Medical Informatics in a United and Healthy Europe K.-P. Adlassnig et al. (Eds.) IOS Press, 2009 2009 European Federation for Medical Informatics. All rights reserved. doi:10.3233/978-1-60750-044-5-527 527 Does Computerised Provider Order Entry Reduce Test Turnaround Times? A Beforeand-After Study at Four Hospitals Johanna I. WESTBROOK a,1, Andrew GEORGIOU a, Mary LAM b a Health Informatics Research & Evaluation Unit, Faculty of Health Sciences, The University of Sydney, Australia b Health Informatics Discipline, Faculty of Health Sciences, The University of Sydney, Australia Abstract. Few multi-centre studies of the impact of computerised provider order entry (CPOE) systems on health care efficiency and effectiveness exist. Further, demonstrating a link between system use and improvements in patient outcomes is challenging. An often neglected step is to characterise the nature of the problem prior to CPOE introduction to ensure that the problem being addressed has a demonstrated impact on the outcome of interest. We undertook a two-staged project to i) investigate the link between test turnaround time (TAT) and length of stay for emergency department patients prior to CPOE; ii) to measure the impact of CPOE on TAT in four Australian hospitals to examine the consistency of findings. We found TAT is a significant contributor to length of stay. All four hospitals experienced a significant reduction in TAT following CPOE. This study presents evidence that TAT is directly related to length of stay and that CPOE systems are effective at reducing TAT. These results add weight to the hypothesis that the introduction of CPOE may positively impact upon patient outcomes. Keywords. pathology, computerised provider order entry, length of stay, emergency department, turnaround times, efficiency, outcomes, computerised medical records 1. Introduction A criticism of the evidence of the impact of computerised provider order entry (CPOE) systems on improving health care efficiency and effectiveness is the lack of multicentre studies to provide an indication of the generalisability of findings [1]. Relatively little research attention has been placed on measuring the potential benefits of CPOE for pathology services in delivering direct improvements in the outcomes of patients [2]. Health systems have invested in CPOE systems with the expectation that they will result in faster delivery of test results to clinicians which will increase the likelihood of quicker diagnosis and treatment decisions, in turn leading to improved outcomes, including shorter lengths of stay. Studies evaluating the impact of CPOE on pathology turnaround times have shown improvements for specific wards or units and for single organisations, but we have 1 Corresponding Author: Prof. J. Westbrook, Health Informatics Research & Evaluation Unit, The University of Sydney, PO Box 170, Lidcombe 1825, NSW, Australia; E-mail: J.Westbrook@usyd.edu.au.
528 J.I. Westbrook et al. / Does Computerised Provider Order Entry Reduce Test Turnaround Times? found no multi-centre studies which demonstrate the applicability of these findings across hospitals of different types in different geographic locations. Reductions in test turnaround times of around 20% have been found in individual hospitals in the US [3] and Australia [4] and smaller units within hospitals have also reported declines in turnaround times following CPOE [5 7]. A further challenge for researchers has been to demonstrate a link between the implementation of CPOE systems and improvements in patient outcome indicators. The research evidence of such links is modest. At best there has been a tendency to extrapolate demonstrated improvements in process indicators. For example, improvements in TAT are assumed to result in faster diagnosis, treatment and outcomes. A major gap in this hypothesis is a demonstrated link between TAT and specific outcome indicators prior to CPOE [4]. Thus in order for CPOE to be seen as an effective solution we need to have a better understanding of the existing problem in the paper-based world. In essence, what is the evidence that test turnaround times contribute to patient outcomes? If there is no association prior to CPOE it is unlikely that CPOE will deliver upon promises of improved outcomes. Improvements in outcomes most likely to flow