DOCSLIB.ORG

Briefing June 2013 Issue 266

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Briefing June 2013 Issue 266 Hospitals Forum briefing June 2013 Issue 266 The non-executive directors’ guide to hospital data Part four: How to make good use of data – quality and safety, including mortality, activity data, contracting and finance Key points Understanding your organisation’s data is an essential part of providing effective oversight. But data may not always give you the complete picture • Many indicators can be used and it is important to first understand what data is available, how it is to monitor quality and safety. recorded and what these records are used for. They should not be considered in isolation. This Briefing will help non-executive directors (NEDs) better understand • The quality of indicators is NHS data and how it can be used to determine what is going on in their inextricably linked to the quality hospital. For the purposes of this Briefing we examine data in the acute of data. care setting only. Data is of course collected in primary care by GPs, • The accurate coding of clinical pharmacists, dentists and opticians, but the various datasets are not activity is fundamental, linked by the NHS. particularly when constructing mortality rates. This Briefing looks at how to make good use of data across: quality and safety, including mortality; activity; contracting; and finance. It also • Good clinical coding is also crucial includes a short technical guide on ICD-10, OPCS-4 and healthcare in determining the right level of resource groups (HRGs). payments for healthcare services, and thus defining NHS tariffs. How to make good use non-executive directors need to of data be aware that they have to ask Data, by its nature, is something the right questions in order to that should be handled carefully. establish what the data is telling NHS data is no exception and them. The questions are simple Produced in association with briefing 266 The non-executive directors’ guide to hospital data: part four ‘There is no single ‘dashboard’ Key questions for NEDs to ask that should be used; the recommendation is that • Do we have a collaborative approach between clinicians and coders to trusts monitor a wide spread ensure the data we hold is accurate? of indicators and consistently • Is there a robust clinical governance process in place regarding the track these over time’ review of mortality to provide assurances to the board? • Do we have mortality sub-groups in place within our organisation? and straightforward. For instance, do our emergency admissions • Are readmission cases a result of the quality of care we provide or a lack figures include mothers and of community services for the treatment of certain conditions? babies? Without answers to these questions the data cannot be • Are we monitoring the same range of quality and safety indicators interpreted correctly. over time? We have already highlighted the • Do we benchmark our performance against other providers? sort of questions that can be asked of NHS data in parts one, trusts to make sure they are characteristics which would two and three of this series. We comparing like with like, and peer affect the indicators (such as now turn to how the data can be used to help trusts improve the group comparison shines a light age, diagnosis, co-morbidities quality and safety of the services on the areas where improvements and procedure). Adjustment they provide. can be made. can be made for a number of indicators, including length Quality and safety However, non-executive directors of stay, readmissions and need to be aware that in order mortality. Although it sounds There are many indicators to benchmark data a number of straightforward, performing that can be used to monitor things have to happen. First is clinically credible risk the quality and safety of care adjustment of the raw numbers. adjustment is difficult and there provided by a hospital trust. At its simplest level, this is for the are different ways of adjusting However, there is no single size of the hospital trust and the for risk. The following example ‘dashboard’ that should be used; number of patients that it delivers explains the basics. the recommendation is that care to in comparison to other trusts monitor a wide spread of trusts (turning “raw” scores into Risk adjusted indicators and consistently track rates, for example, the number of readmissions within these over time. events per 1,000 patients). 