Antibiotic Resistance and the Problem of Susceptibility Testing

While tests for microbe susceptibility were not unknown to bacteriologists prior to the 1950s, their importance increased considerably with the emergence of antibiotic resistance as a major clinical problem. The new anti-infective drugs, and penicillin in particular, had been released for common use in the immediate postwar era and there was a prevailing understanding that these “magic bullets” had turned infection into a mere technical problem, amenable to pharmaceutical intervention (Bud Reference Bud2007, 98). However, in the 1950s it became apparent that antibiotics could not conquer all microbes. During this period resistance to antibiotics, which had previously been an occasionally-occurring problem in individual patients, was becoming more frequent and slowly beginning to gain epidemic proportions. Epidemic strains of the bacterium known as staphylococcus aureus made their appearance in hospitals and nurseries worldwide, causing serious concern because they proved to be resistant to penicillin and other antibiotics, thus resulting in many deaths around the world.

However, it was not evident what kind of an object drug resistance was. For clinicians in hospitals worldwide, resistance was mainly a clinical phenomenon – a failure of antibiotic drugs to act on patients suffering from infectious disease. To laboratory bacteriologists, drug resistance was an object belonging to Petri dishes or broth cultures. Resistance was yet another object to the international and local health policy makers, who wanted operable standards for evaluating the epidemic situation internationally and nationally. The testing of microbe susceptibility to antibiotics was thus a procedure that, although performed in the laboratory, could afford a common language for all these different sites, enabling resistance to be formulated as a collectively intelligible object. However, although collectively intelligible, drug-resistance also needed to be crafted as a multifunctional product that could have different uses and meanings in different sites. It had to be simultaneously shaped as a contribution to scientific bacteriological studies, as a simple diagnostic tool applicable in not too sophisticated clinical laboratories, as a useful test for clinicians in medical practice, and as a marketable system.

This article will explore the processes of standardization taking place in Scandinavia, and particularly in Sweden and Norway. These countries share a more or less common language, and important political ideals and practices, and in the period in question these countries developed the political system commonly known as the “Nordic Welfare State.” Since elements of these practices and ideals will be important to my discussion later on, it seems necessary at this stage to specify this conception of the “Nordic welfare state.”

Nordic Welfare Politics and the “Nordic Welfare State”

The conception of a special Nordic model (one that includes Sweden, Norway, Denmark, Finland, and Iceland) has gained wide acceptance in social science literature (Esping-Andersen Reference Esping-Andersen1990; Titmuss, Titmuss, and Abel-Smith Reference Titmuss, Titmuss and Abel-Smith1974; Furniss and Tilton Reference Furniss and Tilton1979; Andersen Reference Andersen and Graubard1986; Shalev Reference Shalev1983; Seip Reference Seip1994). The Nordic welfare model is often depicted as the “social democratic model” or “social democratic regime,” and during this period social democrats partly monopolized governmental power in Scandinavia; in Norway they did so from 1945 to 1963. However, most analysis includes descriptions of a Nordic Sonderweg, where efforts towards building a broad political consensus on a platform of state-regulated socially modified capitalism were taken before the Second World War and before the dominance of the social democrats. The concept of the Nordic welfare state included a universal approach: the welfare benefits should include the entire population. Thus, global programs should be preferred over selective ones, and a tax-based social security system should underpin the economy-wide, solidaristic wage politics. Unlike in Germany, where insurance contributions and benefits differed according to levels of income, benefits in the Scandinavian countries were paid, as in Britain, at a flat rate for all. But they differed from both the German and the British provisions in that retirement pensions and sickness insurance were financed from general taxation. According to social scientists and historians, the “welfare state” was gradually built up during the 1950s and 1960s, reaching its full appearance as a social and political reality in Norway in 1967 (Seip Reference Seip1994).

Historians of the Nordic welfare state have been criticized for using as their historical material ideals rather than practices, and critics have argued that the Scandinavian welfare state in its ideal form probably never existed (Nordby Reference Nordby1990). Recent research on the concept of the “welfare state” in Norway has documented that the term “welfare state” was “practically absent in the platforms and elections manifestos of the political parties,” including those of the Labor Party, until the early 1980s (Haave Unpublished Paper). However, during the years between 1945 and 1970, there was a general cross-political agreement of the most important principles of welfare politics: that economic growth, social security, and integration should mutually support one another in harmonious cooperation. Referring to the postwar period until the 1970s in both Sweden and Norway, the historian Francis Sejersted argues that “for a long period, it seemed that developments justified the idea [of welfare policies]. This was Social democracy's happy moment” (Sejersted, Adams, and Daly Reference Sejersted, Adams and Daly2011, 266). The thirty years following the Second World War, during which the efforts to standardize a method for susceptibility testing took place, was at the same time the golden age of public welfare expansion.

Let us now take Norway as an example: Even the critics of the historians of the welfare state agree that in Norway, welfare politics were in no other field as strongly adhered to and carried out as within the field of health (Nordby Reference Nordby1989). Karl Evang, the director of Health from 1938 until 1972, used the concept of welfare state already in 1951 (ibid.). His administration continuously fought for and gained acceptance for the principle that free health care of excellent quality should be provided for all in publicly owned institutions, thus reducing the demand for private institutions. The health system was to be egalitarian and universal: all Norwegians, regardless of personal income and place of residence, would be guaranteed both good protection against disease and high-quality treatment of sickness and injury financed by the state. The basis of the public health policy in this period was the expansive definition of health adopted by the WHO in 1948, and with the same Karl Evang as one of its main architects: “a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity” (Berg Reference Berg2002). Achievement of these goals required the expansion of preventive health care, which would be carried out by district medical officers assisted by specially trained public health nurses and supervised by county medical officers. It also required the rationalization of the country's many small, local hospitals into a network of large, central hospitals. Legislation developed in rapid succession during the decades after the war and established a coordinated system of comprehensive public health and related welfare services: school dental services (1947), general nursing (1948), universal sickness benefit (1956), public-health nursing (1957), home nursing (1959), occupational rehabilitation and disability (1960), mental health and psychiatric care (1961), universal social security benefit (1966), hospitals (1969), and public health centers (1972) (Schiøtz and Skaset Reference Schiøtz and Skaset2003). The size of the health sector (public and private) grew considerably. Between 1950 and 1976 its share of the gross national product rose from 3.5 per cent to 8 per cent; the number of certified physicians and nurses doubled; in many communities hospitals and health services became the largest single employer (ibid.).

Welfare policies in Sweden followed more or less the same pattern. For example, Sweden introduced a compulsory national sickness insurance program covering the entire Swedish population for medical care and sickness cash benefits in 1955. Hospital policy in both Sweden and Norway during the postwar period was characterized by efforts to improve centralized planning and coordination and a growing desire to increase the central state's administrative power as a basic feature of social democratic policy. However, Sweden had a more developed public health system, where most health institutions were public, whereas in Norway there was a patchwork of public and private institutions (Haave Reference Haave, Christiansen, Petersen, Edling and Haave2006). However, the majority of the major institutions were publicly owned, and most health care was paid for by the public (Sejersted, Adams, and Daly Reference Sejersted, Adams and Daly2011). It must be emphasized though, that this public ownership was a long-term cultural legacy in the Nordic countries and not an effect of the twentieth-century welfare state reforms (Haave Reference Haave, Christiansen, Petersen, Edling and Haave2006).

Particularly relevant to our context, most of the important bacteriological laboratories in the Nordic countries were state-owned. In Norway, all of the bacteriological laboratories at the time were state owned, and most of them were located within university hospitals. As late as 1970, there were only seven bacteriological laboratories in the entire country.Footnote ¹ In the same manner, bacteriological laboratories were part of or associated with university institutions throughout Sweden. Early examples of bacteriological laboratories, also at the county council level, were Borås (started in 1951) and Sundsvall (1957). The expansion of the Swedish Health Care sector during the 1960s–1980s led to several new laboratories (Hambraeus et al. Reference Hambraeus, Kahlmeter, Malmborg and Nyström2007). This concentration of laboratory services in relatively large institutions and within a public health care system with little opportunity for commercial endeavor had an important consequence: It enabled a system based on labor-intensive and carefully standardized processes in the laboratories.