from decreased test turnaround times are within critical care environments. Here time is central to decision making and fast test results are most likely to influence the management and potential outcomes of patients. Length of stay is one such outcome indicator. We undertook a two-staged project to tackle both these research challenges. First, we aimed to establish whether there is evidence of a relationship between test turnaround times and patients lengths of stays in an emergency department prior to the introduction of a CPOE system. Second, we conducted a before and after study of the impact of the introduction of CPOE on pathology TAT in four Australian hospitals of different size and in different locations. The aim was to determine whether there was consistency in the association between CPOE introduction and reductions in test TAT. 2. Materials and Methods 2.1. Stage 1 Relationship between TAT and Length of Stay in an Emergency Department Stage 1 was undertaken at a major Sydney teaching hospital which operates one of the largest emergency departments in the country with around 50,000 attendances each year. Test turnaround data for the months of August and September 2005 were extracted from the hospital s laboratory system. These data were linked [8] to the emergency department data for the same period to allow examination of the association between TAT and length of stay using regression analyses. Table 3 shows the variables in the model. Table 2 lists definitions used in the research. 2.2. Stage 2 Changes in TAT Following the Introduction of CPOE Data relating to test TAT from four Sydney hospitals for the months of August and September in the year before the introduction of CPOE (Cerner Millenium PowerChart v2001.01 system) and for the same period in the two subsequent years were obtained. The proportions of test orders delivered using the CPOE system for each of the after
J.I. Westbrook et al. / Does Computerised Provider Order Entry Reduce Test Turnaround Times? 529 periods were calculated. All hospitals used paper-based ordering systems in 2005. Table 1 lists details of each of the hospital study sites. Comparisons between total turnaround times were made using the Kruskal-Wallis one-way analysis of variance statistic which is a non-parametric method (which does not assume a normal distribution) for testing equality of population medians among groups. Table 1. Details of hospital study sites Hospital Type Number of beds Hospital A Major acute metropolitan teaching hospital 660 Hospital B Major metropolitan acute hospital 454 Hospital C Acute general hospital 260 Hospital D Acute district hospital 200 To determine the extent to which the profile of hospital patients may have changed over the three year study period we examined the clinical profile of the patient populations using Diagnosis Related Groups (DRGs). We calculated the number and average DRG cost weights for all patients for each hospital for each of the three study years. These data provided an indication of the clinical caseload for each hospital. Test turnaround time Emergency department length of stay Table 2. Definitions used in the research Time from blood collection was made to the time the results were available Time from patient arrival in the emergency department to the time discharged from the hospital or admitted to a ward. 3. Results 3.1. Relationship between TAT and Length of Stay in an Emergency Department Regression analysis demonstrated that test TAT was a significant factor contributing to patients length of stay in the emergency department. The model accounted for 25.4% of the variance (Adj. R 2 =0.254) (Table 3). Table 3. Results of regression analysis B SE t p. Age 0.003 0.000 6.766 <0.001 Total No. Tests 0.019 0.002 11.874 <0.001 Triage1-0.422 0.076-5.551 <0.001 Triage2 0.031 0.034 0.902 0.367 Triage3 0.085 0.025 3.340 0.001 TAT 0.184 0.019 9.764 <0.001 Discharged -0.419 0.020-21.359 <0.001 Transferred -0.020 0.097-0.201 0.841 Discharge Other 0.256 0.081 3.151 0.002 3.2. Changes in Test Turnaround Time by Hospital before and after CPOE For each hospital we found a significant decrease in the median TAT following the introduction of the CPOE system (Table 4). In the second year after CPOE introduction, the proportion of CPOE orders increased and the TAT continued to decline.
