Akershus University Hospital introduced Information and Communication Techno-logy (ICT) to its radiology department in 2005. Both images and reports were stored and communicated electronically instead of as printed film and paper. The images and reports were also made available to clinicians directly from the Electronic Patient Record (EPR).
The objective of this study was to retrospectively assess whether the introduction of ICT improved diagnostic imaging and health care. The objective was addressed by investigating whether the introduction of ICT made radiology reports available sooner to clinicians, whether they read them sooner, whether this had an impact on the length of patient hospital stay, and whether any improvement in reporting was achieved without reducing the diagnostic accuracy.
The basic design of this study was a before-after study using two cross-sectional data collections. The establishment of hypotheses was in part assisted by a data splitting method. Most analyses were based on data retrieved retrospectively from the hospital infor-mation systems; the Radiology Information System (RIS), The Picture Archiving and Communication System (PACS) and the EPR. These data were partly recorded by health care professionals as part of their daily work, and partly created by automatic logging of their activities. Person-identifiable attributes were removed for both pa-tients and health care professionals before the statistical analysis. Supplementary data was collected manually from work lists and routine descriptions. Diagnostic accuracy was addressed by comparing a retrospective classification of lesions reported in the original reports with lesions identified in an independent re-analysis of the images. The data were analysed using the two-sided non-parametric Mann-Whitney U-test for ordinal and the T- test for nominal data.
The median report turnaround time (RTAT) – the time from the images were acquired until they were reported - was initially reduced by 84% for the preliminary version and by 44% for the final version of the reports. Over the observation period, the me-dian RTAT increased slightly for preliminary reports, and was reversed almost back to the pre-ICT level for final reports. However, the percentage of preliminary reports available for the clinical afternoon round increased over the observation period. Radi-ologists used the flexibility of the system to give priority to preliminary ultrasound (US) and all Computed Tomography (CT) reports. Both preliminary and final reports were immediately sent to the EPR. The median time until the final reports were opened by a clinician was 2.8 to 3.9 hours. The use of final reports did not vary much over the observation period. In total, 88% of the final reports had been opened 4 weeks after they became available in the EPR. For prelimi-nary reports, the median time until they were opened was 40 to 50 minutes. Only 42% of them were read. Preliminary CT and US reports were opened sooner than CR re-ports. There was no general reduction in length of patient hospital stay (LOS) after the ICT introduction. There was, however, a reduction in LOS for patients with one or more CT scans, from 5.3 to 3.9 days. This reduction was significant both in itself and rela-tive to the non-CT group. It has been feared that more lesions would be missed than when images were printed on film. Our study did not identify any such deterioration. On the contrary, when both certain and uncertain findings were included, the detection sensitivity was actually improved.
The introduction of ICT led to reports being available for and read by clinicians earli-er than before, however not all effects proved sustainable. We also found that radiolo-gists used the flexibility offered by the system to give priority to certain report catego-ries. The study indicated that length of stay was reduced for patients that had CT scans during their stay. Diagnostic sensitivity of chest radiographs did not deteriorate. Our findings indicate that when ICT is introduced in the radiology department of a large hospital, a few improvements may follow. However, our findings also indicate that an ICT introduction may have an untapped potential, and that not all effects are neces-sarily sustained. We did not observe important adverse consequences of the ICT in-troduction.
List of papers
Articles I, II and IV are removed due to copyright restrictions.
I: Hurlen P, Østbye T, Borthne AS, Gulbrandsen P. Introducing PACS to the late majority. A longitudinal study. J Digit Imaging. 2010 Feb;23(1):87-94. Epub 2008 Nov 1.
II: Hurlen P, Østbye T, Borthne AS, Dahl FA, Gulbrandsen P. Do clinicians read our reports? Integrating the radiology information system with the electronic patient record: experiences from the first 2 years. Eur Radiol. 2009 Jan;19(1):31-6. Epub 2008 Aug 6.
III: Hurlen P, Østbye T, Borthne AS, Gulbrandsen P. Does improved access to diagnostic imaging results reduce hospital length of stay? A retrospective study. BMC Health Serv Res. 2010 Sep 6;10(1):262. Epub ahead of print
IV: Hurlen P, Borthne AS, Dahl FA, Østbye T, Gulbrandsen P. Does PACS improve diagnostic accuracy in chest radiograph interpretation in clinical practice? Eur J Radiol. 2010 Sep 30. [Epub ahead of print]