There was also a profound skepticism towards private initiatives, especially initiatives from the pharmaceutical industry, most pronounced in Norway. In 1953, a governmental monopoly on the wholesale distribution of drugs was established, the Norwegian Medicinal Depot (NMD). In the proposition to the Parliament, drugs were characterized as “products which are very poorly suited for trade following commercial principles. They belong, in a most pronounced way, to necessary consumer articles whose determination of price and sales should be regulated in accordance to pure principles of need” (Norwegian Department of Social Affairs 1952, 4–5). This new governmental agency was given tasks that united wholesale distribution to public health initiatives (including drug information and drug statistics). The governmental health sector thus expanded by including new instruments for Norwegian health politics (Moen Reference Moen1997). In another context, discussing the regulation of patents, the health director Karl Evang explained that this skepticism towards commercial initiatives in the field of health was a long-standing tradition in Norway: “Through regulation and otherwise, the authorities in this country have at all times attempted to counteract any unreasonable commercial and economic utilization of drugs, and to remove unnecessary competitive factors from the trade of drugs.”Footnote ²

I will come back to how these various features of the “Nordic welfare state” can be said to having taken part in constructing the standardization processes regarding sensitivity testing, but also how these practices took part in constructing a common Nordic ground. In order to do that, however, we need to focus first on the concrete practices taking place within the laboratory walls regarding sensitivity testing.

Sensitivity Testing: “A Necessary Evil”

In the immediate postwar era, susceptibility testing was a frail object. There was no agreement concerning what method to use or how to carry out the tests, once chosen, in a uniform way to ensure consistent results. Even for the same test, there were differing traditions in the composition of the medium, the size of the inoculate, and the concentration of the antibiotic (Fairbrother, Martyn, and Parker Reference Fairbrother, Martyn and Parker1951; Boundy and Munch-Petersen Reference Boundy and Munch-Petersen1962).

In the beginning of the 1960s, for example, several methods of susceptibility testing were used in the bacteriological laboratory at Rikshospitalet, the Norwegian National Hospital, which was one of Norway's four bacteriological laboratories at that time (Bøvre Reference Bøvre1962). Most of the methods in use were variations of two basic methods: the diffusion method and the dilution method. The latter was the most time consuming and was therefore impractical for daily use. However, it gradually became the reference method in the 1950s because it could give quantifiable results. In this technique, a series of cultures were tested with various concentrations of antibiotics in order to calculate the minimal antibiotic concentration that impeded bacterial growth in the medium. This concentration was called the Minimal Inhibitory Concentration (MIC).Footnote ³

In the diffusion method the reverse principle was used. Here the antibiotic to be tested was applied on a culture medium seeded with the microbe. If the microbe was sensitive to the antibiotic, a zone of inhibition appeared around the drug. The antibiotic could be applied in ditches on the medium (the ditch plate method), in cylinders (the Oxford cup method), or on the surface of the medium in form of tablets (the tablet method), or paper disks (the paper disk method). The Oxford cup method was still used in Norway early in the 1960s, but the two dominating methods were the tablet method and the paper disk method (Steen Reference Steen1955). Paper disks had first been employed for assays of penicillin during World War II (Vincent and Vincent Reference Vincent and Vincent1944), but the method was quickly adapted to the antibiotic susceptibility test (Thompson Reference Thompson1950).Footnote ⁴

In the beginning of the 1960s at the Wilhelmsen laboratory at Rikshospitalet, the most prestigious bacteriological laboratory in Norway, responsibility for antimicrobial susceptibility testing in the lab was delegated to the research assistants with least experience (Midtvedt Reference Midtvedt2010); such testing was considered a necessary evil. For example, Dick Henriksen, the head of this laboratory, wrote a letter to the hospital manager in which he expressed regret that his laboratory had been designated responsibility for the work on hospital hygiene (including sensitivity testing procedures) in the hospital, because as he explained, “it involves a routine which is demanding, and which not insignificantly reduces our capacity to do other and more interesting work.”Footnote ⁵

Culture media were “home brewed” and would be prepared by technicians in the basement at the laboratory (called “the media kitchen”). For some slow growing organisms defibrinated horse blood that the laboratory bought from the stable at the king's estate in Bygdøy was added to the medium. There were no standards with which to perform the tests, and the procedure would vary depending on the technician.Footnote ⁶ As the task of microbe susceptibility testing became increasingly recognized as one of complexity, several Norwegian bacteriologists began calling for a standardization of the procedure.

Calls for Standardization

The lack of standardized procedures became increasingly problematic during the 1950s. The Norwegian bacteriologist Kjell Bøvre commented that the “technical problems” with susceptibility testing were many, and that “deviant” methodology could lead to erroneous results (Bøvre Reference Bøvre1962; Steen Reference Steen1955). Bøvre was of the opinion that errors in the practical execution of the testing were important causes of incongruity between the test results and the clinical outcome. In 1956, the Norwegian bacteriologist (later otorhinolaryngologist) Emil Steen complained that there was a whole family of related methods in use, and that the same bacterial strain could thus be described as resistant in one laboratory and sensitive in another, resulting not only in the clinicians’ loss of trust in laboratory services but also creating potentially life threatening situations (Steen Reference Steen1955). Moreover, the sampling was often made with what Steen called an “incorrect indication,” i.e. on a mixed culture of bacteria with “doubtful pathogenicity.” An additional and related problem was that etiological inferences were drawn “uncritically” from any pathogenic bacteria the laboratory could find, independent of the clinical picture (Steen Reference Steen1955, 1269; Bøvre Reference Bøvre1962). There had also been discussions regarding the discrepancy between the results of susceptibility testing in the laboratory and those of the treatment in the clinic (Steen Reference Steen1955).

These Norwegian demands for standardization mirrored international discussions on the subject of susceptibility testing during the 1950s (for a discussion of the international discussions on susceptibility testing, see Gradmann Reference Gradmann2013). During this time, the bacteriological laboratory was becoming increasingly relevant for clinical practice. One decade earlier, sensitivity testing had been used in bacteriological research and in pharmacology as a way to examine the quality of antibiotics and as an aid in the search for new effective drugs. However, the epidemics caused by staphylococci that were haunting the hospital wards in the Western world during this decade transformed sensitivity testing into an important clinical tool that became increasingly requested in daily clinical practice (Hillier Reference Hillier2006).

The antibiotic susceptibility of a strain could not be measured directly – in vivo. Rather, it had to be deduced from the (in vitro) activity of the antibiotic in the laboratory. Bacteriologists would argue that due to the worldwide increase of acquired antibiotic resistance in the 1950s, empirical therapy became more and more uncertain, and the foundation for such a therapy needed constant re-evaluation (Fairbrother, Martyn, and Parker Reference Fairbrother, Martyn and Parker1951). To be able to treat the infection one needed to know the sensitivity of the aetiological agent to the different drugs available. Though this test was carried out in the laboratory, it was in fact framed as a clinical one, as a necessary tool for the clinician tailoring treatment: “chemotherapy without bacteriology is guesswork” (Howie Reference Howie1962). The phrases “sensitivity testing” and “susceptibility testing” point to the primarily clinical interest in the testing: their main function was to provide a guide to clinical treatment. In this period, sensitivity testing was predominantly regarded as a clinical tool and not a means for epidemiological measurement of resistance in populations. Its primary aim was to help clinical medicine find appropriate (sensitive) drugs, not to measure resistance.

Yet, and as mentioned previously, problems arose when a lack of testing uniformity resulted in discrepant results, with test results revealing that the same microbe could be categorized as sensitive in one laboratory and resistant in another (Gould and Bowie Reference Gould and Bowie1952 and Expert Committee on Antibiotics 1961). Thus, toward the end of the 1950s, sensitivity testing was increasingly problematized and the trust in such testing was deteriorating (Branch, Starkey, and Rodgers Reference Branch, Starkey and Rodgers1955; Branch, Starkey, and Power Reference Branch, Starkey and Power1959; Expert Committee on Antibiotics 1961).

