Innovation and technology drive P&G’s multi-billion dollar business Dr Peter Ling, Edith Cowan University Procter & Gamble (P&G) has over 20 billion-dollar brands with each generating over US$1 billion sales annually. These brands are Actonel (osteoporosis pill), Always (sanitary pad), Ariel (detergent), Bounty (paper towels), Braun (shaver), Charmin (toilet paper), Crest
Mind the gap 2nd draft livsval’Mind the gap’
The mundane co-ordination for producing scientific
evidence in clinical trials
Work in slow progress. Please comment, but do not quote Claes-Fredrik Helgesson
Department of Thematic Studies - Technology and Social Change
Claes-Fredrik HelgessonDepartment of Thematic Studies - Technology and Social ChangeLinköping University Claes-Fredrik HelgessonDepartment of Thematic Studies - Technology and Social ChangeLinköping University ’Mind the gap’
The mundane co-ordination for producing scientific
evidence in clinical trials1
Can the organising of activities have epistemic outcomes? It might seem reasonable that the organising of any endeavour aiming to produce knowledge might influence the outcome of that endeavour. Yet, it is far less clear how the organising of activities can shape the knowledge produced. This paper concerns the mundane co-ordination of large-scale clinical trials and in what ways this organising might contribute to the scientific authority of the results.
On my first day on the field of large trials I was told that large clinical trials to 90% was a matter of organisation. The remark was made by a woman in a high position at a large pharmaceutical company. She said it in response to my brief presentation of this research project. The setting was an investigators’ meeting for a large trial where personnel from a subset of the involved hospital clinics met with trial management at the pharmaceutical company sponsoring and co- Meetings such as this are just one of the plethora of co-ordination activities taking place within a large trial. A large multi-centre clinical trial mean that many clinics in parallel recruit and treat patients and where the results from this work are gathered and analysed centrally. This merging of data necessitates that all clinics have recruited the same kind of patients, treated them similarly, taken the same kind samples in the same way, registered data in the same fashion, and so 1 The research for this paper that was financially supported by The Bank of Sweden Tercentenary Foundation. This paper was previously presented at 4S in Washington, 21-31 October 2009. Previous versions of this paper was presented at the SOND workshop in Uppsala, Sweden, 25-26 April 2006 and at EASST, Lausanne, 23–26 August 2006. Claes-Fredrik HelgessonDepartment of Thematic Studies - Technology and Social ChangeLinköping University on. The amount of these efforts becomes further clear when considering that a large trial can involve thousands or tens of thousands of patients,who participate in the trial for a few years, regularly visiting one of several hundreds of clinics distributed across several countries. I take a large clinical trial to be a considerable organisational achievement.
Hence, the topic if this paper is the co-ordination of large trials. Co-ordination is especially interesting here given the many spatial and temporal distances contained within a large trial. My interest is hence in the activities that allows for co-ordinating a large number of scattered activities so they can be transformed into a condensed result that regularly is attributed as having significant scientific authority. In short, I’m interested in how the co-ordination efforts within trials contribute to the authority of the results.
I will in the following use material gathered through participant observation at different sites involved in clinical trials. Before turning to the observations, however, I will briefly discuss how we can appreciate co-ordination and distance in the production of scientific knowledge.
Co-ordination and distance in the production of scientific knowledge One important observation in science studies is that knowledge is changed as it is transported from the location where it was produced. Donald MacKenzie noted, for instance, in Inventing Accuracy how those directly involved in the gyro development had a close acquaintance with the subtle considerations involved in producing knowledge. This gave them, he noted, a source of highly intimate doubt in the knowledge produced that were absent to those more distant to the These effects of distance from knowledge production are of course interesting when it comes to large clinical trials and the notion of evidence based medicine. What I take as an important property of large trials is, however, the significant spatial and temporal distances within the trial. What we have here is not a demarcated laboratory within which the scale can be reduced, mistakes can be made, for a later extension of the network (cf. Latour 1983, 164). A clinical trial is a laboratory in the sense that it is an attempt to set up a controlled setting for knowledge production. Yet, the activities involved are far more distributed than the traditional laboratory. This, in my view, makes it appropriate to denote the large clinical trial as a distributed laboratory.
