Tuesday, February 13, 2007
All at Sea
Not much changes in 10 years! Here is an updated and shortened version of a paper, which was first published in the Health Service Journal on 30 Jan 1997.
Until the 18th century, everyone agreed that going to sea was dangerous and navigation was a major problem. The loss of a ship at sea was a human and economic disaster. Most people considered the failing to be one of culture, organisation and education. If only sailors drank less, were more vigilant and better trained, fewer ships would sink. Others claimed that astronomers would produce a solution by tracking the movements of the stars and planets. Indeed, astronomers had already solved the problem of finding a ship’s latitude, by measuring the angle of the sun. Few believed that the problem of navigation would be solved by combining astronomy with a new technology – clock-making. Yet this is what happened when the development of accurate chronometers allowed sailors to calculate their longitudes.
Today, we have more fear about going into a hospital than going to sea or flying. Experienced observers claim that the problems are more of organisation, culture and education than of technology. Others suggest that current technology can deliver what is required, provided implementation is done carefully and with greater commitment to data quality. Such arguments can be supported by evidence, but they do not provide the breakthrough that is so badly needed. Indeed, in the 18th century, experienced, vigilant seamen with some astronomical knowledge were less likely to be shipwrecked than those that were not.
Clinical decisions determine more than 80 per cent of the variance in the cost of non-acute hospital care for patients with similar conditions. All decisions, clinical or otherwise, are based on the information available at the time. It is the therefore remarkable that only about 5 per cent of hospital doctors use computers in their clinics. Compare this with the number of clinic clerks (95 per cent) who use computers. This phenomenon is not confined to the UK. In all countries, hospital doctors have resisted using computers.
Let’s go back to basics. Computers serve two main functions: communication and counting. Communication is a core healthcare activity. Billions of documents are produced every year in the NHS. The main reason why hospitals exist is to facilitate the exchange of communication (verbal and written) between the staff. It is useful to distinguish four levels of communication:
1 within the work-group responsible for the care of individual patients;
2 between specialised diagnostic and treatment departments to request services and to report on procedures;
3 between hospital clinicians, GPs and community staff;
4 from the care provider to purchasers and other agencies.
There are listed in decreasing order of volume of information, but paradoxically, in increasing order of investment in technology. One important reason is that computers are paid for by managers. Managers need information that can be counted – the second basic function of computers. Counting is the basis of all management, accountancy and research.
We need a system that enables both communication (messaging) and counting, but the requirements of messaging and counting cannot easily be reconciled in a single structure. You cannot use e-mail for counting, and you not use a spreadsheet for messaging. They perform different jobs.
One way to resolve the issue is to combine the approaches. In the mid 1990s, the NHS electronic patient record programme defined the electronic patient record as “a dynamic collection of messages, held electronically, created by healthcare professionals principally to inform themselves and others about the provision of health to an individual patient”. This mirrors the paper record, which consists of referral letters, assessments, test reports, nursing notes, clinic letters, discharge summaries and a wide range of other notes.
Well-formed messages (electronic or paper) have a clearly identified originator, a list of recipients and a clear context. The content is determined by the author. If there is any doubt about how to interpret a message, the receiver can contact the author requesting elaboration.
Database structures are designed by a computer professionals, who neither originate nor use the records they contain. The creator of each database record usually has little if any control over who if anyone will ever see the data, or for what purpose. The user seldom knows who entered the data. Lack of context information makes it difficult for the reader to evaluate the trustworthiness of what it contains.
Today, databases provide information for healthcare management, but databases need to be used together with communications and messaging tools, to help clinicians help patients and enable improvements in productivity, for example by performing simple safety checks automatically.
The solution to the maritime navigation problem lay in combining different tools. It took both good technology (clocks) and good science (astronomy) to solve the problem of finding longitude. Similarly we need to focus on communications as well as data base tools to help solve our healthcare problems.
Until the 18th century, everyone agreed that going to sea was dangerous and navigation was a major problem. The loss of a ship at sea was a human and economic disaster. Most people considered the failing to be one of culture, organisation and education. If only sailors drank less, were more vigilant and better trained, fewer ships would sink. Others claimed that astronomers would produce a solution by tracking the movements of the stars and planets. Indeed, astronomers had already solved the problem of finding a ship’s latitude, by measuring the angle of the sun. Few believed that the problem of navigation would be solved by combining astronomy with a new technology – clock-making. Yet this is what happened when the development of accurate chronometers allowed sailors to calculate their longitudes.
Today, we have more fear about going into a hospital than going to sea or flying. Experienced observers claim that the problems are more of organisation, culture and education than of technology. Others suggest that current technology can deliver what is required, provided implementation is done carefully and with greater commitment to data quality. Such arguments can be supported by evidence, but they do not provide the breakthrough that is so badly needed. Indeed, in the 18th century, experienced, vigilant seamen with some astronomical knowledge were less likely to be shipwrecked than those that were not.