Efforts at Standardization

In Sweden, attention had been drawn towards the problematic lack of testing uniformity since the late 1940s. In 1949, the Hygiene and Bacteriology Section of the Swedish Medical Association established a permanent committee to explore the existing guidelines and possibilities for adopting a uniform methodology in antibiotic susceptibility testing (Ericsson et al. Reference Ericsson, Erlanson, Tunevall and Wallmark1953). After having done a research project where they compared the existing procedure in several Swedish laboratories, the committee concluded that there were several methods in use in the country. Since there was no internationally recognized uniform methodology, they did not consider it practical at this stage to recommend one particular method. However, they outlined several general recommendations regarding the choice of method, the media, and the inoculate, in addition to giving advice on how the results of the susceptibility testing should be communicated to the clinician.Footnote ⁷

Hans Ericsson, a young member of this committee who would later become a professor of bacteriology at Royal Caroline Hospital, committed his life's work to the standardization of sensitivity testing, not only in Sweden, but also internationally. Starting his career investigating diphtheria toxin (Ericsson Reference Ericsson1946), he developed an interest for sensitivity testing and by the 1950s he was engaged in an international network of bacteriologists working with sensitivity testing. He was the secretary of the “Second Report” published by the WHO's Expert Committee on Health Laboratory Methods (Expert Committee on Hospital Laboratory Services 1959). Then, in 1961 the Expert Committee on Antibiotics published its second “Technical Report,” exclusively devoted to sensitivity testing (Expert Committee on Antibiotics 1961). The report proposed that the standardization of testing would enable antibiotic research with reproducible results and the comparison of clinical and epidemiological studies from different laboratories and different countries. They issued recommendations for broad governing principles for this testing and urged researchers to conduct further studies to improve the standardization of these methods. However, a lack of agreement remained regarding what to standardize, who should make decisions about standardization methods, and the means by which such methods should be enforced. In what follows, I will focus on such attempts at standardization within the Nordic context, concentrating on the Norwegian and Swedish case. For pragmatic reasons, I will divide the discussion according to what kind of work was done in order to achieve standardization: of materials and methods, that is, of the disks used in the sensitivity tests and how to use them (i.e., the procedures in the laboratory when employing the disks); and of the interpretations of the results. All standards presuppose one another, and in practice, these different types of standards interacted. However, each kind of standard implies a different kind of coordination work, and such a division is useful in so far as it allows for a description of these different kinds of work.

The Swedish Paper Disk Method: Materials and Methods

Commenting on the WHO report on standardization of sensitivity testing in 1961, Kjell Bøvre wrote that if such a “work on a higher level” should bear fruit locally, the laboratories adopting the standardized methods should execute their work in accordance with the regulations. As we have seen, he argued that this was not the case in Norway at that time. Norwegian laboratories lagged far behind when compared to other countries, and in particular when compared to Sweden (Bøvre Reference Bøvre1962; Midtvedt Reference Midtvedt2010). In marked contrast to the situation in Norway, where sensitivity testing was regarded as a mere duty within the research laboratories, in Sweden the development and improvement of such testing was promoted as a field of research. Alberto Cambrosio has described how standardization work in postwar biomedicine amounted to a full-fledged form of research both in itself and because research on standardization affected more traditional forms of research (Cambrosio Reference Cambrosio, Gradmann and Simon2010). The Swedish research in the context of sensitivity testing at Caroline hospital is a good example.

From the beginning of the 1950s, Ericsson and his colleagues took an interest in the paper disks as a tool for routine susceptibility testing. The dominating paper disk method at that time was the multidisk method, where several disks of each antibiotic, each of them with different concentrations, were applied on the plate (Barry Reference Barry1989, 207).Footnote ⁸ Ericsson, however, started to develop a single disk method, where only one disk was applied for each antibiotic. This enabled several antibiotics to be tested by one and the same plate. From 1952, the bacteriologists at Royal Caroline Hospital initiated the production of paper disks for their own use from 1952 (Ericsson Reference Ericsson2011; Ericsson, Hogman, and Wickman Reference Ericsson, Hogman and Wickman1954). After “unacceptable qualities” had been documented in existing disks on the market (Branch, Starkey, and Rodgers Reference Branch, Starkey and Rodgers1955), the production was increased in order to supply other Swedish laboratories (AB Biodisk 1972).

Ericsson would prepare the paper disks by using an ordinary hole-punch to cut out small 6 mm disks from differently colored 428 mm Ford blotting-paper. The diameter of the disks was determined by the size of the holes in the hole-punch. From a burette with a narrow outlet, one drop of drug solution was placed on each disk. A different color was used for each chemotherapeutic agent (Ericsson Reference Ericsson1960a, 37–42; Ericsson, Hogman, and Wickman Reference Ericsson, Hogman and Wickman1954; Ericsson and Löfström Reference Ericsson, Löfström and Hammarsten1952).Footnote ⁹ Thus, an object from ordinary life – the blotting-paper – was transformed via another ordinary object – the hole-punch – to become a crucial object in the laboratory analysis of antibiotic susceptibility, and eventually, and as we shall see, a commercial object produced by Ericsson's successful company Biodisk.

The disks were first produced in the bacteriological laboratory at Royal Caroline Hospital. Initially, the enterprise had an informal, cooperative character, and was run in collaboration with other Swedish laboratories. The surplus earnings were spent on research on laboratory techniques, on sensitivity testing, and “rational use” of antibiotics in the clinical ward, in accordance with Ericsson's belief that sensitivity testing was the prerequisite for rational use of antibiotics. “Rational use of antibiotics” was also the name of his Ph.D. thesis, defended in 1960 (Ericsson Reference Ericsson1960a).

During these years of informal organization, the Swedish enterprise was called Lappverket (lapp in Swedish means disk). In 1960, the paper disk method was the dominating diffusion test, and was recommended in the second WHO technical report on antibiotics (Expert Committee on Antibiotics 1961). The production of disks at Royal Caroline Hospital steadily increased, and, by the end of the 1960s, had grown to such a scale that it required an independent economy. Furthermore, private enterprises within the university had become unpopular by the end of the 1960s, thus providing yet another reason for the establishment of an enterprise outside the university organization (Kronvall Reference Kronvall2007b).Footnote ¹⁰ In 1967, a limited company named AB Lappverket was constructed. Since the name was difficult to translate to international customers (it could be directly translated as “patchwork”), the name was changed to AB Biodisk the following year (Kronvall Reference Kronvall2007b). Until the end of the 1960s, the disks were individually impregnated with antibiotics, in a labor-intensive craft-like way. Even so, AB Biodisk was able to meet consumer demand in Sweden, Finland, and Norway until the end of the 1960s. In the mid-1950s, it had supplied disks to the majority of Swedish laboratories (Referensgruppen för antibiotikafrågor 1977, 4), and already in 1960, disks were manufactured on a fairly extensive scale (1.5 to 2 million disks annually) (Ericsson Reference Ericsson1960b, 412). In 1966, the annual sale was 8.5 million disks. In 1972 the company produced 19 million disks, and had an annual turnover of 1.4 million Swedish kronor (AB Biodisk 1972).

From the beginning of the 1970s, AB Biodisk not only sold disks, they also provided supplementary equipment. Because the commercially available media differed largely in composition, between the different brands but also between batches produced by the same manufacturer, they had long been considered a challenge for standardization efforts. Thus, AB Biodisk started to sell its own media together with the disks. Unlike the disks, the media were not produced in the laboratory. AB Biodisk would import the media, control all the batches, and sell the best ones under their own name PDM-Antibiotic Sensitivity Medium (AB Biodisk 1972; Ericsson and Ericsson Reference Ericsson and Ericsson1975; Midtvedt Reference Midtvedt2010). Lappverket had during the 1960s initiated the production of additional accessories that were supplied primarily to the Nordic market and parts of Germany and Switzerland, a production that was continued by AB Biodisk. AB Biodisk also became the Nordic distributor of the then dominant method of determining bacterial strains, the API system (AB Biodisk 1972).

