Cancer Institute Gustave-Roussy
In 1999, the Cancer Institute Gustave-Roussy, Europe’s largest centre devoted to the fight against cancer, embarked on an innovative and far-reaching reorganisation programme. It was intended to improve the way people were received from the moment they entered the hospital and to transform the management of their diagnoses and treatment. The new arrangements had to refocus completely on the patients, their needs and expectations. These were profound changes that required the complete involvement and cooperation of technicians and medical laboratory staff to be successful.
The automation of laboratory processes was integral to achieving our objectives. Centralising the hospital’s laboratory and pathology services has enabled the hospital to make better use of its funds as well as improve the total turnaround time (TAT) and the consistency of delivery of the results. In one year, we have seen operating costs in immunoassay and biochemistry decrease by almost 10%. Total TAT is down substantially, particularly for immunoassay tests (see Table 1). The gap between the fastest delivery of results and slowest has also been cut – and for cancer patients that is particularly crucial.
Need for integrated working
If we look at how the services were set up before automation we can see that the reforms were well overdue. The medical testing and pathology laboratories were spread out on different levels. When samples were first taken they had to be transported 12 floors to be processed. Consulting facilities and operating theatres were scattered throughout the old building.
While dealing with the design of the new building, we also had to assess the type of analysers that might be needed and how the premises could accommodate them. The laboratory facilities had initially been designed to accommodate hospital wards and were organised with separate areas for each specialty – general biochemistry, specialist biochemistry, haematology and coagulation, microbiology, and genetics. Each specialism had its own reception and sample-taking procedure, one or more centrifuges and systems of varying efficiency for moving samples. They worked separately from each other, with little coordination or sharing of resources.
In the new four-storey building the consulting area is still on the ground floor, but with the laboratory complex immediately underneath at lower-ground level. Intensive care and high-dependency units are on the first floor, with the operating theatres on the second floor. Having consulting areas and operating theatres close to the laboratories made it easier to integrate and transport the samples into the laboratory with a high-speed shoot.
Total savings benefit
The benefits include saving time, making best use of human resources, improving sample quality and achieving a financial saving. It is also a long-term solution that requires in-depth consideration to avoid mistakes. Automating the preanalytical stages, for example, is a complex matter that requires detailed quantitative studies of flows, yields, return on investment and technical reasons for deciding on facilities such as centrifuging and tracking. This is why our first task was to make an inventory of the volume of laboratory work, which included looking at the number of specimens processed, the range of tests, how many processes were done by hand and the length of time involved.
With an average of 900 specimens processed per day (two-thirds requiring centrifugation), a crucial decision was whether or not the preanalytical stage should be automated. If yes, we had to decide what type of automated system – with or without sample tracking, with or without centrifuging – would be most appropriate. One thing was definite: there had to be provision for centrifugation. This was illustrated perfectly by Dr Ralph Dadoun in his article in the Medical Laboratory Observer, talking about the installation of a preanalytical system in the laboratories of his Montreal hospital.(1)
Advantage of centrifugation
One advantage is that automated centrifugation significantly reduces what Dr Dadoun calls the “transition times”. These are the periods between the technical stages such as the time between receiving a specimen and it being centrifuged, or between the end of centrifuging and loading the specimen into the analyser. In most cases, these periods depend on technical staff being available at the exact moment something happens. In our own preliminary workflow study we noticed that transition times were high and could, in extreme cases – notably immediately before and after centrifuging – double the time taken by preanalytical handling.
For automation to succeed it had to offer consistent quality of sample handling and reliability, replacing over 60% of the manual tasks that the staff find repetitive and time- consuming. We wanted to reduce the number of specimens taken from a single patient and group as many investigations as possible on one automated machine. It was important to reduce the time taken to obtain and communicate results to the clinician and, overall, make significant reductions in the cost of the lab services. And we had to have the support of our staff if they were to adopt this new way of working – multitasking.
