Summarisation and Visualisation of e-Health Data Repositories

1
Summarisation and Visualisation of e-Health Data
Repositories

• Catalina Hallett, Richard Power, Donia Scott
• Centre for Research in Computing
• The Open University

2
Overview

• Background
• The need for textual and visual summaries
• Typical input
• Requirements
• Textual summaries
• Visual navigator
• Conclusions

3
Background

• Research conducted under the MRC-funded CLEF
  project
• CLEF aims at establishing a technical
  infrastructure for managing repositories of
  aggregated patient data in the domain of cancer
• Currently, the CLEF repository contains about 22k
  patient records
• The aim of the current research is to provide
  solutions for accessing and visualising
  individual patient histories for use in clinical
  care

4
Rationale

• Clinicians and other health professionals use
  patient health summaries at the point of care,
  where time is a critical resource
• Reports provide quick access to an overview of a
  patients medical history
• Typically, an electronic patient record contains
  around 1000 messages
• Even structured, this volume of data is very
  large
• Access to relevant information about particular
  patients is difficult
• Easy visual access to time-oriented data and in
  particular EPRs has long been a topic of
  research
• Knave-II (Shahar et al 2003)
• Lifelines (Plaisant et al 2001)
• Vie-Visu (Horn et al 2002)
• Our research emphasises the role of textual
  reports, enhanced by visual navigation tools

5
Typical input

• A complex representation of an Electronic patient
  record, termed Chronicle
• Structured data
• Demographics
• Age, gender, postal district, ethnical group,
  occupation
• Laboratory findings
• 32 types of haematology findings
• 51 types of chemistry findings
• Cytology reports
• Histopathology reports
• Imaging studies
• Radiology procedure, site, diagnosis, morphology,
  topography, report, indication, department
• Treatments
• Prescription drugs
• IV chemotherapy
• Radiotherapy
• Surgical procedures
• Diagnoses
• Clinical diagnosis
• Cause(s) of death

Structured data Processed clinical narratives
Inference engine CLEF chronicle
6
The CLEF chronicle - the story of an illness
Human1382
7
Requirements

• Fast, on-the-fly generation of textual reports
  from complex EPRs
• Information extracted selectively and further
  aggregated to allow a 30 second overview of a
  patients history
• Presentation in an easily understandable format
  natural language and graphics
• Alternative views of the patient record, e.g.
• Summaries from various viewpoints
• Graphics, timelines and other visual aids
• Easy visual navigation, with the possibility of
  drilling down into events

8
Text vs Graphics

• Textual reports
• are easy to read and understand - natural
  language is the best medium of aggregating
  events, expressing causality relations,
  describing temporal relations
• can be customised to the type of information
  needed
• provide a quick way of identifying errors in the
  patient record
• However
• If all information is to be included, textual
  reports are likely to be very large
• It can be difficult to find specific pieces of
  information in a large text document
• Large amounts of text can be difficult to
  navigate through

9
Text vs Graphics

• Text is better for
• Presenting snippets of information, particularly
  if they contain relations between facts
• Producing printable documents
• Communicating information between medical
  practitioners
• Graphical visualisations are better for
• Presenting large amount of data in a small space
  (i.e., one computer screen)
• Data navigation

Solution textual summaries backed up by a
visual navigator
10
Textual summaries

• Types of summary
• Longitudinal events are presented
  chronologically and schematically
• Focused events are presented in logical order
  according to the incidence of more important
  events, selected by the user (a summary of
  interventions, a summary of investigations of
  type x-ray)

11
Textual summaries

• Three steps
• Content selection
• - Select important events to include in a summary
• Text structuring
• - Having a pool of events, arrange them in a
  coherent way to convey a certain viewpoint
• Realisation
• realise the defined text structure as paragraphs,
  sentences, phrases
• format according to the summary type

12
Content selection

• Users can select types of information they want
  in a summary (e.g., a summary of interventions, a
  summary of investigations of type x-ray)
• Facts corresponding to these focal events will
  form the spine of the summary
• Facts linked to spine events play a less
  important role in the summary and will be
  selected only if the depth of linking is lower
  than a set threshold

13
(No Transcript)
14
Textual summaries

• Three steps
• Content selection
• - Select important events to convey in a summary
• Text structuring
• - Having a pool of events, arrange them in a
  coherent way to convey a certain viewpoint
• Realisation
• realise the defined text structure as paragraphs,
  sentences, phrases
• format according to the summary type

15
Text structuring

• The discourse structure is governed by two types
  of information
• relations present in the chronicle (19 different
  types of relations), which can be
• Attributive Problem has_locus Locus
• no discourse role, but impose local constraints
  (phrase expressed as a possessive, in the same
  sentence)
• Rhetorical Problem caused_by Intervention
• Define the discourse relation to be used (e.g.,
  causality, consequence, elaboration)
• the type of summary
• Longitudinal events occurring in the same week
  should be grouped together and further grouped
  into years
• Logical arrange chronologically and then group
  similar events (e.g., liver panels, screening
  consults)

16
Textual summaries

• Three steps
• Content selection
• - Select important events to convey in a summary
• Text structuring
• - Having a pool of events, arrange them in a
  coherent way to convey a certain viewpoint
• Realisation
• realise the defined text structure as paragraphs,
  sentences, phrases
• format according to the summary type

17
Realisation

• Aggregation
• Enlargement of the liver No enlargement of the
  spleen
• gt Enlargement of the liver but not of the spleen
• Ellipsis
• Examination of the left breast revealed no
  recurrent cancer in the left breast gt
• Examination of the left breast revealed no
  recurrent cancer
• Syntactic realisation
• Simple phrase structure grammar
• Spine events dictate which facts are most salient
  (i.e., will be in a more salient position in a
  sentence)