Standardizing Procedures: Enrolling Clinical Bacteriologists

However, it was not enough to have the materials in place. In order to succeed, this method needed to be accepted as a reproducible method, applicable for researchers also outside the walls of Royal Caroline Hospital, independent of its creator and the bacteriologists working in that laboratory. Lappverket was not the only producer of materials for sensitivity testing in the Nordic market; its main competitor was the Danish producer of antibiotic tablets, Rosco. The antibiotic tablets produced by Rosco were used in a slightly different method, called the tablet diffusion method. The tablets had the advantage of being more durable and thus could be stored for a longer time. Rosco covered most of the market in Denmark, dominated the veterinary market throughout the Nordic countries, and also sold tablets to a few human laboratories in Sweden and Norway. One way Lappverket differed from Rosco in this period, however, was that they more strongly used science to promote their products. In a Nordic market where the customers were microbiologists at the few primarily state-owned laboratories, and in a context where commercial endeavors within the health field were regarded with suspicion, this was no small advantage. From the beginning of the 1960s, Lappverket developed detailed instructions in order to enable the other laboratories to perform the test in a standardized way (Lappverket n.d.). These brochures had the form of a scientific article, with references, a few tables, a neutral language, and no images. The brochures from Rosco during the same time period were in contrast, according to a Norwegian bacteriologist, “of a more commercial sort,” and “not satisfactory” regarding the description of the methodology and its sources of error (Bøvre Reference Bøvre1962).

During this period, Lappverket started the construction of an informal network of members of laboratories, public health workers, and clinicians from various Nordic countries. These civil servants actively took part in constructing and co-producing the standardization procedures and guidelines. In the before-mentioned Swedish commission on sensitivity testing, 16 laboratories had been included in the investigations in order to explore the existing procedures. This procedure of involving other laboratories, partly to explore existing practice and partly to co-produce standards, was a part of Ericsson's way of working throughout the entire period, and was also his way of working in his local region, the Nordic countries. It also shows how standardization took place as “endogenous” regulatory work pursued not from the outside, by national or regional formal bodies responsible for standardization, but by bacteriologists and scientists as part of their biomedical activities, which in turn established the conditions of possibility for other forms of standardization of the procedures (Keating and Cambrosio Reference Keating and Cambrosio2003; Cambrosio Reference Cambrosio, Gradmann and Simon2010, 253).

Lappverket/Biodisk would arrange training courses in susceptibility testing, give stipends to young microbiologists, and facilitate the establishment of quality control procedures so that each single laboratory could ensure high quality of testing. More importantly, Lappverket introduced a series of regular biannual, two-day meetings, initially at Royal Caroline Hospital, later in the premises of the Swedish Medical Association.Footnote ¹¹ Lappverket/Biodisk would cover the expenses for the participants at these meetings, who included bacteriologists, interested clinicians, public health officials and members of the pharmaceutical industry from all the Nordic countries.Footnote ¹² The conference had no official status, and no mandate to pass binding resolutions for any institution or a single microbiologist. It should instead be a place “to exchange experiences and discuss standpoints for scientists, clinical microbiologists, pharmaceutical industry and manufacturers” (Ericsson and Ericsson Reference Ericsson and Ericsson1975, 1).

Among issues on the agenda in these meetings were evaluation of new antimicrobial agents, discussion of appropriate (or “rational”) antibiotic use, and resistance to antibiotics. More importantly, however, they would discuss what came to be considered crucial factors in the standardization of testing, such as the preparation of antibacterial stock solutions and their storage, the preparation of agar media, the amount of inoculum, the time of incubation, the ideal PH of the solution, and the degree of humidity or temperature.Footnote ¹³ The discussions would result in informal agreements. They were recorded on tape, and then Ericsson would summarize the most important issues in the so-called Minnesanteckningar (minutes or memoranda), that were distributed to each participant after the meetings. For example, in the meeting in 1975, two-hundred transcribed pages were condensed to a 34-page memoranda (Ericsson and Ericsson Reference Ericsson and Ericsson1975). We do not know the extent to which these memoranda were actually used in daily laboratory life in the different local settings. However, we do know that they established a common ground for the participants, formalized the outcome of the meetings, and effectively produced the image of a prevailing consensus. The importance of such memoranda had not been underestimated earlier in Nordic social democratic political history. Einar Gerhardsen, later to become the Labor Party's prime minister in Norway (for a total of seventeen years), claimed in his famous manual for aspiring labor politicians Tillitsmannen (the Shop Steward) that the note taker was one of the most powerful members of the organization (in addition to the treasurer) (Gerhardsen Reference Gerhardsen1930).Footnote ¹⁴ The reason for this was that the note taker constructed the historical events by choosing the information to include, and perhaps more importantly, what not to include in the minutes. Gerhardsen emphasized that this of course did not imply that the note taker could randomly choose what was to be recorded for history, for he was still dependent on trust from the members. Analogically, one might say that the minutes produced after these meetings at Royal Caroline Hospital were decisive. Trust was produced by engaging each individual bacteriologist in the discussions, and letting everyone have their say. However, the minutes were produced by Ericsson alone, giving him the final word. This way of ensuring agreement in print was also likely in accordance with Ericsson's preferred form of debate – oral instead of written disputes. Looking back, he commented many years later that “in my work as a clinical microbiologist I have often found verbal discussions to be both rewarding and stimulating, and sometimes they have sown the seeds of lasting friendship. Discussions in print, on the other hand, tend to generate misunderstandings, even animosity, and therefore are preferably to be avoided” (Arvidson, Dornbusch, and Ericsson Reference Arvidson, Dornbusch and Ericsson1981).

These meetings turned out to be very important, largely defining a common Nordic approach. Through these meetings, persons, knowledge, and things were made to travel in the Nordic region. Media, disks, and strains would be moved from one laboratory to another, scientific papers would be distributed, and the bacteriologists would bring back materials, methods, and skills to their own laboratories. In 1971, Ericsson himself would describe the process in Scandinavia the following way: “Standardization of techniques and materials for the disk sensitivity test has been achieved in Scandinavia by informal collaboration between colleagues and without the intervention of public bodies. Our policy in regard to Single-disk testing is undogmatic” (Ericsson Reference Ericsson1974). Eventually, an entire workforce of clinical bacteriologists was enrolled to carry out this task.

Thus, a Nordic effort of common collaboration and the regular transfer of personnel, things, and knowledge constructed not only a common Nordic platform of working with antimicrobial susceptibility testing, it also helped establish Lappverket, later known as AB Biodisk, as the dominating supplier of material. Though the ICS study carried out by Ericsson and Sherris did not achieve the international recognition it had aspired to, it did become the reference method in the Nordic countries in the 1970s. In Norway, the standardized paper disk method eventually dominated in the Norwegian laboratories. So much so, that at the end of the 1970s, AB Biodisk was totally dominant in the Norwegian human bacteriological laboratories, and until the 1980s entirely displaced its rival in the Nordic market, the Danish producer of antibiotic tablets Rosco (Bergan, Digranes, and Melby Reference Bergan, Asbjørn and Kjetil1996, 7).

Standardizing the Interpretation of Results

Lappverket's version of the disk diffusion method aimed at achieving two objectives that in the beginning of the 1950s had been seen as almost irreconcilable: therapeutic relevance and laboratory reproducibility. The problem with the disk diffusion method was that its results could not be presented in a quantified manner, because many factors other than the susceptibility of a bacterial strain to antibiotics influenced the size of the zone of inhibition. This, in the opinion of many researchers, provided space for subjective interpretation and, consequently, a lack of uniform results. In contrast, the dilution test enabled the quantification of the minimal amount of antibiotics necessary to inhibit bacterial growth (the Minimal Inhibitory Concentration, MIC). In 1954, however, when Hans Ericsson published the first paper of what was to become his doctoral dissertation, he presented for the first time a method that enabled a quantitative estimation of the results of the disk diffusion test. This, he argued, enabled dissemination of valid and reliable information in a standardized way, independent of the individual observer, and thus implied a shift toward a more public form of knowledge.Footnote ¹⁵

At the same time that the randomized controlled trial was being put forth as the gold standard in medical research regarding the testing of new antibiotics, Ericsson tried to eliminate chance in the disk diffusion test by exposing microbes to statistical evaluation.Footnote ¹⁶ In his research, Ericsson would compare large series of dilution tests with diffusion tests for each antibiotic and plot them into a diagram: As he explained, “if the diameter of the diffusion zone is plotted according to a linear scale and the drug concentration according to an exponential (2log), the line obtained is mainly straight” (Ericsson, Hogman, and Wickman Reference Ericsson, Hogman and Wickman1954). Such straight lines could then be directly applied in the laboratories, enabling the direct translation of the size of the zone of inhibition to be read from the diffusion tests and translated to the quantitative value of the Minimal Inhibitory Concentration, without having to perform the laborious dilution tests (“the gold standard”). Bringing mathematical statistics into the disk diffusion test in this way helped promote the creation of the disk diffusion test as a stable entity that could be deployed across distances, independent of the observer.

