APRIL 2001
Web Technology and Laboratory Data
It was recognised when the Internet became a reality that it would affect the way in which many operations were carried out, but few could have imagined the far-reaching changes it would bring to almost every facet of everyday life. One of the most important changes has been in communications - instant e-mail and the ability to send images to any part of the world at the click of a mouse - has brought about a global neighbourhood.
These changes are beginning to impact on the relationship between laboratories and data management (DM) groups in the way they relate with each other. It has enabled barriers to be breached which years of paper transactions have never succeeded in doing. There has been a reluctance on the part of DM to get involved with the laboratory and in spite of the fact that the possibility of direct transfer of data has existed with central labs for some years many DM groups still elect to input data from hard copy rather than go through the process of setting up a system with the laboratory.
The argument is that the data from the CRFs are being input from hard copy so it is no extra burden to input the lab data. An added advantage of inputting from hard copy is that only those results which fit the data base are used and all the extra results which appear on the hard copy e.g. follow up tests not listed in the protocol but performed to evaluate an abnormal test, need not be entered. This is a very reasonable approach, and it is only when one considers that laboratory data can be as much as 70% of all the data generated in the study that one questions the effect that this approach has on the study timelines and the cost in terms of man-hours.
To bring about a change to this reasonable approach, DM and the laboratory must move closer to understanding each other.
Generation of Laboratory Data
One of the things which DM and laboratories have in common is the improvement in recent years in the equipment which each uses to carry out its task. In the laboratory now the biochemistry analysers can almost run themselves, have a wide range of checks of both quality and performance which they carry out automatically quite apart from the analysis of samples. In DM computers are faster and software more sophisticated - a long way from reading punched cards! Neither of these developments has moved these two groups any closer to each other, but with the advent of the Internet the possibility now exists for them to get to know each other in a non-threatening way.
Most of the laboratories analysing samples in clinical trials have never seen a study database and have no idea how it is formatted or what it looks for as a reasonable value to put in a specific field. The people in DM who deal with reference ranges will have had experience of the response of the laboratory when they ask for ranges for certain tests particularly those relating to urine deposits - the lab staff can't believe anyone would want such a thing.
To the DM people it seems ridiculous that a lab could do a test and not have a reference range for it. The fact is that laboratories mainly provide results to clinicians and even in a laboratory which only does clinical trials, the lab staff will have trained in a hospital environment. Clinicians understand these results often on a pattern recognition basis and don't ask for such reference ranges. The lab staff cannot understand the insistence of the DM group for an objective measure of the value.
Another issue is the laboratory changing a reference range during the study. Labs have been accustomed to changing methods and consequently reference ranges without consultation, and this has caused problems for data in pharmaceutical studies. It does not cause any problem for the clinician since he will look at the result compared with the reference range given beside the value. The important difference is that the clinician in a clinical environment is looking at the results to support a diagnosis or to make one. In clinical trials the patients are already diagnosed and the measurements are made to detect change which is why it is so important to control variables such as method of analysis. DM have not managed to get this concept over to the lab perhaps because they assume that that is what every lab user is doing. DM are recording results not interpreting them so they must have a system which allows interpretation to be made by a regulatory body against the appropriate range.
Laboratories could contribute signi-ficantly to the clarification of results if they knew what was required. One of the hardest things to control is the report given when a technologist looks down a microscope and describes what they see. Most labs don't bother to standardise comments for things such as urine deposits or blood films and both of these cause problems for DM. However there is no reason why this should not be standardised. DM and laboratory staff could usefully get together at the beginning of a study and discuss the protocol. The important thing would be for DM to explain to the lab staff the issues which cause problems such as urine deposits, text results where a numeric value was expected, comments such as insufficient, broken in transit etc etc. Everyone who deals with lab data in DM knows what these thorns are, but laboratory staff don't so there is no possibility of improvement. It may be that a particular lab cannot improve every aspect but they at least will understand the issues and do what they can. Having done this on a hands-on basis with one lab and understood the process in the lab, DM could carry out this negotiation for future studies with other labs by e-mail - perhaps with scanned in photographs of the people involved to make the whole thing more human - people are more likely to help other people rather than an e-mail and putting a face to an e-mail helps!
Use of Excel
Many DM groups have begun to communicate directly with labs electronically and one thing they ask for is a regular transmission of an Excel spreadsheet with all the patients who have been recruited to date listed by investigator and visit. This is a useful tracking tool and most laboratories can provide it without difficulty. Many of these DM groups do not go on to have the lab data by direct transfer - this is the only communication they make with the lab apart from requesting reference ranges for tests which don't normally have any!!
Use of the Web
The way the Web is being used has opened up possibilities for transfer of laboratory data. RDE, EDT and all the various acronyms which denote collection of data from investigator sites and transmitting it to a central database has implications for laboratory data. Why should the laboratory not be treated as an investigator site since it generates data exactly as a site does?
Just as the site downloads the CRF pages from the Web under secure password arrangements, the laboratory could download the lab pages and complete them. This would mean that professional lab personnel would interpret those text type results and additional tests and remove the onus from DM. Lab personnel would act as "investigators" in the sense that they would use their experience to decide whether the results of a non-protocol result had any value, where DM would have no option but to record it in the database.
A hard copy of the report could go to the investigator by post or courier but ideally he would receive an electronic copy in his mail box automatically - far more secure than the post office. The investigator could add comments as to the clinical significance of out of range results and DM would then have the complete record. One of the short-comings of using direct transfer of results from labs is that the lab doesn't see the clinical significance comments made by the investigator so DM has the task of adding these to the lab results. Using this system the comments and results would finish up together in an electronic format.
Lab as Investigator Site
In treating the lab as an investigator site there will be the need to monitor source data as with any other site. Audit by the QA team of the sponsor company will have assessed the ability of the lab to identify the sample correctly and transfer results from the analysers to the report. The additional audit for this proposed web-based system will be to check the way results are entered into the electronic report and the degree of validation which has been done. This would need someone who had a clinical understanding but also understood the use to which the data would be put - a data monitor in fact, a hybrid of a CRA and a member of the DM team.
Transfer Formats
Apart from the obvious advantage of using the web for transfer of lab results directly from the lab to the CRF pages there is the benefit that DM could extract the information they need from these pages as they do from the other electronic CRF pages. This would eliminate the need for transfer formats which are a nightmare for laboratories. Every sponsor company has its own format and every therapeutic group, sometimes every study within a therapeutic group, has a different format requirement. It is difficult for the lab to understand the lack of standardisation of transfer formats particularly within the same company, but the data management teams have the same difficulty in understanding why laboratories, when they are all doing the same tests on the same body fluid, can't all have the same reference range.
Use of the Internet and e-mail has begun to break down the barriers between the laboratory and data management.
There is now the possibility of these two groups getting to understand the needs of the other and thus for the lab to provide non standard data in a DM-friendly way.
Treating the lab as an investigator site may have the advantage of removing the need for transfer formats.
Brief Biographical Details:
Ann Speaight trained in biochemistry and haematology in a clinical environment. She set up laboratories for G D Searle in London and Brussels before joining J S Pathology and establishing the Clinical Trials Division. She was one of the founder member of Chandos Clinical Research and is on the main Board. She has a special interest in laboratory results as data items and has written extensively on the subject. She has also lectured on ACDM training courses on interpreting lab data.
Ann Speaight
Marketing Director
Chandos Clinical Research
2 Chandos Rd, London NW10 6NF
Tel: +44 (0)20 8961 5030
Fax +44 (0)20 8961 5399
E-mail ann@ccr.u-net.com