Using Machine Learning to Identify Organ System Specific Limitations to Exercise via Cardiopulmonary Exercise Testing

Cardiopulmonary Exer cise Testing (CPET) is a unique physiologic medical test used to evaluate human response to progressive maximal exercise stress. Depending on the degree and type of deviation from the normal physiologic response, CPET can help identify a patient's specific limitations to exercise to guide clinical care without the need for other expensive and invasive diagnostic tests. However, given the amount and complexity of data obtained from CPET, interpretation and visualization of test results is challenging. CPET data currently require dedicated training and significant experience for proper clinician interpretation. To make CPET more accessible to clinicians, we investigated a simplified data interpretation and visualization tool using machine learning algorithms. The visualization shows three types of limitations (cardiac, pulmonary and others); values are defined based on the results of three independent random forest classifiers. To display the models' scores and make them interpretable to the clinicians, an interactive dashboard with the scores and interpretability plots was developed. This machine learning platform has the potential to augment existing diagnostic procedures and provide a tool to make CPET more accessible to clinicians.

Research Reproducibility in Longitudinal Multi-Center Studies Using Data from Electronic Health Records

A fundamental premise of scientific research is that it should be reproducible. However, the specific requirements for reproducibility of research using electronic health record (EHR) data have not been sufficiently articulated. There is no guidance for researchers about how to assess a given project and identify provisions for reproducibility. We analyze three different clinical research initiatives that use EHR data in order to define a set of requirements to reproduce the research using the original or other datasets. We identify specific project features that drive these requirements. The resulting framework will support the much-needed discussion of strategies to ensure the reproducibility of research that uses data from EHRs.

Health data use, stewardship, and governance: ongoing gaps and challenges: a report from AMIA's 2012 Health Policy Meeting

Large amounts of personal health data are being collected and made available through existing and emerging technological media and tools. While use of these data has significant potential to facilitate research, improve quality of care for individuals and populations, and reduce healthcare costs, many policy-related issues must be addressed before their full value can be realized. These include the need for widely agreed-on data stewardship principles and effective approaches to reduce or eliminate data silos and protect patient privacy. AMIA's 2012 Health Policy Meeting brought together healthcare academics, policy makers, and system stakeholders (including representatives of patient groups) to consider these topics and formulate recommendations. A review of a set of Proposed Principles of Health Data Use led to a set of findings and recommendations, including the assertions that the use of health data should be viewed as a public good and that achieving the broad benefits of this use will require understanding and support from patients.

Discussion of "Biomedical informatics: we are what we publish"

This article is part of a For-Discussion-Section of Methods of Information in Medicine about the paper "Biomedical Informatics: We Are What We Publish", written by Peter L. Elkin, Steven H. Brown, and Graham Wright. It is introduced by an editorial. This article contains the combined commentaries invited to independently comment on the Elkin et al. paper. In subsequent issues the discussion can continue through letters to the editor.

Seamless care: what is it; what is its value; what does it require; when might we get it?

The vision of a seamless care environment enabled through the use of health information technology opens the door to exciting possibilities. The first is the creation of a patient-centric Electronic Health Record that would contain all health-related data about a patient from all sources. That EHR would contribute to highest quality, safe, efficient, and low cost care. That same data would be used for multiple purposes, and, by integrating all requirements at the beginning, would meet all those multiple needs. Key to obtaining seamless care is the creation of a global master registry of data elements with attributes that would provide not only interoperability throughout seamless health care but also permit the coupling of knowledge with data. To accomplish this lofty vision will require making critical decisions, staying to the course, eliminating siloed approaches, and adopting universal, cooperative solutions to the use of HIT.

Realizing the potential of healthcare information technology to enhance global health

For much of the world, truly productive and functional Electronic Health Record Systems (EHRs) remain an elusive goal of the future. Opportunities abound from the visibility provided by the availability of Health Information Technology funding in the U.S. and other countries of the world. Now is the time to seize the initiative to move from the past to the future to design HIT systems that meet the specific needs of each nation of the world in a way that is obtainable and affordable, and that provides an immediate return on investment. We need to move from an electronic system based on the paper-system to an empowering system based on available technology. We need to recognize that the EHR is not just for data storage but needs to become an intelligent, active partner with the healthcare provider and the patient to enhance health. This paper describes the current state of EHRs and addresses challenges for moving into the future.

