Structured data entry conformant to S3-guidelines for breast cancer documentation

There is an increasing demand for several secondary usages of existing electronic patient data like sharing and reuse, data analysis and decision support. While there are some reasons for the use of free text from the perspective of involved human beings like acceptance and expressiveness, free text cannot be processed and interpreted by computers suitably. This is also true for the macroscopic and microscopic documentation of findings in pathology that are mainly created using free-text dictations. With the example of breast cancer diagnosis in this project the following questions are addressed:

  1. Empirical analysis of shortcomings of existing free-text findings of breast cancer diagnostics in pathology with respect to the so-called S3-guideline (see below), e.g. missing or incorrect information and inconsistency.
  2. Implementation of a software system PathIS for the structured documentation of data concerning the patient, case, sender, sending (material) as well as result, including report generation in various formats.
  3. Implementation of a grammar-based structured data entry (SDE) of findings. Furthermore relevant codes like ICD-O, Grading and B-Classification are derived semi-automatically and free text for the corresponding findings sections are generated.
  4. Evaluation of the PathIS system in terms of user acceptance, documentation quality and testing of the integration of the SDE component in the existing commercial Pathology Information System.

There are already first results of the empirical analysis of free text findings, e.g. missing data of microcalcifications  (macroscopic section) or non-uniform specifications for the intraductal component in invasive ductal breast carcinomas (microscopic section). The implemented software PathIS (Java and MySQL database) does not replace the commercial system. Instead it allows the demonstration of the “intelligent” structured data entry of findings that we are mainly interested in this project. It should allow to evaluate the SDE approach (third step) in the fourth step as realistic as possible.

The central purpose of this project is to

 a) define the S3-guidelines-compliant document structures and classifications from a medical point of view and
b) to implement grammar-based input wizards to capture the required characteristics and values with respect to defined dependencies, and generate codes and free text based on the grammar entries (engrams), see examples of screenshots for entries for “submission”, “macroscopic and microscopic data entry” and “code mapping”

Notes

Selected Publications

  1. Büchler A, Graeve L, Ingenerf J, Thorns C. Leitlinien-basierte strukturierte Dokumentation in der Pathologie. In: 58. GMDS-Jahrestagung 2013 in Lübeck. (LINK)
  2. Ingenerf J. Computergestützte strukturierte Befundung am Beispiel der Wunddokumentation. WundM 2009; 3[6]:264-268. (LINK1, LINK2)
  3. Schoech W, Hatje H, Ingenerf J. DICOM Structured Reporting in der Pathologie. In: 53. GMDS-Jahrestagung 2008 in Stuttgart. (LINK)

Projekt Team

Astrid Büchler (Medizin-Doktorandin)
Lars Graeve (Informatik-Diplomarbeit)
Dipl.-Inform. Winfried Schoech
PD Dr. rer. nat. Josef Ingenerf
Dipl.-Inform. Dr. med. Jan-Hinrich Wrage

Partners

PD Dr. med. Christoph Thorns, Institut für Pathologie, Universität zu Lübeck
Harald Hatje, Institut für Pathologie, Universität zu Lübeck