Clinical decisions determine more than 80 per cent of the variance in the cost of non-acute hospital care for patients with similar conditions. All decisions, clinical or otherwise, are based on the information available at the time. It is the therefore remarkable that only about 5 per cent of hospital doctors use computers in their clinics. Compare this with the number of clinic clerks (95 per cent) who use computers. This phenomenon is not confined to the UK. In all countries, hospital doctors have resisted using computers.
Let’s go back to basics. Computers serve two main functions: communication and counting. Communication is a core healthcare activity. Billions of documents are produced every year in the NHS. The main reason why hospitals exist is to facilitate the exchange of communication (verbal and written) between the staff. It is useful to distinguish four levels of communication:
1 within the work-group responsible for the care of individual patients;
2 between specialised diagnostic and treatment departments to request services and to report on procedures;
3 between hospital clinicians, GPs and community staff;
4 from the care provider to purchasers and other agencies.
There are listed in decreasing order of volume of information, but paradoxically, in increasing order of investment in technology. One important reason is that computers are paid for by managers. Managers need information that can be counted – the second basic function of computers. Counting is the basis of all management, accountancy and research.
We need a system that enables both communication (messaging) and counting, but the requirements of messaging and counting cannot easily be reconciled in a single structure. You cannot use e-mail for counting, and you not use a spreadsheet for messaging. They perform different jobs.
One way to resolve the issue is to combine the approaches. In the mid 1990s, the NHS electronic patient record programme defined the electronic patient record as “a dynamic collection of messages, held electronically, created by healthcare professionals principally to inform themselves and others about the provision of health to an individual patient”. This mirrors the paper record, which consists of referral letters, assessments, test reports, nursing notes, clinic letters, discharge summaries and a wide range of other notes.
Well-formed messages (electronic or paper) have a clearly identified originator, a list of recipients and a clear context. The content is determined by the author. If there is any doubt about how to interpret a message, the receiver can contact the author requesting elaboration.
Database structures are designed by a computer professionals, who neither originate nor use the records they contain. The creator of each database record usually has little if any control over who if anyone will ever see the data, or for what purpose. The user seldom knows who entered the data. Lack of context information makes it difficult for the reader to evaluate the trustworthiness of what it contains.
Today, databases provide information for healthcare management, but databases need to be used together with communications and messaging tools, to help clinicians help patients and enable improvements in productivity, for example by performing simple safety checks automatically.
The solution to the maritime navigation problem lay in combining different tools. It took both good technology (clocks) and good science (astronomy) to solve the problem of finding longitude. Similarly we need to focus on communications as well as data base tools to help solve our healthcare problems.
Labels: Computer Projects, e-Health, Healthcare IT
Sunday, October 02, 2005
SNOMED Comes to Hammersmith
The SNOMED CT road-show comes to the Hammersmith Novotel this week, just over three weeks after the SNOMED User Group meeting in Chicago. It all starts Monday, 3 October 2005, with the joint SNOMED/HL7 Terminfo project meeting. The Terminfo project addresses the issue of how best to incorporate SNOMED CT into HL7 Version 3 message structures, where there is more than one way to do the same task, such as including a diagnosis or drug allergy. Specific guidelines are being drawn up.
Tuesday is devoted to SNOMED Working Groups on Anaesthesia, Concept Model and Primary Care. These groups continue on Wednesday, along with the Nursing and Pharmacy Working Groups. Thursday and Friday are devoted to the meetings of the SNOMED International Editorial Board.
Tuesday is devoted to SNOMED Working Groups on Anaesthesia, Concept Model and Primary Care. These groups continue on Wednesday, along with the Nursing and Pharmacy Working Groups. Thursday and Friday are devoted to the meetings of the SNOMED International Editorial Board.
Friday, September 16, 2005
HL7 Certification Grows
More than 1000 candidates will be certified in HL7 Version 2 by the end of 2005. The HL7 Version 2 Certification tests address Chapter 2 (Control) of the HL7 Version 2 Standard and have been available since 2001.
The HL7 Education Committee, meeting this week in San Diego, resolved to move forward with a new project to develop formal curricula and certification tests for HL7 Version 3. The first Version 3 certification tests are likely to cover understanding of the HL7 Version 3 RIM (Reference Information Model).
The HL7 Education Committee, meeting this week in San Diego, resolved to move forward with a new project to develop formal curricula and certification tests for HL7 Version 3. The first Version 3 certification tests are likely to cover understanding of the HL7 Version 3 RIM (Reference Information Model).