530 J.I. Westbrook et al. / Does Computerised Provider Order Entry Reduce Test Turnaround Times? Differences between each year (i.e., 2005/2006; 2006/2007; 2005/2007) were tested and for each hospital site there were significant declines (p<0.001). Table 4. Comparison of TAT in minutes by hospital before and after implementation of CPOE 2005 2006 2007 Kruskal-Wallis Hospital A Median TAT in minutes 77 68 66 P<0.001 Number of tests 362,728 396,878 437,486 % tests ordered using CPOE 75% 80% Hospital B Median TAT in minutes 145 129 108 P<0.001 Number of tests 186,062 178,361 179,196 % tests ordered using CPOE 0 44% 57% Hospital C Median TAT in minutes 138 135 113 Number of tests 115,936 110,623 141,994 % tests ordered using CPOE 29 38% 53% Hospital D Median TAT in minutes 141 139 128 Number of tests 85,506 79,349 94,582 % tests ordered using CPOE 56 71% 74% P<0.001 P<0.001 For all hospitals except Hospital B the number of tests performed between 2005 and 2007 substantially increased. Results from the DRG analyses showed that the average DRG cost weight for this time also increased considerably for all hospitals (Table 5). Table 5. Average Diagnosis Related Group (DRG) cost weight by hospital by year 2005 2006 2007 Average Average DRG cost Number DRG cost Number weight patients weight patients Average DRG cost weight Number patients Hospital A 6,851 1.47 7,834 1.33 7,209 1.56 Hospital B 5,013 1.03 4,992 1.03 4,881 1.23 Hospital C 5,012 0.78 5,552 0.78 5,249 0.98 Hospital D 2,751 0.86 3,054 0.80 3,127 1.08 4. Discussion We found that TAT is a significant contributor to emergency department length of stay which confirms results from a US study [9]. This provides evidence that reducing TAT via the introduction of a CPOE may produce improvements in these patients outcomes in terms of reduced lengths of stay. Thus we have evidence of a mechanism by which CPOE may result in improved outcomes. Future research should include demonstration of change in lengths of stay following CPOE introduction. Our study of TAT following the introduction of CPOE at four hospitals showed significant declines for all in the first and second years after system implementation. The increased proportions of CPOE orders in the second follow-up year may have contributed to the continued improvement in TAT. This suggests consistency in the effects of CPOE on TAT across hospitals of different type and in different geographic locations. Further it presents evidence of the sustained improvements CPOE may bring to TAT, only previously demonstrated at a single hospital site [10].
J.I. Westbrook et al. / Does Computerised Provider Order Entry Reduce Test Turnaround Times? 531 5. Conclusion This study presents evidence that TAT is directly related to patients lengths of stay and that CPOE systems are effective at reducing TAT. These results add weight to the hypothesis that the introduction of CPOE may positively impact upon patient outcomes. Acknowledgements. The research was funded by a Commonwealth Department of Health and Ageing Quality Use of Pathology Program (QUPP) grant and Australia Research Council Linkage grant LP0989144. References [1] Chaudhry, B. et al. (2006) Systematic review: Impact of health information technology on quality, efficiency, and costs of medical care. Annals of Internal Medicine 144:742 752. [2] Georgiou, A. et al. (2007) The impact of computerised physician order entry systems on pathology services: a systematic review. International Journal of Medical Informatics 76(7):514 529. [3] Guss, D.A., Chan, T.C., Killeen, J.P. (2008) The impact of a pneumatic tube and computerized physician order management on laboratory turnaround time. Annals of Emergency Medicine 51(2):181 185. [4] Westbrook, J. et al. (2006) Computerised pathology test order entry reduced laboratory turnaround times and influences tests ordered by hospital clinicians: a controlled before and after study. Journal of Clinical Pathology 59(5):533 536. [5] Mekhjian, H.S. et al. (2002) Immediate benefits realized following implementation of physician order entry at an academic medical center. Journal of the American Medical Informatics Association 9(5):529 539. [6] Ostbye, T. et al. (1997) Introducing a module for laboratory test order entry and reporting of results at a hospital ward: An evaluation study using a multi-method approach. Journal of Medical Systems 21(2):107 117. [7] Thompson, W. et al. (2004) Computerized physician order entry of diagnostic tests in an intensive care unit is associated with improved timeliness of service. Critical Care Medicine 32(6):1306 1309. [8] Lam, M.A. et al. (2008) Record linkage A method to generate an enriched dataset for the study of CPOE systems on test turnaround time. In Grain, H. (Ed.) Proceedings of the 16th National Health Informatics Conference 2008, Health Informatics Society of Australia, Melbourne, Australia, 133 138. [9] Holland, L.L., Smith, L.L., Blick, K.E. (2005) Reducing laboratory turnaround time outliers can reduce emergency department patient length of stay: An 11-hospital study. American Journal of Clinical Pathology 124(5):672 674. [10] Westbrook, J.I. et al. (2008) Computerised order entry systems: Sustained impact on laboratory efficiency and mortality rates? In Andersen, S.K. et al. (Eds.) ehealth Beyond the Horizon Get IT There, Proceedings of the XXI Medical Informatics Europe (MIE 2008), IOS Press, Amsterdam, 345 350.