Data mining methods find demographic predictors of preterm birth

BACKGROUND

Preterm births in the United States increased from 11.0% to 11.4% between 1996 and 1997; they continue to be a complex healthcare problem in the United States.

OBJECTIVE

The objective of this research was to compare traditional statistical methods with emerging new methods called data mining or knowledge discovery in databases in identifying accurate predictors of preterm births.

METHOD

An ethnically diverse sample (N = 19,970) of pregnant women provided data (1,622 variables) for new methods of analysis. Preterm birth predictors were evaluated using traditional statistical and newer data mining analyses.

RESULTS

Seven demographic variables (maternal age and binary coding for county of residence, education, marital status, payer source, race, and religion) yielded a .72 area under the curve using Receiving Operating Characteristic curves to test predictive accuracy. The addition of hundreds of other variables added only a .03 to the area under the curve.

CONCLUSION

Similar results across data mining methods suggest that results are data-driven and not method-dependent, and that demographic variables offer a small set of parsimonious variables with reasonable accuracy in predicting preterm birth outcomes in a racially diverse population.

A national agenda for public health informatics

The American Medical Informatics Association 2001 Spring Congress brought together the public health and informatics communities to develop a national agenda for public health informatics. Discussions on funding and governance; architecture and infrastructure; standards and vocabulary; research, evaluation, and best practices; privacy, confidentiality, and security; and training and workforce resulted in 74 recommendations with two key themes: (1) all stakeholders need to be engaged in coordinated activities related to public health information architecture, standards, confidentiality, best practices, and research and (2) informatics training is needed throughout the public health workforce. Implementation of this consensus agenda will help promote progress in the application of information technology to improve public health.

Characteristics of adult primary care patients as predictors of future health services charges

BACKGROUND

Utilization risk assessment is potentially useful for allocation of health care resources, but precise measurement is difficult.

OBJECTIVE

Test the hypotheses that health-related quality of life (HRQOL), severity of illness, and diagnoses at a single primary care visit are comparable case-mix predictors of future 1-year charges in all clinical settings within a large health system, and that these predictors are more accurate in combination than alone.

RESEARCH DESIGN

Longitudinal observational study in which subjects' characteristics were measured at baseline, and their outpatient clinic visits and charges and their inpatient hospital days and charges were tracked for 1 year.

SUBJECTS

Adult primary care patients.

MEASURES

Duke Health Profile for HRQOL, Duke Severity of Illness Checklist for severity of illness, and Johns Hopkins Ambulatory Care Groups for diagnostic groups classification.

RESULTS

Of 1,202 patients, 84.4% had follow up in the primary care clinic, 63.2% in subspecialty clinics, 14.8% in the emergency room, and 9.6% in the hospital. Of $6,290,775 total charges, $779,037 (12.2%) was for follow-up primary care. The highest accuracy was found for predicting primary care charges, where R2 for predictors ranged from 0.083 for medical record auditor-reported severity of illness to 0.107 for HRQOL. When predictors were combined, the highest R2 of 0.125 was found for the combination of HRQOL and diagnostic groups.

CONCLUSIONS

Baseline HRQOL, severity of illness, and diagnoses were comparable predictors of 1-year health services charges in all clinical sites but most predictive for primary care charges, and were more accurate in combination than alone.

How the past teaches the future: ACMI distinguished lecture

More than 30 years of experience in developing a computer-based patient record system, The Medical Record (TMR), in multiple settings, in multiple specialty groups, and at multiple sites has taught us many lessons. Lessons related to computer-based patient records include the importance of a data model in which input, storage, and planned use are independent; separation of patient-specific data from metadata; a modular design to localize the program code that deals with a set of data; redundant storage to optimize tasks and response time; and integration of decision support into work process. Lessons related to medical informatics include the importance of a clinical-technical partnership, control of tools at the leading edge, and rapid prototyping in the real world. Finally, changes in technology move the challenges but do not eliminate them.

Standards for the electronic health record, emerging from health care's Tower of Babel

This paper considers the standardization of an Electronic Health Record (EHR). Relations between several distinct medical datasets and information systems are mapped in order to derive a more precise definition of the EHR. Two international efforts to establish standards for the EHR are presented and critiqued. Strategies for standardizing the EHR are analyzed and recommendations are provided for approaching the standardization process.