Why was the elaboration of these regression lines so important? Theodore Porter has argued that a general feature in the development of standardized measuring procedures is that it is understood to be more important to measure in the same way than to measure “right,” and furthermore, that there is a strong incentive to prefer readily standardized measurements to highly accurate ones, when these ideals are in conflict (Porter Reference Porter1994a). With Porter then, we might argue that the quantification of the disk diffusion method helped to ensure that all the collaborating laboratories measured in the same way, and that the results were comparable across different environments (Porter Reference Porter1994a).

Establishing these regression lines however, was complicated and labor intensive, and thus an impossible task for small laboratories. From the end of the 1950s, Lappverket began providing such regression lines for each new antibiotic entering the market, making it attractive not only for smaller laboratories but also for scientific laboratories uninterested in spending time on sensitivity testing research, like the laboratories in Norway. According to the Norwegian microbiologist Asbjørn Digranes, who was active in the Nordic context from the late 1960s, Ericsson even began charging the pharmaceutical industry for this work when new antibiotics entered the market (Digranes Reference Digranes2011). An example from the archive of the Board of Specialities in Norway suggests another importance of these regression lines: considering an application regarding an extension of the area of utilization for Rifampicin, lack of regression lines was used as one argument against such an extension: “disks exist, but not reliable regression lines, and then the testing (with the disk diffusion method) remains worthless.”Footnote ¹⁷

Establishing Breakpoints: Enrolling the Clinicians

However, providing quantifiable results of the response of the microbe to the antibiotic “in vitro,” i.e. in the laboratory, was not sufficient to make it directly relevant to the clinician. Christian Bonah has shown how medical or scientific properties or qualities of therapeutic principles and products “need to be maintained and adapted over time and space within a network of actors and material constraints in order to guarantee the enduring successful use of the product” (Bonah Reference Bonah, Gradmann and Simon2010, 204). The standardization process concerned mediation between the different worlds of the laboratory and the medical therapeutic practice. The test had to be made useful for the immediate clinical context. For the clinician, the important information was how the microbe would respond “in vivo,” that is, in a particular patient. In his first description of his method, Ericsson portrayed sensitivity testing as a cooperative project between clinicians and laboratory technicians. He emphasized that it was necessary through clinical experience and close cooperation between the clinician and laboratory “to provide experience about the relationship between experimentally proved degree of sensibility or resistance and the expected clinical results” (Ericsson et al. Reference Ericsson, Erlanson, Tunevall and Wallmark1953). These arguments are in line with the arguments found in the WHO report on hospital laboratories, for which Ericsson had been secretary. Ericsson's project was intended to promote hospital laboratories as scientific institutions, discontinuous with the mere service functions that such laboratories had often had. However, it is important to underline that Ericsson's approach did not imply hostility towards the medical officers. On the contrary, hospital physicians were given a prominent role in Ericsson's ideal portrait of the hospital laboratory, not only because they helped turn these laboratories into research laboratories, but also because their involvement would strengthen the bond between these laboratories and clinical practice. As Kathryn Hillier has shown, clinical bacteriology arose as a special field in this period (Hillier Reference Hillier2006). The specialists in clinical bacteriology, persons sharing their time between the hospital ward and the laboratory, were Ericsson's favorite members of the hospital laboratories, and to him they were necessary also for the development of sensitivity testing as a scientific endeavor.

Since the microbe's response to the antibiotic, as it was measured in the in vitro conditions in the laboratory, was only one element among many variables in the clinical situation in the living bodies, a series of other factors had to be taken into account when providing the results. The concentration at the infection site was important – in the blood, the urine, or other body fluids. In his doctoral dissertation, Ericsson had compiled data on serum and tissue levels obtainable with ordinary doses of antibiotics, which provided an alternative, clinically relevant, sensitivity criterion. Using these data, the laboratory results could be translated to categories relevant to the clinicians, thereby establishing sensitivity as a clinical, not a bacteriological phenomenon: a bacterial strain could be considered sensitive to an antibiotic when the minimal inhibitory concentration (MIC) of the agent was lower than commonly achievable blood or serum levels (Ericsson Reference Ericsson1960b). Already in the report given from the Swedish committee in 1953, the critical values of concentration of the different antibiotics in different parts of the body after the then common dosage were provided.

Further, since the clinician could not be expected to have a command of bacteriology to such an extent that she would be able to directly make sense of the information from the laboratory, it was suggested that the results be translated into categories that she could easily read. In the Swedish report from 1952, the committee recommended that “four different groups of resistance” should be the basis for the information given to the clinicians (Ericsson et al. Reference Ericsson, Erlanson, Tunevall and Wallmark1953; Ericsson Reference Ericsson1960b). The aim was to convert MIC values into susceptible or resistant categories, i.e., to assess whether it was possible to treat an infection by a given antibiotic. In the report from 1952, the categories were labelled I (sensitive [känslig]), II (relatively sensitive [tämligen känslig]), III (hardly sensitive [föga känslig]), and IV (resistant [resistent]). The boundary values (MIC values) would divide the susceptible from the resistant bacteria, and the assignment to one category would enable the prediction of the patient's response when treated with the antibacterial drug.

In an international report published almost 20 years later, Ericsson and his co-author Sherris emphasized that those categories should be worked out and reviewed from time to time between bacteriologists and infectious diseases clinicians, thus ensuring the tight connection between the clinician and the clinical bacteriologist (Ericsson and Sherris Reference Ericsson and Sherris1971). They would still defend the use of four sensitivity categories, although four categories were not found within any of the international standards at that time. Rather, most laboratories around the world were using the three-categories system, introduced in an article from 1966 on a standardized single disk method – interestingly, an article which Sherris himself had co-authored. This method, which later came to be referred to as the Kirby-Bauer method, defined three categories of sensitivity instead of four: sensitive (S), intermediate (I) and resistant (R), and became widely accepted. However, Ericsson and Sherris in the report from 1971 decided to keep the four categories, mainly because they found it important that there was reserved a place for those infections where the concentration “in vivo” could be sufficiently high. In this group, mainly consisting of urinary tract infections, an aetiological agent which in the other groups would be categorized as resistant to an antibiotic, could be reported as sensitive (because the antibiotics could achieve a very high concentration at the infection site, e.g. in the urinary tract). As in the report twenty years earlier, the four categories of sensitivity were identified by Roman numerals (the reason for using numerals, not names, was again related to the issue of standardization: names would make them difficult to compare across national languages): Group I: High degrees of bacterial sensitivity that made in vivo [clinical] response probable when mild to moderately severe systemic infections were treated with the usual dosage of the antibiotic. Group II: Degrees of sensitivity which made in vivo response probable in systemic infections when the antibiotic was given in high dosage or up to the limits of toxicity. Group III: Degrees of sensitivity which made in vivo response probable in the treatment of localized infections at sites where the agent could be concentrated by physiological processes or local application [e.g. urinary infections]. Group IV: Organisms of a degree of resistance which made in vivo response improbable (resistant).

In the 1971 report, Ericsson and Sherris for the first time introduced the word “breakpoint” as a concept referring to the boundary values between the different categories. Since resistance in the laboratory (in vitro) and resistance in the clinic (in vivo) was not the same thing, since many more factors influenced the concentration of the antibiotic at the infection site than in the laboratory, they argued that breakpoints should be established on the basis not only of the quantitative estimation of the response in the laboratory (the Minimal Inhibitory Concentration), but also on the basis of clinical considerations. Kirby, Bauer, and Sherris had not used the word breakpoint in their article from 1966 which gradually in the 1970s came to be the standard point of reference (Bauer et al. Reference Alfred, Kirby, Sherris and Turck1966), but they were offering a table in the article based on the boundary values of 20 different antibiotics. In the 1970s, these boundary values were subjected to much discussion, as they were increasingly considered tools for epidemiological management of resistance.

