The Word Processor as a Data Collector for Quality Assurance in Anatomic Pathology Reports: Tracking the Use of a Text-Insertion Feature
Martin C Chang ; Brigham and Women's Hospital;
Content:
Anatomic pathology reports are documents generated with word processor software, and integrated into the laboratory information system. Commonly used words or phrases may be inserted into the word processor document using abbreviations, with the full text being subsequently substituted automatically. The purpose of text insertion is to reduce typing and errors in transcription.
Technology:
At Brigham and Women's Hospital, the Anatomic Pathology Information System (PowerPath 8.3, IMPAC, Sunnyvale, CA) includes a built-in software interface to a word processor (Word 2003, Microsoft, Redmond, WA), in which all anatomic pathology reports are generated. The text insertion feature and the recording of its use are built into this software interface.
Design:
Each instance of text insertion using an abbreviation is recorded by the word processor software interface. We analyzed 1000 consecutive text insertion events, and compared them to the final reports in which they were used, with particular attention to erroneous uses.
Results:
The text insertion events varied from the replacement of a word to the construction of an entire synaptic report. Both simple text and other retrievable information were inserted using the function. In most cases, the function was used appropriately. Three main types of erroneous use were apparent: 1) instances of text-insertion that were subsequently deleted from the final report, 2) incomplete use, in which retrievable diagnostic information was omitted, 3) failure to use the text insertion function in a field required for subsequent retrieval. All types of errors were more prevalent in longer synoptic reports.
Conclusion:
The use of automatic text-insertion in creating anatomic pathology reports saves time and allows important diagnostic data/statistics to be retrieved more readily. In longer reports, errors may result from unclear text or illegibility in the draft diagnosis, and/or errors in transcription from this draft. Efforts to reduce errors should therefore aim for improvements in the transcription process.