30 days Although this intelligence is The next stage of refinement is This indicator is the relative risk useful, it is one dimensional. the construction of risk-adjusted of readmission within 30 days. The board can only see how the indicators. Risk adjustment is It is the ratio of the observed trust is performing in isolation. the process of adjusting for risk number of readmissions Benchmarking is required to factors (which might explain to the expected number of assess performance within, and variation in outcome) so that readmissions (taking into across, the healthcare system. comparison can be made. account the various risk factors). There is, understandably, a In essence, the adjustment This ratio is then multiplied growing market for benchmarking is not only for the volume of by 100. The factors that will services. Providers work with patients seen but also key influence the ratio are: 02 briefing 266 The non-executive directors’ guide to hospital data: part four 1. The accuracy of the observed When looking at quality of care, number of readmissions indicators like readmissions ‘Non-executives should be aware This sounds like it should be a within 30 days and length of that a longer than expected simple count: how many people stay might be considered useful. length of stay can be a result of have come back as an emergency It is possible that a higher than poor coding of co-morbidities’ admission within 30 days of expected level of readmissions discharge? However, the rules for within 30 days in a given specialty this indicator, used for the current (such as orthopaedics) indicates • HSMR – Hospital Standardised Mortality Ratio (Dr Foster financial penalty, include people that operations are not being Intelligence). who end up being admitted to carried out as successfully as they a different hospital within the are elsewhere. Likewise, a longer SHMI is the only measure now 30 day period, which your staff than expected length of stay on published on NHS Choices. will not be able to ‘see’ until the a specific ward might indicate central data lets them know. problems with the quality of care All three measures are usually provided – perhaps a few patients expressed as a value of 100. An 2. The way the expected number developed pressure ulcers. index above 100 indicates more of readmissions is calculated deaths than expected, whilst a There is no one single ‘correct’ However, non-executives should be lower index indicates fewer deaths way to predict the expected aware that a longer than expected than expected. The Health & number. It relies on having length of stay can also be a result Social Care Information Centre a good understanding of the of poor coding of co-morbidities. publishes SHMI slightly differently. way services are provided. For This means hospital data will It uses 1.00 as the average, which instance, are there any sub- not record how sick the patient is often multiplied by 100 to allow groups who should be excluded is. Alternatively, it could signal a comparison. for specific reasons? Looking pathway issue where a consultant at readmissions, many cancer might be making a judgement The NHS Medical Director has treatments are not carried out to a based on years of practice such made it clear that mortality fixed timetable so patients might as: “I always keep my patients in ratios are one of a number appear as readmissions when overnight”. Since indicators can of indicators that should be they are on a known treatment be skewed in this way, it is always monitored, and hospital trusts pathway and potentially should in the board’s interest to seek should not rely on a single be excluded from this indicator. assurance from divisional teams indicator. In addition, six factors The calculation then requires the where there is an outlier and, if need to be taken into account use of good statistical methods appropriate, instigate consultant- which can have a direct impact to make the ‘best’ prediction level investigation. on mortality ratios. These are: – but it is possible for two • percentage of the population different answers to be produced, Mortality indicators have taken depending on the methodology. who die in hospital as opposed on a particular importance in to outside the last few years. The three main measures currently used in • population demography ‘A high mortality ratio is England are: not necessarily a sign of • different pathways of care a poor performing trust • SHMI – Summary Hospital-level • zero length of stay emergencies from a safety perspective... Mortality Indicator (Heath & • palliative care a number of factors can Social Care Information Centre) affect an individual figure’ • data quality. • RAMI – Risk Adjusted Mortality Index (CHKS) (For more detail see www.chks.co.uk) 03 briefing 266 The non-executive directors’ guide to hospital data: part four continued discussion about their Activity data, contracting ‘One of the most effective usefulness as a predictor of the and finance ways to use mortality ratios is safety of care, it is clear that to develop a process of review We have already explained that they are a useful ‘smoke alarm’ around one measure’ the NHS contracting rules set to trigger further investigation. down the principles for which Boards must ensure they have payments should be made. Under an understanding of what the A high mortality ratio is not the standard contract, hospital indicators say about their own necessarily a sign of a poor income is determined by payment organisation and have the performing trust from a safety by results (PbR), which effectively perspective.
