Shannon Haymond (Presenter)
Northwestern University
Authorship: Shannon Haymond and Faye Vicente
Ann & Robert H. Lurie Children's Hospital of Chicago
| Short Abstract Though automation of sample preparation and analyses are becoming more common in clinical mass spectrometry laboratories, much of the data review and transfer are still performed manually. This is time-consuming and error prone, easily exceeding the limits of human capacity for effective parallel processing. Other problems exist when sample data fields and results are manually transcribed, some repeatedly across multiple points in the testing process. Therefore, systematic, automated methods for reviewing and transferring data throughout the testing process have a tremendous effect on operational efficiency and quality. This presentation will describe our efforts to automate data flow in our LC-MS/MS lab from test order to result. Approaches for the various points of automation that we and others have used will be described with discussion of associated challenges. |
Long Abstract
Introduction:
Though automation of sample preparation and analyses are becoming more common in clinical mass spectrometry laboratories, much of the data review and transfer are still performed manually. This is time-consuming and error prone, easily exceeding the limits of human capacity for effective parallel processing. For example, even in a lab like ours with relatively small testing volumes, technologists may manually review upwards of 200,000 peaks per year. Each result may be associated with more than 10 simultaneous assessments for quality. In fact, a large percentage of these peaks and results meet all defined quality criteria and require no further intervention. Other problems exist when sample data fields and results are manually transcribed, some repeatedly across multiple points in the testing process. Manual steps for transcription and decision making introduce error and lead to variability. Therefore, systematic, automated methods for reviewing and transferring data throughout the testing process have a tremendous effect on operational efficiency and quality.
Methods:
We implemented lab order and result interfaces, data file transfer protocols, and data review software to automate steps in clinical mass spectrometry lab workflows. We performed time studies for the manual and automated workflows to determine effects on technologist time. Patient results generated using automated and manual workflows were compared for accuracy and data quality and review parameters.
Results:
Implementing a process for automating only the data review and resulting steps for immunosuppressant drugs resulted in an auto-verification rate of 96%. Data is automatically and systematically reviewed for 24 quality flags versus 16 for the manual process, which relied largely on visual and memory cues. Flags firing most often are those involving parameters related to concentration, peak area, and peak quality. The automated process allows for greater visibility of manual corrections of raw data that affected results. Comparison of 4-month windows before and after implementation showed a 40% decrease in the number of manual interventions prior to data release. Further efficiency is achieved when the data transfer is automated from lab order to result, as we have done for vitamin D testing. Per batch of 84 samples, we realize technologist time-savings of approximately 2 hours related to steps involving data. This includes time saved during the sample preparation and analysis steps requiring repeated tracking or entry/transcription of sample information and during data review, manual entry and verification procedures prior to result release. Data entry errors are reduced since manual transfer is replaced with barcode scanning or automated data file creation and movement protocols. Data and metadata, once digitized versus manually recorded, are also increasingly available for quality assurance and process monitoring and improvement efforts.
Conclusions:
Automation of data transfer and review is not yet routine practice in clinical mass spectrometry laboratories, due to a number of challenges. However, the gains in efficiency and quality through implementing automated data handling procedures should justify a shift away from manual processes. Data review and release is increasingly being automated with various approaches available. But more work is needed for LC-MS/MS labs to match the high levels of automation common to general clinical chemistry labs where information from lab order to result is electronically transferred throughout the testing process.
References & Acknowledgements:
| Description | Y/N | Source |
| Grants | no | |
| Salary | no | |
| Board Member | no | |
| Stock | no | |
| Expenses | no |
IP Royalty: no
| Planning to mention or discuss specific products or technology of the company(ies) listed above: | no |