Toward vocabulary domain specifications for health level 7-coded data elements

The "vocabulary problem" has long plagued the developers, implementers, and users of computer-based systems. The authors review selected activities of the Health Level 7 (HL7) Vocabulary Technical Committee that are related to vocabulary domain specification for HL7 coded data elements. These activities include: 1) the development of two sets of principles to provide guidance to terminology stakeholders, including organizations seeking to deploy HL7-compliant systems, terminology developers, and terminology integrators; 2) the completion of a survey of terminology developers; 3) the development of a process for HL7 registration of terminologies; and 4) the maintenance of vocabulary domain specification tables. As background, vocabulary domain specification is defined and the relationship between the HL7 Reference Information Model and vocabulary domain specification is described. The activities of the Vocabulary Technical Committee complement the efforts of terminology developers and other stakeholders. These activities are aimed at realizing semantic interoperability in the context of the HL7 Message Development Framework, so that information exchange and use among disparate systems can occur for the delivery and management of direct clinical care as well as for purposes such as clinical research, outcome research, and population health management.

Object technology: raising the standards for healthcare information systems

Netscape and the public Internet have accelerated the acceptance of many different open "Internet standards". Through wide acceptance of its browser, Netscape gave a boost to the Java programming language helping it become truly platform independent. Objects written in Java are ideal building blocks for application components. CORBA gives such objects the ability to communicate and operate over networks. Applications built with these distributed objects become the services in an Internet-wide healthcare framework. The convergence of object technologies has raised the standards for modern healthcare information systems. To illustrate the relationship among such technologies, this paper presents an architecture for a Universal Healthcare Information System (UHIS) in terms of its web, Java and CORBA components.

Managing healthcare: a view of tomorrow

This paper presents a vision of the future in which standards exist at all levels necessary to accomplish true interoperability. The infrastructure has been established to support connectivity among all healthcare-related institutions as well as the population at large. Provider and patient care integrated in the process of an individual's care.

Data mining issues for improved birth outcomes

Issues obstructing progress in data mining for improved health outcomes include data quality problems, data redundancy, data inconsistency, repeated measures, temporal (time-contextual) measures, and data volume. Related issues involve theoretical and technical problems involving uncertainty management, missing data and missing values, and matching appropriate data mining techniques to patient data sets. Results of data mining research in progress are reported for Duke University's perinatal database that contains nearly a decade of clinical patient data, 71,753 database (patient) records and 4-5000 variables per patient.

A proposal for incorporating health level seven (HL7) vocabulary in the UMLS Metathesaurus

The HL7 Vocabulary Technical Committee (TC) was organized to select and maintain the vocabulary used in HL7 messages. The goal is to make implementations of the Version 3 HL7 Standard more plug-and-play compatible. In order to make the vocabulary readily accessible to the public, HL7 is collaborating with the U.S. National Library of Medicine (NLM) to include HL7 vocabulary in the Unified Medical Language System (UMLS) Metathesaurus. This article describes a proposal for how HL7 data elements and coded values can be represented accurately in the relational tables of the UMLS Metathesaurus.

A comparison of neural network models for the prediction of the cost of care for acute coronary syndrome patients

Acute coronary syndromes have remained the focus of many clinical economic studies due to the increasing prevalence of the disease and the tightening of cost controls. An accurate descriptive cost model for this population would be a valuable tool for clinical researchers. With such a model, the relative importance of different factors upon the total cost of care could be determined through computer simulation. This study explored the use of different neural network architectures in creating a descriptive cost model. This was a difficult problem in that the costs span 3 orders of magnitude but the output variable of the neural network must be restricted to the range 0-1. Models that used logarithmic transformations and multiple modular networks were created and analyzed. It was found that the model with a single network and logarithmic transformation performed significantly better than other more complicated networks.

Duke Case-Mix System (DUMIX) for ambulatory health care

The Duke Case-Mix System (DUMIX), which combines age, gender, patient-reported perceived and physical health status, and provider-reported or auditor-reported severity of illness to classify patients by their risk of high future utilization, explained 17.1% of the variance in future clinic charges and 16.6% of the variance in return visits. When a random half of 413 ambulatory adults were classified into four risk classes by predictive regression coefficients from the other half, there was a stepwise increase in actual future utilization by risk class. The most accurate classification was for Class 4 (highest risk) patients, with a sensitivity of 40.8%, specificity of 82.1%, and likelihood ratio of 2.3. These 23.7% of patients accounted for 44.2% of charges for all patients. When predictive coefficients from this population were used to classify a different group of 206 ambulatory adults, past utilization also increased in stepwise order by case-mix class.