The distances between different activities means that much of the work to be performed at the clinics are pre-described in texts. Focussing on the aspects of co- ordination and distance, I would suggest, is to focus on the collective forms of practices that constitutes what has been coined the regulatory objectivity of medicine (Cambrosio et al. 2006). Yet, what I focus on here is not only the regulatory aspects of knowledge production but also how the practices relate to, and deviate from, the prescribed. In her study of two large clinical trials, Petra Jonvallen observed and examined the large amount of localising work necessary to be able to perform the centrally prescribed activities (Jonvallen 2005, 175-180). The marked centre-periphery structure of multi-centre trials makes room for a variation in how clinics adopt to the prescribed procedures. What is interesting, then is how co-ordination over distances might involve not only programmatic behaviours, but also various deviations that nevertheless cannot be counted as outright non-programmatic (cf. Helgesson and Kjellberg 2005). Moreover, it can be interesting to observe the extent to which such deviations are registered and relayed to other places.
Co-ordination in a distributed laboratory An abundance of organisations, rules and places Each clinical trial involves a large number of rules. One set of rules often referred to is the internationally agreed upon rules for clinical research named Good Clinical Practice (GCP). There are, in addition, several rules and prescribed procedures for each clinical trial. Many of these are documented in the trial’s protocol, which is the central document describing the purpose, design and The large array of organisations involved is also widely apparent. I have during a patient visit more than once seen how the research nurse within an hour have interacted with representatives of several different organisations. The nurse has faxed a form to the lab situated in another country, registered the patient visit with an interactive voice response system operated by another company, made a call to the local office of the CRO company to make an inquiry, and so on. Sometimes even the patients have had to make a phone call during their visit, for instance to ask their wife what medications they take at moment.
There are thus in connection to a patient visit a large number of highly specific communications performed between the clinic and other places and organisations involved in the trial. Distances are hence endlessly being traversed through I have on countless occasions observed instances where distance, the lack of proximity, appears as an obstacle to the conduct of the trial. In the meeting referred to in the beginning, there was a conversation about the recruitment rate Sara [representative of the sponsor]: ‘I want to ask you: Why do the recruitment curves look like this? Are there no [suitable] patients? Or is it that you cannot find them? Excerpt from field-notes [001:002, #rekrytering] Sara expressed a frustration that she, from her position, could not know the nature of the problem. To facilitate recruitment, they had now engaged a few more clinics for identifying possible patients by having these clinics examining their patient records and calling possible candidates thus identified. However, the checklist developed for this had been copied from the checklist used when screening patients on site. This caused a problem as one of the physicians dryly Patrick [a physician]: ‘It is difficult to assess the patient’s veins over the phone!’ Sara [representing the sponsor]: ‘Oh dear, that is a problem. The checklist has already been sent to the ethics committee [and cannot hence be changed]. Assessing the veins was important here since it checked whether a patient was suitable with the particular way in which blood samples was to be taken. Here a new distanced routine for recruiting patients was introduced, causing a new and apparently unforeseen problem related to a lack of proximity to the potential I also encountered frustrations expressed by those meeting patients that others did not know what they knew. During a telephone conference involving the personnel at several clinics as well as trial management, a physician wanted to relay that many patients had said that the pills were difficult to swallow: The physician … then takes on an acid tone and asks the international management and the representatives of the pharmaceutical company whether they ever have seen the pills and how large they are. This leads to a concession that the management team has discussed to re-coat the pills, but that nothing has been decided yet.
There are of course a large number of devices that in many ways facilitates traversing the distances of a trial. A routine patient visit exhibits many such devices. The patient may first meet the physician for a short talk and a physical examination. At the same time the nurse prepares the material for the taking of specimens. While the nurse later takes the physical specimens, the physician might register some data about his examination in the patient record, as well as in the case report forms for the trial, the CRFs for short. (The CRFs contains the data that is to be sent to the trial’s data management unit.) After the specimens have been taken, the nurse will put some blood samples into a centrifuge to separate out the blood serum as well as fill in some additional CRFs. In such a routine visit there are a number of distance-traversing devices implicated: the patient record, the CRFs (with three carbon copies), the test-tubes with easily transportable serum, a form faxed to the laboratory, and so on.
