Collaborative Care Team in Open Source Medical Challenges and requirements analysis

Version 7 Sept 2016, Etienne Saliez /

• Introduction:

  • The essential goal is to solve patient's health problems, as far as possible.
  • Therefore the question is how informatics and telecommunications can help in the care process, i.e. more than just a simulation of the traditional way of working with papers.
  • The first step is to identify the current issues in general.
  • Next we have to formulate a vision of the care process.
  • Next a break down of the necessary tasks.
  • This chapter is dealing with WHAT to do, while the next one will deals with HOW to achieve solutions taking advantage of sharing know-how in Open Source and managing professional support for large medical organizations.
  • Models can be seen from different points of view, "task model" and "data model".
  • See the project Interactive Medical Mind Maps.

• Current Issues:

  • Issues

• "Task Model"

  • Introduction

    • The essential question is to try to solve health problem of the patient. Therefore it is necessary to analyze the care process.
    • Look at what most matter for the patient. A patient has usually several "Health Issues" needing solutions.
    • Understanding and agreement on task or service models are very critical for the design of informatics solutions..

  • Iterative Care Model

    • Health Issues:
      • Therefore the first step is to identify the problems of the patient. An overview as a complete "Problem List":
        • What matter are "Health Issues", a broader concept than "Diagnoses" ?
        • An essential responsibility of the doctor while many other tasks can be delegated.
        • In most cases a patient has a care team of several professionals, GP, specialists, nurses. At any time they need to share up to date information.
        • They need to understand the information and a standardization of "Health Issues" is necessary.
        • The information must be complete including clinical qualifiers, as degree of belief, importance, etc...
    • Observations:
      • Make explicit on basis of which observations these problems have been identified. A factor of quality of care.
      • Patterns of observations could lead to problem suggestions.
    • Actions:
      • Make explicit what has been decided by who and why. A factor of quality of care.
      • An approach which could provide data for cost/benefit analysis.
    • More on http://www.chos-wg.eu/Models/iterative-care-model.html .
    • Moreover the task model can also be used for training purposes.

  • Logical Components

    • Introduction:
      • In informatics parlance " Objects" are logical entities containing both data and functional aspects. They are the logical building blocks of applications. Such block contain closely related information about a topic and is relatively autonomous.
      • We need a large global vision, but we must breakdown the question to be analyzed. Lets look at one component at a time, about requirement and specifications. Introductions are proposed but much more analysis will be necessary.
    • Components can be shared at different levels:
      • ( A )  Common Generic component, as being necessary in most patient care situations.
      • ( B )  Additional requirements for specialized situations, in function of specific disease or professional specialties. For example GP practice, emergency entrance, medical image processing, lung function tests, drug interaction checks, etc ...... and which need anyhow most of the common components.
      • ( C )  Cultural differences
      • ( D )  Freedom leaving space for local exceptions
  • "Care Team":
    • There is a trend to more and more specialization. In most situations a patient need several health professionals. The difficulties can be alleviated by means of telemedicine.
    • The project should be analyzed in function of the motivations of every stakeholder, patient, community GP, specialists, nurses, paramedical, hospital management.
    • In general sharing up to date information on-line seems better than document exchanges.
  • Semantics:
    • The "Problem Oriented Medical Record" can be seen as a graph. Indeed the natural human way of thinking is based on a kind of N-dimensional network between many concepts. The advantage of new graph technologies should be explored, http://www.iweee.org/2015-las_palmas/presentations/files/Etienne-IWEEE2015.pdf .

• "Descriptive Data Model":

  • Data models are intended to describe static data. Large medical classifications are already available, for example "WHO ICD", "HL7", "Snomed", "ICPC", "UK Read Code", "OpenEHR", etc... https://en.wikipedia.org/wiki/Medical_classification .
  • The task model should be able to accept most usual data types. On condition that the data are properly prefixed, it does not much matter in which of these data format the information is available, even in plain text. Btw plain text is also a useful set of conventions between the members of a linguistic community.
  • Medical information is more than rough data. The Task Model need rigorous standardization of "Clinical Qualifiers":
    • The goal is that the complete information must be shared and well understood between the member of the patient care team, across telemedicine network. Concepts and relations are meaningful as far as they they can be qualified. For example "degree of belief", "importance", "complication of", "risk factor", etc ...
    • Medical information reduced to codes truncated to "yes or no", well or not present, could be misleading for the medical decision process. ( Note: this in contrast to information intended for epidemiology on large populations, in which case the granularity must of course be reduced. )
  • Research:
    • Anonymized patient informations are very useful byproduct for management and research purposes, but avoid to charge all the users by data collection exclusively intended for these purposes.