Wednesday, September 14, 2005
Welcome Jeremy Thorp to BSI IST/35 Chair
Jeremy Thorp, Head of Business Architecture at NHS Connecting for Health, is to become the new Chair of BSI IST/35, the UK's national health informatics standards committee, it was announced at the IST/35 meeting on 7 September. He replaces Ray Rogers, who has held the post since 1991. Pat Village, who has served as secretary since 1991, is also moving on. BSI IST/35 is important because it is the one and only body that represents the UK on international standards bodies such as CEN TC251 (Europe) and ISO 215 (International). Jeremy Thorp is one of the few people in NHS Connecting for Health who has direct experience of healthcare standards development. Jeremy is a key-note speaker at HL7 uk 2005
e-GIF Technical Standards Catalogue 6.2
The latest version of the Technical Standards Catalogue (6.2) is now available on GovTalk. The main changes from 6.1 are mainly related to web services, smart cards and other fast evolving technologies. In Healthcare the 3 "adopted" standards (the highest level) are HL7 V3, the NHS Data Dictionary and SNOMED CT. Download from: e-Government Unit e-GIF TSC 6.2.
Tuesday, September 13, 2005
HL7 uk 2005 Conference Sponsors
BT, CSW Group, InterSystems, iSOFT, Microsoft and Oracle are sponsoring the HL7 uk 2005 conference to be held at the Hotel Russell on 2-3 November 2005. See HL7 uk 2005 Web-site
Monday, September 12, 2005
Brailer's Wake-Up Call to Standards Developers
Today at the HL7 Annual Plenary Meeting in San Diego, David Brailer, President Bush's National Health Information Technology Coordinator issued a wake-up call to the healthcare standards community. He stated that the standards development infrastructure was "fatally broken", with some 30 different and competing standards development organizations. Yet he claimed that interoperability is his no. 1 priority and depends on standards.
He demanded "use-case driven" standards. If the standards community cannot agree on what is needed to achieve a business task, then he is putting in place a process that will choose "singular" solutions for each use case.
He observed that people take risks when they adopt standards! Ultimately standards are about money. They change the power structure by giving more power to ultimate customers (and their intermediaries) and away from system suppliers.
He demanded "use-case driven" standards. If the standards community cannot agree on what is needed to achieve a business task, then he is putting in place a process that will choose "singular" solutions for each use case.
He observed that people take risks when they adopt standards! Ultimately standards are about money. They change the power structure by giving more power to ultimate customers (and their intermediaries) and away from system suppliers.
Friday, September 09, 2005
New SNOMED Standards Development Organization
Kevin Donnelly, Vice President and General Manager of SNOMED International, today told the SNOMED User Group about the plan to create a new SNOMED Standards Development Organization (SDO), which has been approved by the College of American Pathologists (CAP) Board of Directors. The SDO is to be responsible for all components of SNOMED CT needed for International use. National Centres will develop extensions for local use. The plan will be implemented over the next 6-12 months. The CAP will continue to provide services for the new SDO for a transition period of up to 5 years. The aim is to make SNOMED CT freely available throughout the world and to encourage international uptake, collaboration, alignment, contributorship and the ability to meet local terminology needs.
Few Brits at SNOMED International User Group
Why are so few Brits here at the 7th Annual SNOMED User Group Meeting in Chicago? I can only find two other paying delegates on the delegate list, in addition to David Markwell and Ed Cheetham who are speaking, yet Iceland has 6 people here.
SNOMED CT is one of the mission-critical standards on which the whole NHS Connecting for Health programme is based. It tackles the most complex issue in health informatics – structured clinical terminology – and is inevitably complex itself. As Kent Spackman pointed out yesterday, even implementing a problem list is "non-trivial". Do people recognise that they do not know what they do not know!
SNOMED CT is one of the mission-critical standards on which the whole NHS Connecting for Health programme is based. It tackles the most complex issue in health informatics – structured clinical terminology – and is inevitably complex itself. As Kent Spackman pointed out yesterday, even implementing a problem list is "non-trivial". Do people recognise that they do not know what they do not know!
Thursday, September 08, 2005
Data Quality and Coding Rules
Kent Spackman – Chair of the SNOMED International Editorial Board – expounded two new rules in his Tutorial for the SNOMED User Group in Chicago today:
1. The First Rule of Data Quality is that the quality of data collected is directly proportional to the care with which options are presented to the user.
2. The First Rule of Coding is that yesterdays data should be usable today.
He explored these rules by considering the challenges and obstacles involved in using SNOMED CT for creating Problem Lists, Medication Lists and Medication Allergy decision support. And he showed that the key to making this work properly is the skillful use of SNOMED CT Subsets and the appropriate use of Post-coordinated concepts.
Subsets are lists of SNOMED concepts for a specific purpose.
Post-coordination is the ability to link concepts together to create more complex ones, using qualifying relationships such as body site, laterality or severity.
1. The First Rule of Data Quality is that the quality of data collected is directly proportional to the care with which options are presented to the user.
2. The First Rule of Coding is that yesterdays data should be usable today.
He explored these rules by considering the challenges and obstacles involved in using SNOMED CT for creating Problem Lists, Medication Lists and Medication Allergy decision support. And he showed that the key to making this work properly is the skillful use of SNOMED CT Subsets and the appropriate use of Post-coordinated concepts.
Subsets are lists of SNOMED concepts for a specific purpose.
Post-coordination is the ability to link concepts together to create more complex ones, using qualifying relationships such as body site, laterality or severity.