Load more

Recommended publications
  • A Reason for Visit Classification for Ambulatory Care
    DATA EVALUATION AND METHODS RESEARCH Series 2 Number 78 A Reasonfor Visit Classification for AmbulatoryCare This report presents a classification system developed to code rea- sons for seeking ambulatory medical care. DHEW Publication No. (PHS) 79-1352 U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE Public Health Service Office of the Assistant Secretary for Health National Center for Health Statistics Hyattsville, Md. February 1979 Library of Congress Cataloging in Publication Data Schneider, Don. A reason for visit classification for ambulatory care. (Vital and health statistics: Series 2, Data evaluation and methods research; no. 78) (DHEW publication; no. (PHS) 79-1352) Includes bibliographical references and index. 1. Nosology. 2. Ambulatory medical care. I. Appleton, Linda, joint author. II. McLemore, Thomas, ,joint author. III. Title. IV. Series: United States. National Center for Health Sta- tistics. Vital and health statistics: Series 2, Data evaluation and methods research; no. 78. V. Series: United States. Dept. of Health, Education and Welfare. DHEW publication; no. (PHS) 79-1352. [DNLM: 1. Ambulatory care. 2. Classification. W2 A N148vb no. 78] RA409.U45 no. 78 [RB115] 312’.07’23s [616’.001’2] 78-11549 NATIONAL CENTER FOR HEALTH STATISTICS DOROTHY P. RICE, Director ROBERT A. ISRAEL, Deputy Director JACOB J. FELDMAN, Ph.D., Associate Director for Analysis GAIL F. FISHER, Ph.D., Associate Director for the Cooperative Health Statistics System ELIJAH L. WHITE, Associate Director for Data Systems JAMES T. BAIRD, JR., Ph.D., Associate .Directorfor International Statistics ROBERT C, HUBER, Associate Director for Management MONROE G. SIRKEN, Ph. D., Associate Director for Mathematiccd Statistics PETER L. HURLEY, Associate Director for Operations JAMES M.
    [Show full text]
  • International Classification of Diseases- 10Th Revision (ICD-10)
    Massachusetts Deaths 1999 Frequently Asked Questions: International Classification of Diseases- 10th Revision (ICD-10) Effective with data year 1999, the International Classification of Diseases, Tenth Revision (ICD-10) is used to code and classify causes of death. This change in coding systems will affect the comparison of mortality data between 1999 and previous years. Definitions and Concepts What is ICD-10? ICD-10 is an abbreviation for the International Classification of Diseases- Tenth revision. The International Classification of Diseases is a classification system developed by the World Health Organization (WHO). The United States uses the ICD in accordance with an international agreement. The purpose of an international classification system is to promote international comparability in collecting, classifying, and tabulating mortality statistics. Why has the ICD been revised? The ICD is revised to reflect changes in medical classification practices. The ICD was first implemented in 1900, and has undergone revisions approximately every ten years, except for the Ninth revision which was in effect between 1979-1998. There has been an increase in the number of specific cause of death codes from about 4,000 codes in ICD-9 to about 8,000 codes in ICD-10. Beginning with 1999, mortality data are coded according to the Tenth revision of the ICD. Can I compare data classified in ICD-10 to data classified in ICD-9? Differences in the coding between ICD-9 and ICD-10 make direct comparisons between the two classification systems difficult for three reasons: 1) there have been changes made in the codes that are assigned to causes of death; 2) there have been changes to the rules used to determine the underlying cause of death; and 3) there have been changes in the codes that comprise the leading cause of death categories.
    [Show full text]
  • ICF: REPRESENTING the PATIENT BEYOND a MEDICAL CLASSIFICATION of DIAGNOSES Posted on July 17, 2009 by Administrator
    ICF: REPRESENTING THE PATIENT BEYOND A MEDICAL CLASSIFICATION OF DIAGNOSES Posted on July 17, 2009 by Administrator Category: Clinical Terms & Vocabularies Tag: ICF Page: 1 by Kathy Giannangelo, RHIA, CCS; Sue Bowman, RHIA, CCS; Michelle Dougherty, RHIA, CHP; and Susan Fenton, MBA, RHIA Abstract The International Classification of Functioning, Disability and Health (ICF) is a component essential to ensuring the collection of accurate and complete healthcare data that correctly reflect the care provided to individuals. In fact, many countries outside of the United States have found uses for ICF. While research continues in the United States on the potential value of implementing ICF, deliberations are establishing the need to implement ICF to develop knowledge about the physical, mental, and social functioning of patients. In the course of these deliberations, issues related to current data collection activities, the use of ICF in an electronic health record (EHR) system, training requirements, and terminology maps are beginning to emerge. Introduction The International Classification of Functioning, Disability and Health (ICF) is one of the World Health Organization’s Family of International Classifications (WHO-FIC).1 The purpose of WHO-FIC is to promote the appropriate selection of classifications in the range of settings in the healthcare field around the world. Since a single classification system cannot encompass all types of healthcare information or provide the level of detail desired for various uses of healthcare data, multiple classifications have been developed to meet specific user requirements. WHO-FIC provides a framework to code a wide range of information about health (for example, diagnosis, functioning, disability, and reasons for contact with health services) and uses standardized language, permitting communication about health and healthcare around the world and among various disciplines and sciences.