Traceability of samples
Some suppliers of preanalytical systems also offer the flexibility of connecting them to automated analysers from other firms, if those machines have sampling systems. Others only allow connection to their own automated machines. The project group decided on the Power Processor (Beckman Coulter) with tracking, which allowed more flexibility in connecting to analysers. Using the Power Processor for all the specimens arriving at the laboratory allows comprehensive recording of the time taken to receive and handle them. We added an interface with the Remisol 2000 technical validation system to enable us to work out the analysis times. The coagulation and chemistry analysers have closed-tube sampling technology, so their centrifuged tubes do not need to be decapped before they are sorted.
Use of the preanalytical module also makes it possible to deal more quickly with urgent results, and the system can handle virtually all of the samples. It is possible to have dedicated locations reserved for specimens that need to be processed urgently, and to programme the analyser to carry out the investigation urgently.
Automating other sections of the lab – haematology, clinical chemistry and immunoassay – also involved Beckman Coulter analysers. We selected two SYNCHRON analysers, an LX 20PRO clinical chemistry system plus an LXi 725 – both with closed-tube sampling technology – and a UniCel DxI 800 immunoassay system. In haematology we used an LH 755 analyser with slide-making module, plus an HmX analyser with autoloader for backup (both Beckman Coulter). Cost constraints meant we had to choose initially to connect only one – the immunoanalyser – though this will shortly be extended to the other automated systems.
Priority to reduce transition time
This was supported by our workflow study that had shown we could make significant reductions in TAT, in this specialty, because the maximum transition time was particularly varied. It was a particular priority to reduce the TAT in immunoassay because of a new initiative to provide “diagnosis in a day” for certain cancers such as thyroid and breast. Since tumour markers detected in a laboratory were one of the indicators aiding clinical diagnosis, we had to be able to reduce the time taken to communicate the results.
In relation to total laboratory tests, we noted an unchanged level of activity compared with previous years. Our reimbursement for the tests we ran in 2003 was 29.9 million “B” (“B” refers to the level of reimbursement from the French government, and there is a set amount in euros for each test) and 30.1 million “B” in 2005 based on around 750 test requests per day. Of course, each request is not for a single test but for a number of individual tests. This constant level relates in particular to the activities affected by the introduction of automation. On the other hand, the number of technical staff needed to carry out this volume of work has fallen significantly. Overall, we have reduced technical staff by 25%, with 32% specifically attributable to automation. By reducing the effective number of hours per year undertaken by the technical staff we have been able to redeploy them, strengthening other areas or establishing new activities such as an oncogenetics service (requiring five fulltime-equivalents [FTEs]).
Standardisation of analysis procedures
Automating the preanalytical and analytical steps is also a major advantage for standardising centrifuging procedures. Refrigerated centrifuging (standardised at 12–14°C) for 10 minutes at 3,000rpm can accommodate all the preanalytical requirements in biochemistry, immunoanalysis and basic coagulation. A double centrifugation will be required for any specialised coagulation. Automating the preparation of slides significantly reduces the manual steps from 667 procedures to 65. When combined with data communication (using DiffPad) of preliminary results and returning data-processed results from the manual count, automation produces a mean time of 56 minutes for communication of the result, with a minimum of 36 minutes and a maximum of 76 minutes (based on a study of 86 specimens).
Improving patient care
Overall, we have seen automation help the laboratory work proceed more efficiently. Standardising the preanalytical and analytical procedures has given us a more stable range of TATs. Another benefit is the reduced number of samples needed for a comprehensive laboratory assessment. Automation also improves the quality of the results. The traceability that can be provided by the preanalytical information system and by the automated analyser itself makes it possible to demonstrate to the hospital managers how we are able to meet our targets. But clearly the main benefits go to our patients. By reducing the waiting time for results, not only do we help to reduce the stress at a very difficult time, but patients are able to start lifesaving treatment as early as possible.
- Dadoun R. Case study: automation’s impact on productivity and turnaround time. MLO Med Lab Obs 2002;34:36-8.