Integration of a computer-based patient record system into the primary care setting

The use of computer-based patient record systems (CPRS) in the primary care setting will increase significantly over the next few years. Real-time, point-of-care use of such systems must provide adequate payback to justify the intrusion into the provider/patient relationship. The authors describe, through the transition of a legacy system into a state-of-the art system, how such a system might be integrated into the primary care setting. Information flow is organized around an event, such as a patient encounter, or around the patient. The Medical Record (TMR) optimizes the provider/computer interaction through the use of protocols and clinical guidelines. Documentation is enhanced through the use of computer-generated progress notes.

Medical data mining: knowledge discovery in a clinical data warehouse

Clinical databases have accumulated large quantities of information about patients and their medical conditions. Relationships and patterns within this data could provide new medical knowledge. Unfortunately, few methodologies have been developed and applied to discover this hidden knowledge. In this study, the techniques of data mining (also known as Knowledge Discovery in Databases) were used to search for relationships in a large clinical database. Specifically, data accumulated on 3,902 obstetrical patients were evaluated for factors potentially contributing to preterm birth using exploratory factor analysis. Three factors were identified by the investigators for further exploration. This paper describes the processes involved in mining a clinical database including data warehousing, data query and cleaning, and data analysis.

Computerized decision support based on a clinical practice guideline improves compliance with care standards

PURPOSE

Clinical guidelines are designed to assist in the management of specific diseases; however, these guidelines are often neglected in the delivery of care. The purpose of this study was to determine whether clinician use of an clinical practice guideline would increase in response to having, at the patient visit, a decision support system based on a practice guideline that generates a customized management protocol for the individual patient using data from the patient's electronic medical record.

SUBJECTS AND METHODS

In a 6-month controlled trial at a primary care clinic, 58 primary care clinicians were randomized to receive either a special encounter form with the computer-generated guideline recommendations or a standard encounter form. The effect of computer-generated advice on clinician behavior was measured as rate of compliance with guideline recommendations. Data from 30 clinicians were analyzed; data from 28 clinicians were excluded because these clinicians did not meet predefined criteria for minimum exposure to diabetic patient care.

RESULTS

Availability of patient management recommendations generated by the decision support system resulted in a two-fold increase in clinician compliance with care guidelines for diabetes mellitus (P = 0.01). Median compliance for the group receiving the recommendations was 32.0% versus 15.6% for the control group.

CONCLUSION

Decision support based on a clinical practice guideline is an effective tool for assisting clinicians in the management of diabetic patients. This decision support system provides a model for how a clinical practice guideline can be integrated into the care process by computer to assist clinicians in managing a specific disease through helping them comply with care standards. Use of decision support systems based on clinical practice guidelines could ultimately improve the quality of medical care.

Implications of the Java language on computer-based patient records

The growth of the utilization of the World Wide Web (WWW) as a medium for the delivery of computer-based patient records (CBPR) has created a new paradigm in which clinical information may be delivered. Until recently the authoring tools and environment for application development on the WWW have been limited to Hyper Text Markup Language (HTML) utilizing common gateway interface scripts. While, at times, this provides an effective medium for the delivery of CBPR, it is a less than optimal solution. The server-centric dynamics and low levels of interactivity do not provide for a robust application which is required in a clinical environment. The emergence of Sun Microsystems' Java language is a solution to the problem. In this paper we examine the Java language and its implications to the CBPR. A quantitative and qualitative assessment was performed. The Java environment is compared to HTML and Telnet CBPR environments. Qualitative comparisons include level of interactivity, server load, client load, ease of use, and application capabilities. Quantitative comparisons include data transfer time delays. The Java language has demonstrated promise for delivering CBPRs.

Documenting data delivery: design, deployment, and decision

Developing and deploying informatics solutions which are useful and acceptable to busy physicians are challenging tasks. We describe the design, deployment, and evaluation process by which the delivery of routine clinical laboratory reports is automated using electronic mail. Data from TMR, an operational computer-based patient record (CPR), are presented to providers using an individualized, modern interface. This system is compared to the existing, paper-based system for delivery of data from the same CPR. Differences between the two systems of data delivery are analyzed, with emphases on 1) electronic documentation of data delivery and receipt, 2) electronic and/or paper documentation of clinical action taken as a result of laboratory reports, 3) timeliness of report availability, 4) costs, 5) workflow compatibility, and 6) physician satisfaction. The new delivery system employs inexpensive, commercially available software applications and entails only trivial changes to the proprietary CPR. Built into the new system are features which allow quantitative measurements of its performance for analysis along with survey-based user satisfaction data. The open systems design is deliberately non-proprietary, inexpensive, and generalizable. Accordingly, it offers practical possibilities for settings in which clinical information systems are just being planned, as well as for those in which such systems are already established.

