare well with patterns exhibited by actual archaeological deposits belonging to California's Milling Stone Horizon. These results suggest that current notions concerning the Milling Stone Horizon and other aspects of California prehistory may require revision, and that more emphasis must be placed on formation process research in such settings.

Abstract

For the past decade, several archaeologists have advocated the development of middle-range theory as a way to give objective meaning to the archaeological record (e.g., Bettinger 1987; Binford 1977, 1983b; Thomas 1983, 1989; Torrence 1986). They argue that we must translate the static archaeological record into behaviorally dynamic terms by documenting causal linkages between relevant behaviors and their static material by-products. This is accomplished, they argue, by making observations today that establish signature patterns allowing the unambiguous recognition of particular dynamics from their static by-products, and inferring past dynamics from identification of signature patterns in the archaeological record. Further, it has been emphasized that the operations and products of middle-range theory must remain logically independent of the general theory we use to explain the past to avoid automatically confirming our ideas about the past through a tautology. This approach to middle-range research is flawed in two major respects. First, the justification of inferences relies on the establishment of universal behavioral laws and unambiguous signature patterns to validate the use of uniformitarian assumptions, neither of which can be accomplished in the manner proposed. Second, the tautological relationship between description and explanation is not only an unavoidable, but also a necessary aspect of science. Solutions to these problems lie in using the physical characteristics of the archaeological record itself as our source of knowledge about the past rather than translating the record into untestable behavioral reconstructions.

Abstract

The Density Dependent Method consists of a coordinated group of techniques that are designed to measure large- and small-scale spatial variability contained in the archaeological record. This is accomplished with a series of nested sampling schemes that focuses on the individual artifact or feature in its environmental context. The primary objective of this paper os to present the details of the density-dependent method as it was developed to deal with the Archaic hunter-gatherer record in the northern Southwest. The method facilitiates quantiative analyses of the relationships between artifact spatial distributions and the local and regional resource framework as defined by a landform environment classificaiton.