    [Show full text]
  • Toward a Preliminary Layered Network Model Representation of Medical Coding in Clinical Decision Support Systems
    146 Int'l Conf. Health Informatics and Medical Systems | HIMS'15 | Toward a Preliminary Layered Network Model Representation of Medical Coding in Clinical Decision Support Systems H. Mallah1, I. Zaarour2, A. Kalakech2 1 Doctoral School of Management Information Systems, Lebanese University, Beirut, Lebanon 2 Management Information Systems Department, Lebanese University, Beirut, Lebanon Abstract - The aim of this research is to describe the impact increase in medical errors, poor trainings to human resources of Medical Codes (MC) on building a successful Clinical and their derivatives. Many studies were done on how to Decision Support System (CDSS) that helps in structuring a computerize medical processes, especially critical decisions beneficial Information Model, we highlight a preliminary in medical fields, using Artificial Intelligence. From these Architectural Layered network model to assist health care needs, Clinical Decision Support System (CDSS) and givers and researchers in representing their work in a Medical classification or medical coding gave birth after the unified manner. Via a descriptive methodology, we development of the Electronic Health Record (EHR) and the incorporate the relationship between (MC) and (CDSS), we Computer-based Physician Order Entry (CPOE). CDSS is discussed a brief state of art of (CDSS) as well as medical the key of success of clinical information systems designed coding as a process of converting descriptions of to assist health care providers in decision making during the medical diagnoses and procedures into universal medical process of care. This means that a clinician would cooperate codes and numbers. Integrated with CDSS, medical codes with a CDSS to help determine diagnosis, analysis of patient are illustrated as an advanced search engine in the data.
    [Show full text]
  • The Concept of Mental Disorder and the DSM-V
    ORIGINAL ARTICLE The concept of mental disorder and the DSM-V MASSIMILIANO ARAGONA Chair of Philosophy of Psychopathology, Sapienza University, Rome, Italy In view of the publication of the DSM-V researchers were asked to discuss the theoretical implications of the definition of mental disorders. The reasons for the use, in the DSM-III, of the term disorder instead of disease are considered. The analysis of these reasons clarifies the distinction between the general definition of disorder and its implicit, technical meaning which arises from concrete use in DSM disorders. The characteristics and limits of this technical meaning are discussed and contrasted to alternative definitions, like Wakefield’s harmful/dysfunction analysis. It is shown that Wakefield’s analysis faces internal theoretical problems in addition to practical limits for its acceptance in the DSM-V. In particular, it is shown that: a) the term dysfunction is not purely factual but intrinsically normative/evaluative; b) it is difficult to clarify what dysfunctions are in the psychiatric context (the dysfunctional mechanism involved being unknown in most cases and the use of evolutionary theory being even more problematic); c) the use of conceptual analysis and commonsense intuition to define dysfunctions leaves unsatisfied empiricists; d) it is unlikely that the authors of the DSM-V will accept Wakefield’s suggestion to revise the diagnostic criteria of any single DSM disorder in accordance with his analysis, because this is an excessively extensive change and also because this would probably reduce DSM reliability. In conclusion, it is pointed up in which sense DSM mental disorders have to be conceived as constructs, and that this undermines the realistic search for a clear-cut demarcation criterion between what is disorder and what is not.