The status of healthcare standards in the United States

Healthcare standards in the US are produced by six standard developers organizations: ACR/NEMA, ASC X12N, ASTM, HL7, IEEE, and NCPDP. The activities of these groups are coordinated through the ANSI HISPP. While considerable progress has been made in the area of data interchange standards, little progress has been made in the area of vocabulary standards.

Computer-prompted diagnostic codes

BACKGROUND

The purpose of this study was to develop and evaluate a computer system that would translate patient diagnoses noted by a physician into appropriate International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes and maintain a patient-specific up-to-date problem list.

METHODS

The intervention consisted of a computerized list (dictionary) of diagnoses, including practice-specific synonyms and abbreviations, linked to their corresponding ICD-9-CM codes. To record the diagnoses for the office visit before the intervention, physicians used International Classification of Health Problems in Primary Care (ICHPPC-2) codes. After the intervention, physicians used their own words or checked previously identified diagnoses on the computer-generated problem list. The computer then identified the correct ICD-9-CM code. Accuracy of coding was compared before and after the new computerized system was implemented.

RESULTS

Visits in which all diagnoses matched increased from 58% to 76% (P < .001) with use of the computer system. Visits in which no computer diagnoses matched the chart decreased from 22% to 8% (P < .001). Errors of omission declined from 38% to 18% (P < .001). Errors of commission decreased from 19% to 11% (P = .006). Overall accuracy increased from 62% to 82% (P < .001).

CONCLUSIONS

Outpatient medical diagnosis coding can be simplified and accuracy improved by using a computerized dictionary of practice-specific diagnoses and synonyms linked to appropriate ICD-9-CM codes. Such a system provides a computer-generated problem list that accurately reflects the chart and assists with prompted coding on subsequent visits.

Converting a legacy system database into relational format to enhance query efficiency

The analysis of clinical data collected over time can provide important insight into the health care process. Unfortunately, much of the electronic clinical data that exists today is stored in legacy systems, making it difficult to access and share the information. An approach is needed to improve the accessibility of electronic data stored in legacy system databases. In this study, a legacy database is converted into a relational format in the personal computer environment. The impact of such a conversion on query performance is evaluated, and issues that need to be considered when converting a legacy system database are identified.

Feasibility and potential clinical usefulness of a computerized severity of illness measure

OBJECTIVE

To assess the feasibility and potential clinical usefulness of the computerized Duke Severity of Illness Checklist (DUSOI).

DESIGN

Cross-sectional study of patients whose severity of illness was measured with the DUSOI. Providers assessed the clinical usefulness of the DUSOI and recorded the length of time required for rating severity. Auditors rated severity using progress note information. Demographic and financial data from clinic records were also obtained.

SETTING

University-based family practice clinic with 64,621 annual visits.

PATIENTS

Convenience sample of ambulatory patients.

MAIN OUTCOME MEASURES

Clinical usefulness and time required to rate severity.

RESULTS

For 117 patients (63.3% female; mean age, 46.3 years), the mean charge was $105.38, the mean number of health problems was 2.0, the mean overall provider DUSOI score was 33.7, and the mean auditor DUSOI score was 34.0 (scale = 0 to 100). There was excellent agreement between provider and auditor DUSOI scores (intraclass correlation coefficient, .77). Providers required 1.1 minutes to record severity; the principal auditor required 1.6 minutes. Providers found the DUSOI potentially useful in 30.3% of patients. Usefulness was greater in women (38.2% of women vs 18.2% of men), older patients (mean age, 54.5 years in useful group vs 41.9 in nonuseful group), and sicker patients (mean DUSOI score, 55.1 vs 25.9). The DUSOI was more clinically useful in patients with health problems such as type II diabetes mellitus (75.0%) than in those with problems such as tobacco use (25.0%). Higher charges correlated with a higher number of health problems and with female gender but not with severity scores.

CONCLUSIONS

The computerized DUSOI is feasible for all patients and is potentially useful for women, older, and sicker patients.