The WHO Study

The international collaboration study mentioned above, which Ericsson and Sherris published in 1971, was actually a result of a work commissioned by the WHO in 1962. This year, at the International Congress for Microbiology, it was decided that a working group should start a project, to be jointly organized by the WHO, the FDA, and the Royal Caroline Hospital.Footnote ¹⁸ The goal was the universal adoption of “reliable methods, standardized as far as possible” (Ericsson Reference Ericsson1964). Like the study conducted by the Swedish committee that Ericsson had been a part of ten years earlier, this research consisted of a comparative study of the methods employed in different laboratories – but this time on a global scale. In the WHO study, 16 laboratories, from Eastern and Western Europe and the United States, agreed to collaborate. Of the factors involved in sensitivity testing, they chose to study the culture medium, the inoculums (influence of heterogeneity, size), the addition of the antibiotic to the medium, the reading of the end point the Minimum Inhibitory Concentration (MIC), and the incubation temperature and time. Already in the preliminary report of 1964, the working group had published detailed instructions on how to standardize the test procedures (Ericsson Reference Ericsson1964). Here they concluded, as the Swedish committee had done more than ten years earlier, that none of the laboratories had the same standard procedure.

The final report was not published until 1971, and did not in the end gain international support as the reference method. (For a thorough discussion of the international discussions related to the WHO report, see Gradmann Reference Gradmann2013.) Already the publication channel signalled what was to come: the report was published not as a part of a WHO series, but as a supplement to the Scandinavian journal Acta Pathologica et microbiologica Scandinavica. The method that finally gained ground when standards for susceptibility testing were first put in place in the United States (the NCCL and FDA) and Europe during the 1970s, was the Kirby-Bauer-method (Bauer et al. Reference Alfred, Kirby, Sherris and Turck1966). This method was, except for a few technical details,Footnote ¹⁹ identical to the method that Ericsson had promoted in his earlier version of the single disk testing, and also identical to the method later promoted in the ICS report. One of the major criticisms towards the ICS report was that the standardization procedures were far too difficult to carry out for the laboratories (Garrod and Waterworth Reference Garrod and Waterworth1971). However, the report studiously avoided recommending standard methods for universal use. Rather, reference methods were proposed against which locally accepted techniques could be evaluated (Ericsson and Sherris Reference Ericsson and Sherris1971). The report recommended that national reference laboratories should issue full instructions on methods and their interpretation in each country, to enable comparison between the results of workers in different parts of the world who were investigating new or old antibiotics, doing epidemiological studies, or reporting results on clinical data. They did not recommend development of internationally accepted standard methods for day to day application, because any such suggestion would “unlikely be acceptable” to many laboratory workers, and could discourage “methodological development and thus entrenching concepts which may become obsolete in the future” (Ericsson and Sherris Reference Ericsson and Sherris1971). However, they did state certain guidelines for the principles on which such reference methods should be based.

Thus, although the criticism raised might seem unfair in the aftermath, there are interesting aspects of this debate to take note of. Many of the international collaborators commented on the fact that Ericsson tried to export the Scandinavian model onto the world, and that the detailed processes of standardization they suggested would be impossible to carry out in smaller laboratories (Gradmann Reference Gradmann2013). While this international attempt at standardization is not the subject of this article, it is interesting to note that the similarities to the Nordic standardization endeavors are striking. The International Collaborative Study (ICS), like the Nordic standardization endeavors, aimed at a standardization of the entire sensitivity testing procedure, including the paper disks, but also the media, the time of incubation, the pH level, and the system of reference strains. In the next paragraph I will discuss how these procedures were made possible by particularities of the Nordic countries such as the existence of few and public laboratories, relations of trust, and the ideal of equality. Ericsson's attempts at universal application of the suggestions in the WHO study, however, did not take into account that these elements did not translate easily to the rest of the Western world. The failure of the ICS system was, however, also a result of broader geopolitical arrangements. The American economic and scientific hegemony at the time was after all, as John Krige and Dan Carpenter have pointed out, partly a result of the standardization of many procedures and objects that had before been left to individual, experience-based decision-making (Krige Reference Krige2006; Carpenter Reference Carpenter2010).

While Ericsson and his collaborators did not achieve the widespread application they hoped to achieve for the ICS project, the Scandinavian standardization collaborations continued. In fact, given the fate of the WHO report, the Nordic region was probably the only geographical area where the ICS method was employed extensively. Being the method used in the Scandinavian countries also before the publication of the report in 1971, the ICS system was used in all Swedish and most Norwegian and Finnish bacteriological laboratories until 1979 (Alestig et al. Reference Alestig, Bertler, Brorson, Burman, Dornbusch, Ericson, Kallings, Kamme, Malmborg, Nordbring, Norrby, Olsson-Liljequist and Wallmark1981). In the following section, I will discuss the context in which these standardization processes took place, and, conversely, how these processes were at the same time constructing the same contexts.

Constructing Standards, Constructing the Nordic Region

The efforts to construct a common Nordic approach to susceptibility testing are hardly thinkable without considering the central characteristics of Nordic societies, first of all, by the network of a few, publicly sponsored, elite laboratories. This enabled a uniform method for proceeding and facilitated the formal and informal exchanges necessary for the longevity of the project. As one editorial commented in the British Medical Journal in 1974: “Not all laboratories can be trusted to conform to every detail of such a technique [the ICS method]” (Editorial 1974). Ericsson envisaged the new laboratories as places of science. In his view, they should be located at medical schools, incorporating medical expertise, and be places for research and training. Bacteriological laboratories attached to university clinics had of course, also been considered scientific institutions. Ericsson emphasized that scientific aspirations should also include clinical bacteriology, focusing, e.g., on the improvement and standardization of procedures necessary for clinical work, such as sensitivity tests, not merely traditional bacteriological research. These comprehensive standardization procedures were far easier to carry out in the Scandinavian system of very few state owned, well-equipped laboratories filled with personnel doing research within their testing procedures. In fact, it might be argued that Nordic standardization procedures suggested an organization similar to that already established in Nordic countries.

Because most health services provided in Norway and Sweden at this time were publicly financed, health-care workers in the Nordic region were not used to commercial competition, and they were therefore probably more likely to accept the provision of instruments and equipment from one contractor than from several. This last aspect was important, as the replicability of the results and the validity of the regression lines depended on the use of identical methods, including identical disks and identical media. The method suggested in the Nordic standardization processes made it necessary to ensure that each step in the overall process was managed and regulated so that the integrity and efficiency of the production process as a whole could be maintained. Again, this way of proceeding was facilitated by the existence of public, centralized quality laboratories, which were assigned an important role in maintaining appropriate standards.

By contrast, in other countries, and in the United States in particular, disks were often provided by private manufacturers that seldom had the ability to do the groundwork, e.g., by supplying regression lines that the Nordic standardization procedures presupposed. This aspect was also highlighted in one of the critical commentaries on the WHO report from 1971: “It should be clearly understood that the preparation of satisfactory regression lines is an enormous task, and that they are only valid for tests on precisely the same medium as that used in constructing them. Those who are now asking disk manufacturers to provide them with regression lines are presumably unaware of this” (Garrod and Waterworth Reference Garrod and Waterworth1971, 780). Thus, to a certain extent, the agreement of the Kirby-Bauer method by the NCCL in 1976 can actually be seen as a decision also on what NOT to standardize, given that one of the arguments raised against the ICS was that it was too standardized.

Finally, I would like to argue that the standardization procedures in the Nordic laboratories embodied values related to the Nordic postwar political project of trust, universality, and equality. As Fujimura has argued, “scientific technologies are highly elaborated techno-symbolic systems” (Fujimura Reference Fujimura1996, 31). They are, in her opinion, meaning-laden and meaning-generating systems, just as language and writing. Scientific technologies represent values and commitments to the ideals of invariability, homogeneity, standardization, and reproducibility. Such values are also real laboratory facts embodied in these standardized procedures. In the Nordic setting, an additional value that may be the most central value in the Nordic model related to the ideal of the standard, is the ideal of equality and the public good. Many researchers have suggested that egalitarian individualism has been a particularly characteristic aspect of Nordic countries (Barnes Reference Barnes1954; Gullestad Reference Gullestad1984). The most common translation of the Nordic and Swedish concept of likhet is equality, but it would probably better be translated as “likeness, similarity, identity, or sameness” (Gullestad Reference Gullestad1992). An important part of postwar welfare practices in Nordic countries was that health services should be delivered in an equal (standard) way to everyone. The comprehensive standardization procedures in the laboratory, including every single step of the process, where every single worker had their own task which had to be fulfilled in order for the procedure to be completed in accordance with the standardization demands, was, as the editorial in BMJ referred to above, not easy to accomplish. Standardization of the whole process, implying that comprehensive procedures depending on all the workers in the laboratory regardless of position could thus be portrayed as a real laboratory fact embodied in the procedures. On the other hand, these procedures also actively produced the ideals of equality and comprehensiveness by making everyone in the laboratory a crucial member of the procedure.