18
(No Transcript)
19
Radiotherapy on the breast is initiated to treat
cancer in the breast. A first radiotherapy cycle
is performed. The radiotherapy causes skin
ulceration. The patient receives hyperbaric
oxygenation to treat the ulcer.
20
A mammogram is performed because of pain in the
left breast, which identifies a lump in the
breast. A biopsy of the lump identifies cancer in
the left breast.
21
Typical output of the NL generator
Long chronological report

• Year 1
• Week 0
• A mammography screening was scheduled at the
  clinic.
• Week 1
• Primary cancer of the right breast
  histopathology invasive tubular adenocarcinoma.
• YEAR 2Week 131
• Xray revealed no cancer of the right breast.
• YEAR 5Week 287
• Xray revealed no cancer of the right breast.
• YEAR 8Week 443
• Xray revealed cancer of the right breast.
• Week 446
• Examination (indicated by primary cancer of the
  right breast) revealed no enlargement of the
  liver or of the spleen, no recurrent cancer of
  the right breast and no lymphadenopathy of the
  right axillary lymphnodes.
• Testing (indicated by primary cancer of the right
  breast) revealed no abnormality of the
  haemoglobin concentration and no abnormality of
  the leucocyte count.
• An Xray (indicated by primary cancer of the right
  breast) was performed.
• Very high level of the ESR concentration.
• Very high level of the Creatinine concentration.
• Very high level of the Alkaline Phosphatase
  concentration.
• Very high level of the Bilirubin concentration.

22
Typical output of the NL generator
Compact reports
Focus on Problems In week 0, the patient is
diagnosed with primary cancer of the right
breast, histopathology invasive tubular
adenocarcinoma. In weeks 131 and 287 Xray
revealed no cancer of the right breast. In week
446, there was no enlargement of the liver or of
the spleen, no recurrent cancer of the right
breast and no lymphadenopathy of the right
axillary lymphnodes revealed by examination.
There was no abnormality of the haemoglobin
concentration or of the leucocyte count, no
abnormality of the platelet count, very high
level of the GGT concentration, of the Bilirubin
concentration, of the Alkaline Phosphatase
concentration, of the Creatinine concentration or
of the ESR concentration. In week 449, excision
biopsy revealed no metastatic lymphnode count of
the right axilla. Histopathology revealed primary
cancer of the right breast. Lumpectomy was
performed on the right breast. Hormone
anatagonist therapy was initiated to treat
primary cancer of the right breast. In weeks
457 to 737, there was no enlargement of the liver
or of the spleen, no recurrent cancer of the
right breast and no lymphadenopathy of the right
axillary lymphnodes. There was no abnormality of
the haemoglobin concentration or of the leucocyte
count, no abnormality of the platelet count, very
high level of the GGT concentration, of the
Bilirubin concentration, of the Alkaline
Phosphatase concentration, of the Creatinine
concentration and of the ESR concentration. In
weeks 457 to 893, Xray revealed no cancer of the
right breast.
Focus on Interventions In week 0, the patient
is diagnosed with primary cancer of the right
breast, histopathology invasive tubular
adenocarcinoma. In week 449, excision biopsy
revealed no metastatic lymphnode count of the
right axilla. Histopathology revealed primary
cancer of the right breast. Lumpectomy was
performed on the right breast. Hormone
anatagonist therapy was started to treat primary
cancer of the right breast.
Focus on Investigations In week 0, the patient is
diagnosed with primary cancer of the right
breast, histopathology invasive tubular
adenocarcinoma. In weeks 131 and 287 Xray
revealed no cancer of the right breast. In week
446, examinations revealed no enlargement of the
liver or of the spleen, no recurrent cancer of
the right breast and no lymphadenopathy of the
right axillary lymphnodes. Testing revealed no
abnormality of the haemoglobin concentration or
of the leucocyte count, no abnormality of the
platelet count, very high level of the GGT
concentration, of the Bilirubin concentration, of
the Alkaline Phosphatase concentration, of the
Creatinine concentration or of the ESR
concentration. In week 449, excision biopsy
revealed no metastatic lymphnode count of the
right axilla. Histopathology revealed primary
cancer of the right breast. In weeks 457 to
737, examinations revealed no enlargement of the
liver or of the spleen, no recurrent cancer of
the right breast and no lymphadenopathy of the
right axillary lymphnodes. Testing revealed no
abnormality of the haemoglobin concentration or
of the leucocyte count, no abnormality of the
platelet count, very high level of the GGT
concentration, of the Bilirubin concentration, of
the Alkaline Phosphatase concentration, of the
Creatinine concentration and of the ESR
concentration. In weeks 457 to 893, Xray
revealed no cancer of the right breast
23
Visual navigator

• A single screen high level overview of a
  patients medical history
• Various types of facts are displayed along
  timelines and colour-coded
• User interacts with the visualiser by selecting
  either
• Individual facts (e.g., a specific instance of
  radiotherapy)
• Fact categories (e.g., Treatments)
• Groups of facts
• Each selection is reflected in a generated
  textual summary build around spine events
  identified through the user selection
• Relations between facts can be highlighted on
  demand
• Complex facts can be further delved into,
  revealing component facts
• Charts displaying the trend of numerical values
  are available for data that is suitable for such
  representation (i.e., blood test results)

24
Visual navigator
25
Conclusions

• A method of presenting complex clinical data for
  use in medical care, using a combination of text
  and graphics
• Unlike previous visualisation tools for medical
  data, our system works on a much richer input
  the CLEF chronicle
• No complex domain knowledge required
• No complex system of data mining and inferences