    [Show full text]
  • Thesis Submitted to the University of Manchester for the Degree of Doctor of Philosophy in the Faculty of Medicine
    Structured Methods of Information Management for Medical Records A thesis submitted to the University of Manchester for the degree of Doctor of Philosophy in the Faculty of Medicine 1993 William Anthony Nowlan Table of Contents Structured Methods of Information Management for Medical Records 1 Table of Contents 2 List of Figures 8 Abstract 9 Declaration 11 Acknowledgements 12 The Author 13 1 Purpose and Introduction 14 1.1 The purpose of this thesis 14 1.2 Motivations for the work 14 1.2.1 Research on clinical information systems: PEN&PAD 14 1.2.2 Characteristics of clinical information systems 15 1.3 What is to be represented: the shared medical model 15 1.3.1 An example: thinking about ‘stroke’ 16 1.3.2 Medical limits on what can be modelled 16 1.4 Scope of the work in this thesis: the limits on context and modelling 17 1.5 Organisation of the thesis 17 2 Coding and Classification Schemes 18 2.1 The changing task: the move from populations to patients 18 2.2 Current methods: nomenclatures and classifications 19 2.2.1 Nomenclatures. 19 2.2.2 Classifications 19 2.2.3 Classification rules 20 2.3 An exemplar classification: the International Classification of Diseases (ICD) 20 2.3.1 History of the International Classification of Diseases 20 2.3.2 The organisation of ICD-9 21 2.3.3 Classification principles within ICD-9 23 2.3.4 Goal-orientation 24 2.3.5 The use of the classification structure: abstraction 24 2.3.6 Extending the coverage of a classification and the ‘combinatorial explosion’ 25 2 Contents 3 2.4 Multiaxial nomenclatures and classifications: SNOMED 26 2.4.1 Organisation of SNOMED 26 2.4.2 SNOMED coding of terms 27 2.4.3 Sense and nonsense in SNOMED 28 2.5 The structure within SNOMED fascicles 28 2.5.1 Abstraction in SNOMED 29 2.5.2 Relationships between topographical concepts in the SNOMED topography fascicle 31 2.6 SNOMED enhancements.
    [Show full text]
  • 2020 Reporting Accurate Diagnosis Quick Reference Guide
    2020 Reporting Accurate Diagnosis Quick Reference Guide Iowa Total Care strives to provide quality healthcare to our membership. We created the Reporting Accurate Diagnosis Quick Reference Guide to help you correctly report the conditions affecting our members with each visit. Please always follow the Official ICD-10-CM Guidelines, State and/or CMS billing guidance and ensure the billing codes are covered prior to submission. For more information, visit www.CMS.gov 1-833-404-1061 IowaTotalCare.com TTY: 711 © 2019 Iowa Total Care. All rights reserved. Quick Reference Guide Reporting Accurate Diagnosis Why is it important? Conditions that go undocumented usually also go untreated. Comprehensive documentation and coding provide a complete and thorough insight into a patient’s health profile for all those participating in their care. Coded data translates into: Identification of members who may need disease management intervention. Accurate and complete claims for appropriate reimbursement. Tailored care management programs based on the specific conditions affecting our members. Better coordination of care between member, provider, and health plan More meaningful data exchanges between health insurance plans and providers, which helps members by: o Identifying new problems early o Reinforcing self-care and prevention strategies o Coordinating care collaboratively o Avoiding potential drug/disease interaction Provider Role Providers are the largest source of medical data for accurate, comprehensive coding. It is important for providers to document each existing medical condition and ensure patient charts are thorough. Providers should take every face-to-face encounter as an opportunity to assess the patient’s health and comprehensively document chronic conditions, co-existing conditions, active status conditions, and pertinent past conditions.