The medical record: a comprehensive computer system for the family physician

BACKGROUND

Despite the early excitement regarding the possible uses of computers in medical care in the 1980s, the computer has not had much effect on routine outpatient medicine except for billing and accounting.

METHODS

An emerging comprehensive ambulatory care computer system, The Medical Record (TMR), is used extensively in a large family practice, the Duke Family Medicine Center. TMR is the central system for accounting, appointments, billing, and reporting of laboratory results, radiographic findings, and medications. TMR also records problem lists and generates prompts to the clinicians for needed health maintenance, laboratory tests, and reminder letters. The most innovative function of TMR is the computerized obstetric patient record, which can be accessed from multiple sites. Cost savings compared with a manual system were found to be in excess of $7 per patient visit or approximately $500,000 per year for the Duke Family Medicine Center.

RESULTS AND CONCLUSIONS

A comprehensive computer system in a large family practice is cost effective and facilitates better patient care through improved access to patient data.

The role of standards in creating a health information infrastructure

In the belief that the existence of a variety of standards is an absolute necessity for health care professional workstations to work, this paper provides a detailed overview of the standards efforts of a number of groups. According to the International Standards Organization (ISO) Reference Model, workstations require a full level of standards from the physical level through and beyond the applications level. Rapidly changing technology challenges acceptance of standards at the lower levels. Current recommendations include fiberoptic media using certain protocols. Other standards in these lower levels also have support. At the applications level, data messaging standards are being developed by six groups. The consensus standards body for the United States is coordinating the efforts of these groups in order to produce a harmonized effort, and is coordinating the effort with Europe for an international effort. Work on the development of the full set of standards necessary for workstation implementations is lagging. Accelerating the process is mandatory if we are to achieve the necessary seamless interoperability required by workstations for ubiquitous intelligent communications between the workstations and the sources of data.

Hospital Information Systems: A Review in Perspective

The demand for information for health care is increasing exponentially in volume, content and the number and geographical distribution of users. Most HIS systems commercially available today are based on designs and philosophies of the 1970s. Even though new technology has improved these systems' performance, to meet current demands, concepts must be shifted from the paper-driven system to an electronic system in which the patient is the focus. We need a merger of the functionality of existing systems along with new functionalities and a computer-based patient record. These new health care information system must have no boundaries; data collection must permeate all locations at which a patient receives care, and seamless linkages must connect all individuals who contribute to that patient's care. Even though systems are being designed today to meet these expanded informational requirements, we will not see such systems in use before the next century.

Development and evaluation of a Computer-Assisted Management Protocol (CAMP): improved compliance with care guidelines for diabetes mellitus

Disease-specific standards for directing patient management are becoming increasingly important. These standards, however, are often not followed because they are not sufficiently integrated into the clinical care setting. In this study we describe the development and evaluation of a Computer-Assisted Management Protocol (CAMP) of care guidelines for diabetes mellitus. While other studies have shown improved compliance with rule-based reminders, the CAMP customizes disease-specific care guidelines to individual patients over time. We evaluated the effect of the CAMP on compliance with guidelines in a prospective, randomized controlled study. The study was performed at a family practice clinic where much of the patient record is maintained electronically on The Medical Record (TMR). The management protocol was developed from standards published by the American Diabetes Association. Fifty-eight providers were randomized to either receive or not receive the CAMP for diabetes. Compliance with standards was assessed by chart audits of all encounters with diabetic patients during the study interval. The following conclusion was made: the Computer-Assisted Management Protocol resulted in a statistically significant improvement in compliance with diabetes care standards.

Acute childbirth morbidity: its measurement using hospital charges

OBJECTIVES

An analytic descriptive analysis of acute childbirth morbidity was carried out at Duke University Medical Center, comparing patients delivered by primary cesarean section with those delivered vaginally.

STUDY DESIGN

All primary cesarean deliveries and vaginal deliveries from July 1, 1981, through June 30, 1986, were combined with maternal and infant charge data. A total of 7256 patients were analyzed. A description of the charges for the associated diagnoses was carried out. A morbidity index was used to identify differences in predicted median hospital charges with 95% confidence intervals.

RESULTS

The ratio of mean primary cesarean delivery to mean vaginal delivery total charges was 2.5:1. The magnitude of the mean hospital charges was inversely related to the frequency of the indication with the lowest charges associated with dystocia and the highest with multiple pregnancy. Antepartum risk factors (increased maternal age, patient referral) were associated with increases in maternal and infant morbidity as measured by the morbidity index. Chronic maternal hypertension resulted in decreased maternal morbidity but increased infant morbidity when primary cesarean delivery was used. Although preterm delivery was associated with large increases in charges, it was not significantly altered by using primary cesarean delivery. Risk factors associated with the management of abnormalities of labor were associated with decreases in maternal and infant morbidity when primary cesarean delivery was used.

