The tests delivered in this setting focused on loss of functional ability or on level of impairment, but since their administration is standardized, the results will not capture the individual characteristics of the patient . Yet, this study’s findings suggest that individual patient characteristics are noticed and made relevant in the clinical use of test information.
The two components of test information
The primary component of test information is gathered in the test situation, where it is apparent that therapists are not only test administrators; they are also observers. The therapists see the individual they test; they see their patients. During testing, they take in the patient’s physical and cognitive functioning, emotional state, coping strategies, conduct, and ability to take instructions. In fact, these observations are, in the therapists’ accounts, often presented as the therapists’ key concerns and they can be used to support or challenge decisions regarding patients’ forthcoming activity and treatment plan: the patient is sad, the patient needs to use a walker, or the patient is slow/fast and careful/reckless. Such concerns and typologies resonate with Thornquist’s  portrayal of therapists as attentive to patients’ subjective experiences and to their functional abilities. On the other hand, concerns, such as the ones presented here, may also influence the clinicians’ ability to score their own patients accurately [37, 38]. It may be the therapists’ twin position, as test administrators and as the particular patient’s therapists, that makes them attuned to collecting information that extends beyond what standardized testing deems significant. One example of therapists’ collecting information that extends beyond the standard is provided when therapists note patients’ coping strategies; another example is provided by the therapist who claimed patients’ malingering in tests was discovered “because we see the patient during the whole day” (PT13). If tests were delivered by a technician, as suggested by DeLuca and Putnam , this information would likely be lost. In actual fact, the therapists’ broad approach to test data suggests that they do not heed the underlying distinction between testing and assessments; the subjective component present in health assessments should be absent in standardized testing .
The secondary component of information falls, principally, in the category of quantifiable test data: end scores. Scores and end scores provide the health care professional with quantifiable documentation on patients’ status quo functional ability. Insights on how therapists use quantifiable data can be summed up by Fujiura and Rutkowski-Kmitta’s statement: “Numerical associations facilitate independent verification, standardization, and economy of communication” (p92) . There were no independent verification procedures in this clinical setting, because no disinterested third party was involved in test interpretation. However, involvement by interested third parties could occur when therapists discussed observations and test scores among themselves or in the multidisciplinary team. In addition, therapists expressed a notion of trust in standardized tools as objective. They compared, for instance, findings from standardized tests to pathological manifestations visible in blood samples and CT scans. Trust in standardization was also demonstrated when scores from previous hospital stays were compared to the patient’s new scores.
Understanding the numbers
The two components presented in our study find a parallel in Polanyi’s  distinction between tacit and explicit knowledge. Tacit knowledge is subjective and created through direct experience . Tacit knowledge, thus, embraces an array of conceptual and sensory information and images (we know more than we can tell (p4) ), whereas explicit knowledge is the knowledge we are able to articulate, standardize, codify, and store. In line with Greenhalgh et al. , our study brings to the fore the interaction between tacit and explicit knowledge in the use of test information. The guiding role of subjectivity in the therapists’ use of “objective” information illustrates how “facts,” such as test scores, do not speak for themselves, but instead are interpreted and translated [2, 43]. Thus, information is not given or is not “the outcome of individual minds, operating in a social vacuum” (p54) ; rather it is the result of a continuous collective interactional activity that produces, interprets, and translates it from one setting to another.
At the outset, numbers are considered objective, and in the therapists’ accounts, objective data, that is, numerical data, are associated with quality, reliability, and credibility. A similar association between objective data, reliability, and quality is noticeable in the rhetoric surrounding continuous upgrading of health care provision (see, for instance, [5, 6, 44]). Still, with regard to the expressed credibility of objective data found in our material, we noted ambivalence among participants: Objective data, numerical data, were often depicted as mere black and white and of limited use or value to clinicians – simultaneously, scores were frequently used in communication. This brings us to a main finding regarding the secondary component of information, a finding that concerns what Fujiura and Rutkowski-Kmitta label “the economy of communication.” The economy of communication on the wards studied here is seemingly sustained in a multifaceted communication practice that, in fact, goes beyond numerical representation. Thus, we argue that to the therapist familiar with the specific standardized test, the score numbers contain information that goes beyond mere numerical representation. Test scores state level of impairment, often in relation to a normative sample and are, as such, encoded . Knowledge of a test’s scoring system and its normative sample is necessary in recognizing the level of impairment indicated by the end score . Atkinson  describes information or encoded knowledge as embodied in different forms of representation (test scores, laboratory test results, MR printouts). His perspective underlines not only that tacit knowledge is key to the production of scores, but also that it is key in generating and maintaining the scores as explicit knowledge. Although encoded knowledge does not preserve the tacit skills of the individuals generating it , it provides the therapists with a common language, essentially a shared understanding, of scores. This common understanding facilitated communication with colleagues, as well as communication with the districts’ needs assessment offices, but it seemed to fail in communication with patients. Patients, as opposed to colleagues, had no understanding of the message in numbers, and therefore had to have them explained. Therefore, in providing test feedback to patients, the primary component of information was used as the main information source. In practice, the therapist would communicate a contextualized image of a decontextualized test to the patients [16, 45].
The ambivalence noted among the participants regarding numerical representation was not directed at the scores’ inability to provide insight into level of impairment; it was directed at the scores’ inability to capture patient’s characteristics [3
]. Our analysis shows that, in line with research conducted from an interactional perspective, patients’ characteristics and the context are relevant in face-to-face standardization. In interactional-oriented research, test administrator characteristics, patient characteristics, wording, and context have been shown to affect test results. This study, however, suggests that only patient characteristics and context are made relevant when test administrators justify their use of test information. The fact that test observations routinely were made known in the form of written caveats illustrates the therapists’ wish to contextualize patient performance. In practice, caveats render visible tacit knowledge in standardized outcome measures: the manner in which clinicians’ intuitive judgment, reasoning, and expertise are used to supplement, dismiss, or adjust scores [27
]. Thus, caveats highlight what the end score could not: the patient’s problem – “we are more concerned with the problem than with the actual end score” (OT6).
This practice also underlines the therapists’ pragmatic stance towards testing. A similar approach to test interpretation is found in Dingwall et al. [18
]. Caveats were especially important when a mismatch between patients’ observed behavior and the end score was noted. Therapists’ uses of caveats provide an example of how:
External clinical evidence can inform, but can never replace, individual clinical expertise, and it is this expertise that decides whether the external evidence applies to the individual patient at all and, if so, how it should be integrated into a clinical decision. (p71) 
The objective contribution of standardized tests proved to be moderated by caveats. Caveats were actively used in seeking agreement between the subjective and objective components of test information. Therefore, therapists challenge the sole use of one of the components.