    [Show full text]
  • Cac 2010-2011
    CAC 2010 –11 Industry Outlook and Resources Report PREPARED BY: June Bronnert, RHIA, CCS, CCS-P Mark Morsch, MS Bonnie S. Cassidy, MPA, RHIA, Kathleen Peterson, RHIA, FAHIMA, FHIMSS Kozie V. Phibbs, MS, RHIA Shirley Eichenwald Maki, MBA, Philip Resnik, PhD RHIA, FAHIMA Rita Scichilone, MHSA, RHIA, CCS, Heather Eminger CCS-P, CHC James Flanagan, MD, Ph Gail Smith, RHIA, CCS-P CAC 2010 –11 Industry Outlook and Resources Report TABLE OF CONTENTS Transitioning Coding Professionals into New Roles in a Computer-Assisted Coding (CAC) Environment ......................................................1-15 A Coder’s Guide to Natural Language Processing and Coded Data ......................16-18 Top Ten Questions to Ask the Vendors about Computer-Assisted Coding (CAC) ........19 Show Me the Benefits: Capabilities of CAC Products ............................................20-24 SPONSOR COMMENTARIES ..........................................................................25 American Health Information Management Association (AHIMA). © Copyright 2011. All rights reserved. Transitioning Coding Professionals into New Roles in a Computer-Assisted Coding (CAC) Environment Transitions are common in the healthcare marketplace today. Encoding of clinical data is a key element of the health information management (HIM) profession. It binds the provision of care to the systems serving both clinical and administrative needs. Coding professionals are experts at adapting to change, since the code sets and coding guidelines require frequent updates to keep pace with the science of medicine. Clinical terminologies and classification systems are an essential and compelling part of an optimal healthcare delivery system. The people who manage and use code sets can expect significant changes ahead, creating new roles and demanding new competencies. This paper addresses the recent shift to automated workflow and innovation in information systems designed to improve the traditional coding process.
    [Show full text]
  • Mapping SNOMED CT to ICD-10-CM By
    Mapping SNOMED CT to ICD-10-CM By Junchuan Xu A Dissertation Submitted to Rutgers Biomedical and Health Sciences School of Health Related Professions Department of Health Informatics [In partial fulfillment of the requirements for the Degree of Doctor of Philosophy] January 2016 Mapping SNOMED CT to ICD-10-CM By Junchuan Xu Dissertation Committee: Shankar Srinivasan, Ph.D., Advisor Masayuki Shibata, PhD. Frederick Coffman, Ph.D. Approved by the Dissertation Committee: _________________________ Date __________ _________________________ Date __________ _________________________ Date __________ i ABSTRACT A SNOMED CT-encoded problem list is required to satisfy the Certification Criteria for Stage 2 “Meaningful Use”. ICD-10-CM has replaced ICD-9-CM as the reimbursement code set in 2015. Having a cross-map from SNOMED CT to ICD-10- CM would promote the use of SNOMED CT as the primary problem list terminology, while easing the transition to ICD-10-CM. There is no established principle and methodology on systematically and semantically linking SNOMED CT to ICD-10-CM. This research project describes the development of mapping principle, mapping guidelines, mapping tools and mapping methodology for a rule-based crosswalk to support semi-automatic generation of ICD-10-CM codes from SNOMED CT-encoded data. A series of mapping guidelines were developed based on the clinical use case, SNOMED CT modeling convention, and ICD-10-CM classification guidelines. One of the important methodology in developing the map set is using triangulation in generating legacy maps. Using the SNOMED CT to ICD-9-CM map and General Equivalence Mappings sequentially, Indirect Map was generated from SNOMED CT to ICD-10-CM for 96.2% of the SNOMED CT concepts within the scope of the study.
    [Show full text]
  • Whofic2011eicstatusreport20
    WHOFIC 2011/ C002b WHO-FIC Education and Implementation Committee: A Status Report, 2010-2011 Cassia Maria Buchalla Sue Walker Co-Chairs Abstract Enter Abstract here The WHO-FIC Education and Implementation Committee (EIC) was established in October 2010, when the Education Committee and Implementation Committee were merged. The EIC assists and advises WHO in implementing the WHO Family of International Classifications (WHO-FIC) and improving the level and quality of their use in member states. The primary function of the Committee described in the 2010-2011 terms of reference is to develop strategies for the implementation of the WHO-FIC with an integrated educational strategy for the Reference Classifications. This is achieved by developing implementation, education, training and certification strategies and tools for the WHO-FIC; identifying best training and implementation practices and providing a network for sharing expertise and experience on training and implementation. EIC is continuing a Joint Collaboration (JC) with the International Federation of Health Information Management Associations and joint projects with the Functioning and Disability Reference Group in addressing tasks for ICD and ICF, respectively. The EIC and JC held three teleconferences in 2011 and a mid-year meeting in Budapest, Hungary to discuss and plan future steps for the web-based training tools for ICD-10 and for ICF, ICD-10 and ICF implementation databases and to make progress on other components of its work plan. This paper and several related papers and posters describe the status of each element in the EIC’s work plan . 1- Introduction The WHO-FIC Education Committee was established at the 2003 WHO-FIC Network meeting in Cologne, Germany, as a successor to the Subgroup on Training and Credentialing of the WHO-FIC Implementation Committee; the latter had been established in Cardiff, Wales in 1999.