CONCLUSION

Analysis of acute childbirth morbidity, as measured by hospital charges, showed marked variation of diagnosis and risk-specific charges for patients delivered by primary cesarean section.

Health Level 7: an application standard for electronic medical data exchange

The development of the HL7 is proceeding along two lines. Much effort is being spent on solving the ambiguities of version 2.1 and in extending the standard to areas that are not yet well defined. For example, at a recent working group meeting new messages were defined for dealing with pharmacy records, specifically those relating to IVs. As the number of vendors implementing the standard and as the number of organizations using the standard increase, the standard will continue to be refined and expanded along current formats. As often as possible, these efforts will be "backwards compatible" and may be implemented by vendor choice as the need arises. These changes will be distributed as chapter updates to the current release. A parallel effort is underway that will add formality to the development of the standard. These efforts incorporate a number of case tools, including data modeling. The biggest advantage of this approach will be to reduce significantly the ambiguity in data element definitions. Completeness and correctness of data relations will be enhanced as well. New approaches to documentation, including an object-oriented specification of the data model, will increase understandability. There are still a number of domain-specific areas in which the standard needs to be defined and examples generated to show how the standard applies. Forms of data transfer other than text must be accommodated. HL7 again will take advantage of existing work in these areas, such as the work of American College of Radiology/National Equipment Manufacturers Association (ACR/NEMA) in standards relating to image transfer. HL7 will continue to move toward OSI-compatibility.(ABSTRACT TRUNCATED AT 250 WORDS)

Adopting TMR for physician/nurse use

A major problem which must be overcome before on-line medical records become widely accepted and used is the interface between the human and the computer. This paper presents the evolution of an interface, over time, which recognizes and addresses important characteristics a system must exhibit to encourage human acceptance. Areas discussed include data entry, displays, and function and parameter specifications.

Bedside computerization of the ICU, design issues: benefits of computerization versus ease of paper & pen

This paper describes the design considerations for the implementation of a bedside computer system in an intensive care environment. Specific issues discussed include application design, implementation problems, design revisions, and design solutions. Specific examples of the above issues include passive data acquisition from bedside devices and computer systems, automated nursing assessments, respiratory therapy assessments and integrated reports.

Changing rates of cesarean delivery: the Duke experience, 1978-1986

There was a steady increase in the yearly cesarean delivery rate, from 14.0 to 24.8%, between July 1, 1978 and June 30, 1986 at Duke University Medical Center; this was associated predominantly with an increase in the rate of primary cesarean deliveries. The three most frequent major diagnoses associated with primary cesarean delivery changed significantly over the study period. Fetal compromise became the most commonly associated diagnosis (from third), dystocia second (from first), and maternal disease third (from second). The categories of fetal positional abnormalities (fourth), abnormalities of placentation (fifth), and multiple pregnancy (sixth) did not change in rank. Primary cesarean delivery patients were compared with patients who delivered vaginally using odds ratios, prevalence, and population-attributable fractions. The risk factors of nulliparity, gestational age less than 37 weeks, late decelerations, and referral had the largest impact on the primary cesarean rate. Decreases in rates related to an increased tolerance of abnormalities of labor were overshadowed by the effects of increased concerns related to fetal health.

Improving drug prescribing in a primary care practice

A model for improving physician prescribing that utilizes computerized feedback was studied in a family medicine residency practice. Resident and faculty physicians were stratified by level of experience and randomized into two groups. For 9 months the experimental group received monthly printouts identifying drugs they had prescribed by brand name with estimates of cost savings that might have been realized by prescribing generic drugs. The control group received no feedback. Prescription monitoring of both groups continued for 12 months after all feedback had ceased. Median weighted rates of generic prescribing for the experimental physicians were 14% for the baseline, 67% for the feedback, and 54% for the follow-up periods. Rates for the control physicians for the three periods were 32%, 37% and 31%, respectively. The increase in generic prescribing by physicians in the experimental group was significantly greater than for control physicians (P = 0.01). The feedback model improved rates of generic prescribing but should be evaluated for broader areas of physician prescribing.