Distance and the organising of information A central facet of clinical trials is the control regarding who is to know what. One containment of information concerns the identity of the patients in terms of their name and social security number. This information is not allowed to leave the clinic. The records and samples sent from the clinic are therefore coded by a patient ID unique for the study. Data recorded for a particular patient ID are only possible to concatenate with an identifiable patient by looking at records kept at The control of who is to know what is particularly obvious when the trial is randomised and double-blind. This study design implies that neither the patient nor the personnel know what treatment – usually a pharmaceutical – the patient receives. The information about the allocation of treatments is kept in a dedicated The procedures for assigning treatments in a double-blind randomisation is designed to allow for auditing how the assignment has been made at the clinic, while still keeping the actual assignment opaque for everyone involved. There are, furthermore, procedures devised to enable the personnel at a clinic to un-blind and “break the randomisation code” in cases of emergency to quickly find out what substance a patient has been randomised to take.
Apart from the devices used to contain information, distance is in itself also a useful resource. This is, for instance, true for containing the patient identity. The people working at a clinic usually does not meet patients from the other clinics, and people working at the sponsor does not meet the patients. By using a specially assigned patient-ID to identify specimens and CRFs leaving the clinic, distance is put to good use in keeping the name and identity of the patients local.
There are, however, one group of people that makes this a bit more problematic, namely the monitors. The monitor works at a CRO or pharmaceutical company and visits the clinics regularly to examine the site. The monitor examines, for instance, that the patient records have the right information and that data from a given visit is the same in the patient record and in the CRFs. Through this work, hence, they learn about patient identities while Another aspect of ‘who is to know what?’ is the conception that the staff involved in a trial should know certain things. At every site there is an investigator’s file. This is one or several binders where a copy of the trial’s protocol, copies of possible amendments to the protocol, important correspondence etc. are kept. The personnel involved in the trial receives a steady stream of new documents from the CRO, the sponsor, governmental agencies etc. to be filed in the investigator’s file.
A well-kept and ordered investigator’s file is seen as an indication that the site staff has received the information. The order of these binders are also examined by the monitor and can be subject to a possible audit by governmental agencies. Yet, in the studies I have observed, I have seen how the monitor in his or her examination of the binder regularly orders the content in the binder as well as supplements the binder with missing documents. To assist in this, they keep a ghost copy of each clinic’s investigator’s file at the co-ordinating office. This allows the monitor to keep track of what ought to be in the real investigator’s file: Mats [the monitor] goes to examine and update the investigator’s file. //He then tells me that these documents actually ought to be put into the binders by them [the personnel at the clinic], but that he usually brings extra copies and puts them into the binders if they are missing. This is something that he has learnt.
Adapted from field-notes [018:001, p. 32:11-12] I’ve also observed how the monitor, who’s task is to examine the forms, also aid in getting the forms completed in a satisfactory way: Monica [the monitor] says that she prepares what he [the physician] shall fill in. She does not herself fill in the forms, but prepares what he shall fill in. ‘This is something that he actually ought to do himself, but it is easier to give him a helping hand.’ Monica says. ‘In this way it becomes correct from the start,’ she continues. She then asks me: ‘Do you have children?’ I reply that I do. ‘Then you understand what I mean,’ she says and gives me a look full of meaning.
This, again, is an example of how practice consciously deviates from how it ‘actually ought to be done.’ Here it is done for the sake of recording correct data. Organising for verifiability and consistency A lot of work is spent by the monitors to ensure a high quality of the collected data. I was, however, repeatedly struck by how these efforts often were paired with a highly pragmatic attitude as to what was to count as factual circumstances. When it comes to verification, the focus was largely on how information in one place, such as on a CRF, corresponded to what was entered in another place, such as on the patient record. This is also the kind of verification that occur in a possible audit by a governmental agency.