The Nordic countries have consistently been among the highest-ranking countries in cross-national surveys measuring trust and/or civic engagement in the past 30 years (Delhey and Newton Reference Delhey and Newton2005; Skirbekk and Grimen Reference Skirbekk and Grimen2012; Trägårdh Reference Trägårdh2009; Wollebæk Reference Wollebæk, Wollebæk and Segaard2011). The proportion of people agreeing that “most people can be trusted” is almost twice as high in Norway as in the United States (Delhey and Newton Reference Delhey and Newton2005). Explanations for this high level of trust have included a high degree of cultural and economic homogeneity in the Nordic societies as well as a gradual democratization based on peaceful mass mobilization (Delhey and Newton Reference Delhey and Newton2005; Wollebæk Reference Wollebæk, Wollebæk and Segaard2011), laying the foundations of a strong civil society. Norway is also characterized by a strong relationship between interpersonal and institutional trust (Wollebæk Reference Wollebæk, Wollebæk and Segaard2011). Trust was crucial in the procedures of standardization that I have described here, and the informal character of the meetings and the development of standards is a characteristic of the whole process. Of course, the limited amount of personnel in the rather small Nordic countries made it easier to maintain personal informal relations in scientific networks. However, also at a more institutional level, informal procedures were common. In this period, there was an emphasis on Nordic cooperation and harmonization at the inter-state level. All attempts at grand design, such as a defense union or a customs union, failed; but more informal attempts at cooperation were extremely successful. The Nordic Council (1952) and the Nordic Council of ministers (1972) are examples of very successful regional arrangements, including for instance a common labor market and a passport union (Markkola and Christiansen Reference Markkola, Christiansen, Christiansen, Petersen, Edling and Haave2006, 28). Another example in the health sector is the Pharmacopoeia Nordica, which was the first pharmacopoeia to be authorized in more than one country.

However, it might be argued that not only was the Nordic context essential for the initial success of the standardization procedures, these standardization procedures actually contributed to the construction of this very context as well. The important role of materiality in building communities in the tradition of cultural studies has been emphasized. For instance, for the Swedish case, it has been claimed that electrical currents and sewage systems were more important elements in the Swedish nation building than more iconical versions of this process, such as Per Albin Hansson's appeals in the National Assembly (Ehn, Frykman, and Löfgren Reference Ehn, Frykman and Löfgren1993). Kristin Asdal has shown how measurement of water pollution took part in constructing nature as a general and comprehensive category in Norway (Asdal Reference Asdal2004 and Reference Asdal2008). In a similar vein, one might argue that the production and circulation of standardized susceptibility testing techniques were small steps towards drawing Scandinavia together, creating a Nordic common territory, which was perhaps more important than the public rhetoric of the importance of a common Nordic health policy. Through binding people, laboratories, materials, and methods together in tight cooperation, the efforts towards standardization were factors that helped to put such a common policy in place. In fact, these new efforts at standardization helped create a whole workforce of people from different fields of medical practice – public health personnel, bacteriologists, infection specialists to name only a few – working together towards a common goal. Thus, in an analogous manner, Scandinavia was constructed as the most unified region taking place within Nordic politics at the time.

The Nordic Standardization Project Fails: Drug Regulation Policies and Breakpoint Controversies

In the 1970s, the main controversy in the negotiations of standardization in Nordic countries was the definition of clinical breakpoints. AB Biodisk held two large Nordic meetings in the 1970s,Footnote ²⁰ and they marked the end of the era of AB Biodisk's dominance. At the conference in 1975, a discussion arose concerning group classification of sensitivity. The background was AB Biodisk's presentation of their “adaption to Scandinavian norms” of Pfizer's new semiautomatic system for determining antibiotic sensitivity, the Autobac System (Ericsson et al. Reference Ericsson and Ericsson1975). A suggestion for the establishment of a reference group for medical microbiology from the microbiologist Gösta Wallmark was accepted, and the mandate was given to the Society for Medical Microbiology within the Swedish Medical Association to establish such a group. In the discussion that followed there was agreement to include not only bacteriologists, but also clinicians in all of the Nordic countries.

This group – the Swedish Reference Group for Antibiotics, Referensgruppen för Antibiotikafrågor, (RAF) as it was eventually called – was formally established in 1976, as a joint venture by the Societies of Medical Microbiology and Infectious Diseases of the Swedish Medical Society. The members of the group were from several sections within the Swedish medical society: infectious diseases, clinical microbiology, pediatrics, otorhinolaryngology, clinical pharmacology, and later general medicine. Hans Ericsson was conspicuously absent. The main aim of the group was to re-evaluate the ICS system after five years, especially with regard to the breakpoints. One of the arguments was that the amount of antibiotics had increased considerably since 1953, when the four-category-system was introduced. In 1977 the reference group presented a revised system for sensitivity testing, which was applied in Sweden from 1979. Two main changes were made to the ICS system. It introduced a 3-group system (SIR – Susceptible, Intermediate, and Resistant) describing the different categories of resistance. The argument was that this system conformed better with the other systems used in Europe and the United States. Secondly, the group introduced new border values or breakpoints (Alestig et al. Reference Alestig, Bertler, Brorson, Burman, Dornbusch, Ericson, Kallings, Kamme, Malmborg, Nordbring, Norrby, Olsson-Liljequist and Wallmark1981).

In the minutes from AB Biodisk's conference in 1975 it was emphasized that members of Nordic countries should be included in this group (Ericsson and Ericsson Reference Ericsson and Ericsson1975). Already a week after this conference, the boards of the Norwegian Society for Infectious Disease and the Norwegian Society for Pathology (Norwegian microbiologists were members of this Society) sent a letter to the Swedish Medical Association, referring to the discussions at the conference. They expressed interest in participating in a Nordic collaboration with the aim of developing common breakpoints and a common system of sensitivity classification.Footnote ²¹ The Swedish group had commented in their first report that they had discussed the “Nordic question” thoroughly (inngående). They agreed that the current Nordic collaboration should continue, and that it should also include the clinicians. However, because treatment regimes were different across the Nordic countries, and because it was considered desirable to keep the group small, the reference group decided not to include the other Nordic countries in their work (Referensgruppen för antibiotikafrågor 1977). As a result, in 1978 the Microbiological and Infectious Disease Associations created their own national body, the Norwegian Working Group for Antibiotics (Arbeidsgruppen for Antibitokaspørsmål, AFA). This was the end of the efforts to arrive at a common Nordic platform on issues of standardization. From then on, the Norwegian and the Swedish groups worked separately on these issues.

One reason for the different treatment regimes was that in spite of profound similarities regarding the public health system, there were also considerable differences, especially regarding drug policy. The common features of the Nordic social models often results in a picture of the Nordic countries as one homogenous block. Historically, however, there are significant differences among Nordic countries, also in this respect. The industrialization in Sweden took the form of relatively large-scale capitalist manufacturing, which enabled the establishment of a considerable international pharmaceutical company like Astra. In Norway however, the Norwegian authorities developed a very restrictive practice regarding drug regulation (Dølvik Reference Dølvik2007, 10).

Sweden, with a larger and more independent pharmaceutical industry, was open to innovation and research to a much larger extent. Norwegian politicians and public drug regulatory authorities, on the contrary, exercised a strong influence over the pharmaceutical market from the early postwar years until the mid-1990s. Since 1951, Norway had a public monopoly on the wholesale distribution of drugs, the Norwegian Medicinal Depot. The country had a centralized drug administration in the Control of Specialities (after 1974, the Norwegian Medicines Agency) (for Norwegian drug regulation policies, see Dyrdal Reference Dyrdal2001, Reference Dyrdal2002, Reference Dyrdal2004; Moen Reference Moen1997). Prior to approval of drugs, the manufacturer had to make sure that the drug was not only effective and safe, but also that it was needed (the so-called “need clause”). This clause, which dates back to 1938, existed only in Norway (and to a lesser extent in Iceland), and was used actively to delimit the amount of available drugs in Norway (Jøldal Reference Jøldal1986). A profound skepticism toward market regulation in its entirety and the idea that professional judgments related to health and medicine should not be affected by private economic interests was the background for this very strict regulation of drug policy (Moen Reference Moen1997).