    [Show full text]
  • Measuring Diversity in Medical Reports Based on Categorized Attributes and International Classification Systems Petra Přečková*†, Jana Zvárovᆠand Karel Zvára†
    Přečková et al. BMC Medical Informatics and Decision Making 2012, 12:31 http://www.biomedcentral.com/1472-6947/12/31 RESEARCHARTICLE Open Access Measuring diversity in medical reports based on categorized attributes and international classification systems Petra Přečková*†, Jana Zvárovᆠand Karel Zvára† Abstract Background: Narrative medical reports do not use standardized terminology and often bring insufficient information for statistical processing and medical decision making. Objectives of the paper are to propose a method for measuring diversity in medical reports written in any language, to compare diversities in narrative and structured medical reports and to map attributes and terms to selected classification systems. Methods: A new method based on a general concept of f-diversity is proposed for measuring diversity of medical reports in any language. The method is based on categorized attributes recorded in narrative or structured medical reports and on international classification systems. Values of categories are expressed by terms. Using SNOMED CT and ICD 10 we are mapping attributes and terms to predefined codes. We use f-diversities of Gini-Simpson and Number of Categories types to compare diversities of narrative and structured medical reports. The comparison is based on attributes selected from the Minimal Data Model for Cardiology (MDMC). Results: We compared diversities of 110 Czech narrative medical reports and 1119 Czech structured medical reports. Selected categorized attributes of MDMC had mostly different numbers of categories and used different terms in narrative and structured reports. We found more than 60% of MDMC attributes in SNOMED CT. We showed that attributes in narrative medical reports had greater diversity than the same attributes in structured medical reports.
    [Show full text]
  • ICD-10 Auto-Coding System Using Deep Learning
    ISBN 978-981-14-4787-7 Proceedings of 2020 the 10th International Workshop on Computer Science and Engineering (WCSE 2020) Yangon (Rangoon), Myanmar (Burma), February 26- February 28, 2020, pp. 46-51 ICD-10 Auto-coding System Using Deep Learning Ssu-Ming Wang 1, Feipei Lai 1,2, Chang-Sung Sung 2 and Yang Chen 2 1 Graduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan Univ., Taipei, Taiwan 2 Department of Computer Science and Information Engineering, National Taiwan Univ., Taipei, Taiwan Abstract. In this research, we aim to construct an automatic ICD-10 coding system. ICD-10 is a medical classification standard which is strongly related to scope of payment in health insurance. However, the work of ICD-10 coding is time-consuming and tedious to ICD coders. Therefore, we build an ICD-10 coding system based on NLP approach to reduce their workload. The result of f1-score in whole label prediction task is up to 0.67 and 0.58 in CM and PCS, respectively. In addition, recall@20 in whole label prediction task is up to 0.87 and 0.81 in CM and PCS, respectively. In the future, we will keep working on combining the current work with the rule-based coding system and applying the other brand new NLP techniques to improve our performance. Keywords: Deep learning, Deep Neural Network, Natural Language Processing (NLP), ICD-10 1. Introduction Auto-coding for ICD-10 (The International Statistical Classification of Diseases and Related Health Problems 10th Revision, ICD-10) based on free-text electronic health records (EHR) has drawn great attention in the field of clinical management system.
    [Show full text]