Although the CRF pages are heavily standardised to suit the trial, the same is not true for the patient records. Since the patient records also have a role in clinical trials as a source for verification, there are attempts to make the physicians and nurses enter data in the patient records that makes them suitable for this purpose. I have several times seen how the trial co-ordination office has issued text-file templates helping physicians and nurses in what what to enter into the patient records regarding a given visit. However, having templates for what is to be entered in the patient records is somewhat of a sensitive issue, since no one can dictate what they are to write in the patient record. This is also recognised since the templates are described as mere suggestions. (However, the monitor can in face-to-face meetings with the site staff be more candid regarding the need to The emphasis on verification, that two sources indicate the same thing, can also When they come back Natalie [a nurse] asks Samir [a patient] about his date of birth: ‘Is it correct that you are born 194X-XX-XX?’ [A date in the 1940s.] Samir responds: ‘ That corresponds to what is noted in my passport. My mother says that I’m a few months older. You know, in our countries they regularly do not note the correct date.’ Natalie accepts this as an affirmative answer to her question.
Here, then, the true birth date was taken to be the date noted in the patient’s passport and hence the date noted in the patient record. Entering the date reported by the patient’s mother into the CRF would have made it differ from what is noted in the patient record. This had made the date less easily verifiable lest one contacted the patient or his mother. It is precisely such remote referents that are avoided when emphasising the verifiability across written sources. Sometimes scarce patient records can also allow for entering little data into the study. This is a conversation between two monitors who are at a clinic to gather Monica tells Catharine that the summary of the patient record related to the death was short. // She is particularly pleased that the patient record summary does not contain any note of pharmaceuticals. ‘No pharmaceuticals!’ Catharine exclaims! The monitors’ relief was in this case apparently connected to the fact that the immediately available record did not state any pharmaceuticals had been given to the now diseased during his or her last days. This meant that no pharmaceuticals had to be entered into the CRF, saving them a lot of otherwise tedious work. Again the focus is on aligning the data in the patient record and the CRF. Sometimes, however, missing data has to be compensated for since it has to be entered on the CRF. Concerning the same patient as the previous quote, I observed how one of the monitors discussed with the physician what date he should put on a CRF-page given that the available patient record only stated the Monica [the monitor] tells Daniel [the physician] that they only know 2004, and that it is reasonable to put it to the middle of the year, in the middle of a month. They agree that the middle of June, 15 June, is a good date.
There are hence several different activities used to make the trial’s data consistent and verifiable. It is my clear impression that the focus is on achieving a consistency between different sources on paper, rather than a consistency between the recorded data and what they represent outside the records. This points to the fact that the activities to make the data verifiable simultaneously stabilises the A traditional explanation as to why the results of large clinical trials carry scientific authority emphasises their experimental design, the wealth of data generated, as well as the statistical techniques used to unveil the important My focus on the many distances and the efforts of co-ordination within a trial provides a different picture. Viewing a large clinical trial as a distributed laboratory I have provided examples of efforts and devices aiming to overcome obstacles posed by the distances within trials. More importantly, I have provided examples of co-ordination where the resulting behaviour cannot be counted as outright non-programmatic, yet clearly not fully aligned and compliant with the formal rules and standards of trials. Think, for example, on the cases where something was carried out in a pragmatic way as opposed to how things ‘actually ought to be done.’ Here distance appears more as a resource rather than an obstacle. Distance is what keeps the intimate knowledge about these deviations I have also provided examples of more covert efforts of co-ordination. This includes some of the work done to facilitate verification by aligning what is recorded on patient records with what is recorded on CRFs. This further includes the dissemination of templates for entering information into the patient records, or the monitors preparing what is to be filled in on the CRFs. Again, distance here appears more as a resource than an obstacle, allowing for the intimate knowledge about this covert co-ordination to remain more local than the results We have thus, visible efforts of co-ordination, such as rules and standards, that are not as effective as they appear to be; and more covert efforts of co-ordination that contributes to the ordering of the trial. A common denominator between many of these examples is that they display how intimate knowledge about possible uncertainties and deviations are kept local, while allowing the certain and ordered to travel. I take this to be a significant effect of the efforts to co- ordinate large trials. By being a distributed laboratory, the many activities co- ordinating a large trial allows for harnessing uncertainties and deviations to the local, while making the certain and ordered to travel within the trial. And At the outset of this paper I ventured to ask whether the organising of activities can have epistemic outcomes. I would now like to suggest that the co-ordination within large trials foster the production of certainty and order. This is an effect not stemming from scale. An epistemic effect of scale would stem from variations cancelling out one another. The effect of the co-ordination within the distributed laboratory, on the other hand, comes instead from arrangements that reduce some such noise to be aggregated in the first place. It is, hence, not the scale that is the key, but the gaps. And indeed, the efforts that deal with and make good use ReferencesAbramson, John. 2004. Overdo$ed America: The broken promise of American medicine. Bloor, David. 1991. Knowledge and Social Imagery. 2nd ed. Chicago: University of Bosch, Jackie, Salim Yusuf, Janice Pogue, Peter Sleight, Eva Lonn, Badrudin Rangoonwala, Richard Davies, Jan Ostergren, and Jeff Probstfield. 2002. Use of ramipril in preventing stroke: Double blind randomised trial. British Medical Journal 324:699–702.