These differences in drug regulation policy were to be of importance for the fate of the Nordic susceptibility testing standardization project. In the 1970s, single-dose antibiotics had been introduced to the market, but they were still seldom used in Denmark and Norway. In Norway, several of these new drugs had been declined by the Specialities Commission, due to their failure to satisfy the “need clause.” The old penicillins were still popular with the health authorities, although the doses had to be scheduled three to four times a day. Besides being cheaper, they did not promote resistance in the same manner as the new, broad-spectrum drugs. Thus, although the new antibiotics were easier to use for the patient because some of them required only one dose a day, they were regarded with skepticism by the Norwegian drug authorities.Footnote ²²

However, in Sweden, the Swedish drug regulation authorities were less skeptical towards these pharmaceutical newcomers. As breakpoints depended not only on the in vitro sensitivity of the microbe to the antibiotics, but also on the drug dosage and local therapeutic traditions, this was a factor to take into account. Secondly, according to the microbiologist Tore Midtvedt, one of the central actors in antibiotic policy in this period both in Norway and in Sweden, susceptibility breakpoints were often used to influence prescribing patterns and were therefore considered essential (Midtvedt Reference Midtvedt2010). Agreeing on breakpoints thus had other than technical interest.

But there are also other possible explanations for the failure of this project that do not relate directly to internal Nordic affairs. As the ICS report failed to be the universal standard method, the Nordic countries were increasingly marginalized in regard to the rest of the world of clinical microbiology. In the report from their first meeting, the Swedish reference group for antibiotics emphasized the importance of adjusting Swedish practices to conform to international agreements. The introduction of a 3-group system by the Swedish group equivalent to the American categories (SIR – Susceptible, Intermediate, and Resistant) is one example. Moreover, the 1970s were also a difficult time for the efforts at a Nordic integration, which had been an important part of postwar Nordic policy. Denmark entered the European Economic Community in 1972, and increasingly, emphasis was put on European, rather than on Nordic collaboration.

Producing Standards in the Nordic Welfare State

Producing standards are never easy. In this article, I have tried to show how the method that eventually became known as the Paper Disc Method gradually developed as the standard method in Sweden, Finland, Norway, and parts of Denmark, but also how it eventually failed. The initial success of the Nordic standardization project depended on a set of circumstances, including the strong Swedish research tradition in clinical microbiology, and the successful enrollment of researchers, materialities (paper disks and culture media), and microbes in a Nordic network.

In a way, this is a story about Hans Ericsson, a very successful network-builder, who aspired to acquire control of bacteriologists, microbes, paper discs, pharmaceutical representatives, antibiotics, and health policy makers by enrolling them all in a stable network in which they acted in concert, while still striving for their own interests. It might be argued that Ericsson, like Louis Pasteur in Bruno Latour's discussion of him (Latour Reference Latour1988), acted as a spokesperson in control of his network. For a considerable time, this network was stable enough to create not only a common Nordic standard of sensitivity testing, but also to create a successful manufacturing company for the production of paper disks. Ericsson and Lappverket/Biodisk translated their discoveries into operational diagnostic tools, and this process implied the establishment of proper production methods, common measurement techniques, and product standards. Although Hans Ericsson and Lappverket/Biodisk were central in the process, their success depended upon their ability to mobilize these other human beings and entities. Together, they organized and guaranteed circulation and control of tools for antibiotic sensitivity testing. They thus played a crucial role in the standardization of research practices, linking the laboratory and the clinic to industry.

The standardization process produced not only a set of standardized procedures; in turn it also influenced and changed the context of its origin. The establishment of agreed methods for standardizing the testing provided a platform for negotiating common interests between different groups of practitioners. The formation of networks of alliances and exchanges between industrial and academic scientists and the clinical institutions within which they worked, was crucial in establishing an effective basis for collaboration and exchanges across the Nordic countries, and thus in constituting the Nordic infection-control specialist.

However, towards the end of the period this struggle was no longer successful; the different interests of all the network elements could no longer be aligned. The paper disks produced by Biodisk were still dominating the market, and the procedures of performing it were still followed. However, other elements of the “standardized package” were increasingly contested. The most controversial element was the interpretation of the test results. Ericsson ceased being the spokesperson of his own network, and new national networks were formed. This nationalization included even the microbes, which, sensitive to the national drug regulation polices and drug use, had developed (or could develop, so it was argued) different sensitivities to antibiotics in the different Nordic countries, and thus demanded different sets of breakpoints.

The idea of a central spokesperson does not entirely capture the process in the Nordic countries however. It fails to acknowledge several pivotal points in these efforts at standardization: First of all, it fails to recognize how elements and characteristics of what have been called the Nordic welfare states influenced the process both by contributing environments that facilitated the introduction and maintenance of these comprehensive standardized procedures (a few state-owned medical research laboratories without their own commercial interests) and by the inclusion of values such as equality and relationships of trust as real laboratory facts embodied in the standardization procedures. Secondly, and related to the first, these standardization procedures were not characterized primarily by struggle, but by co-production. Crucial for the initial success of the standardization procedures was their character of being pragmatic, local, and flexible. The standardization procedures were successful not because a stable network was created, but because the creation of the network was itself a collective effort. Ericsson and his enterprise Lappverket were facilitators by arranging these conferences and supplying the materials, but the actual performance of the procedures were decided upon collectively in the conferences, and locally in the practice in the laboratories. Together, they developed shared languages, ways of communicating, shared tools and theories (Fujimura Reference Fujimura1996, 12). Stefan Timmermanns and Marc Berg have shown how the origin of universal standards is the result of historically situated, distributed work of a multitude of actors (Timmermans and Berg Reference Timmermans and Berg1997). In this line of argument then, it might be argued that when the Nordic standardization project failed, a main reason was that the standardization work was no longer coordinated and maintained sufficiently. As Fujimura emphasizes, standardization is a work that has to be continually maintained – it has to be worked with carefully and continuously.

However, it was increasingly difficult for Ericsson and Biodisk to be the center of these ongoing efforts of standardization, for different reasons. As we have seen, Biodisk was crucially involved in the production and circulation of standardized materialities (paper disks and culture media), procedures and methods, as well as in the commercialization of instruments for the performance of these tests. Thus, commercial interest promoted standardization of work, and the standardization of work promoted commercial interest. The link between Biodisk, the laboratories in the Nordic countries, and the social system in the Nordic welfare states were thus manifold and complex. However, in the Nordic welfare states in the 1970s, there was an increasing skepticism towards instruction manuals and guidelines produced by commercial bodies. For example, in the field of pharmaceuticals, increasing emphasis was put on corporate independent information on drugs in this period.Footnote ²³ Guidelines should no longer be left to commercial interests; they should be performed by formalized regional and national bodies. Whereas Ericsson had the ambition to enroll the whole world in his standardization process, his company Biodisk increasingly came to be seen as an actor with a commercial interest, not as a guarantor of scientific truth. However, although Biodisk ceased being the center of development of standards, it continued as an increasingly successful commercial manufacturer. Actually, it became a worldwide leading supplier of susceptibility testing material after the development of a further test, the Epsilometer testing method (E-test) in 1991.

The efforts at reaching a consensus on international guidelines regarding susceptibility testing were partly given up during the 1970s. As Alberto Cambrosio has pointed out, regulatory decisions include decisions about what to standardize and what NOT to standardize (Cambrosio Reference Cambrosio, Gradmann and Simon2010). The NCCL decision to standardize only the disks, not the whole procedure, in 1975, is a typical example. Calls for a comprehensive international standardization, however, continued to arise, and they reached the political agenda anew when antibiotic resistance became acknowledged as a major global health challenge during the 1990s. It was not easy this time either. Although the need for international, uniform, and standardized methods were increasingly asked for, the first European guideline for standardization of disk diffusion tests only appeared in 2009 (EUCAST), and it is still contested. Consensus is not an achievement that can be taken for granted. But that is another story.