Busfield, Joan. 2006. Pills, Power, People: Sociological Understandings of the Pharmaceutical Industry. Sociology 40:297–314.
Cambrosio, Alberto, Peter Keating, Thomas Schlich, and George Weisz. 2006. Regulatory objectivity and the generation and management of evidence in medicine. Social Science & Medicine 63 (1):189-199.
Collins, Harry M. 1985. Changing Order: Replication and Induction in Scientific Practice. Beverly Hills: Sage.
Earl-Slater, Alan. 2002. The Handbook of Clinical Trials and Other Research. Oxon: Helgesson, Claes-Fredrik. 2004. Mellan vetenskap och verklighet: Om vad riktlinjer är och gör. In Lychnos: Årsbok för idé och lärdomshistoria, 131–149. Uppsala.
Helgesson, Claes-Fredrik, and Hans Kjellberg. 2005. Macro-actors and the sounds of the silenced. In Actor-Network Theory and Organizing, edited by Barbara Czarniawska and Tor Hernes, 145–164. Malmö & Copenhagen: Liber & Copenhagen Business School Press.
Jonvallen, Petra. 2005. Testing Pills, Enacting Obesity: The work of localizing tools in a clinical trial. PhD diss., Department of Technology and Social Change, Linköping University, Linköping.
Latour, Bruno. 1983. Give Me a Laboratory and I will Raise the World. In Science Observed: Perspectives on the Social Study of Science, edited by Karin Knorr-Cetina and Michael Mulkay, 141–170. London: Sage.
———. 1987. Science in Action: How to Follow Scientists and Engineers Through Society. Cambridge: Harvard University Press.
Lilford, Richard, and Andrew Stevens. 2001. Clinical Trials: Introduction. In The Advanced Handbook of Methods in Evidence Based Healthcare, edited by Andrew Stevens, Keith Abrams, John Brazier, Ray Fitzpatrick and Richard Lilford, 7–9. London: Sage.
MacKenzie, Donald. 1990. Inventing Accuracy: A Historical Sociology of Nuclear Missile Guidance. Cambridge: MIT Press.
Marks, Harry M. 2000. Trust and mistrust in the marketplace: statistics and clinical research, 1945-1960. History of Science 38:343–355.
Porter, Theodore M. 1995. Trust in Numbers: The Pursuit of Objectivity in Science and Public Life. Princeton: Princeton University Press.
Star, Susan Leigh, and James R. Griesemer. 1989. Institutional Ecology, ‘Translations’ and Boundary Objects: Amateurs and Professionals in Berkeley's Museum of Vertebrate Zoology, 1907-1939. Social Studies of Science 19:387–420.
Timmermans, Stefan, and Marc Berg. 2003. The Gold Standard: The Challenge of Evidence-Based Medicine and Standardization in Health Care. Philadelphia: Temple University Press.
Outcomes after Transplantation of Cord Blood or Bone Marrow from Unrelated Donors in Adults with LeukemiaMary J. Laughlin, M.D., Mary Eapen, M.B., B.S., Pablo Rubinstein, M.D., John E. Wagner, M.D., Mei-Jei Zhang, Ph.D., Richard E. Champlin, M.D., Cladd Stevens, M.D., Juliet N. Barker, M.D., Robert P. Gale, M.D., Ph.D., Hillard M. Lazarus, M.D., David I. Marks, M.D., Ph.D., Jon J. van Rood, M.