Article Text

Download PDFPDF

Acting by persuasion— values and rhetoric in medical certificates of work incapacity: A qualitative document analysis
  1. Guri Aarseth1,
  2. Bård Natvig1,
  3. Eivind Engebretsen2,
  4. Anne Kveim Lie3
  1. 1 Institute of Health and Society, Department of General Practice, University of Oslo, Oslo, Norway
  2. 2 Institute of Health and Society, Department of Health Sciences, University of Oslo, Oslo, Norway
  3. 3 Institute of Health and Society, Department of Community Medicine and Global Health, University of Oslo, Oslo, Norway
  1. Correspondence to Dr Guri Aarseth, Institute of Health and Society, Department of General Practice, University of Oslo, Oslo 0318, Norway; guri.aarseth{at}


When the patient applies for disability benefit in Norway, the general practitioner (GP) is required by the National Insurance Administration (NAV) to confirm that the patient is unfit for work due to disease. Considering the important social role of medical certificates, they have been given surprisingly little attention by the medical critique. They may make essential differences to peoples’ lives, legitimise large social costs and, in addition, the GPs report that issuing certificates can be problematic. This article explores values, attitudes and persuasive language in a selection of medical certificates written by GPs. We direct attention to such texts as significant social actors using a mixed rhetoric including certain values and attitudes. When arguing for granting the patient disability benefit, some GPs emphasised the ‘worthiness’ of the patient by pointing to positive attitudes approved by the national insurance: a will to work and participate, to cooperate and be motivated. Others pointed out the patient’s positive character in terms of universally accepted values, called for the reader’s (the NAV official) sympathy , understanding and helpfulness or appealed to his/her willingness to be realistic and pragmatic and grant disability benefit (DB). The dialogic style varied: some certifiers—although they argued for disability benefit—showed openness to possible opposing or alternative voices by displaying their own uncertainty. Others addressed the reader to share responsibility, demanding or urging for DB. This shifting rhetoric, we believe, mirrors that the GPs see themselves as the patient’s advocate, and that they may find themselves conflicted. We propose further studies within qualitative research to investigate the effect of this rhetoric on the reader, the decision-makers. In addition, to improve the quality and accuracy of these important documents, we suggest that medical schools introduce students to the making of text as a specific skill of medical practice.

  • primary care
  • medical humanities
View Full Text

Statistics from


Medical texts, whether they are clinical case reports, referrals, prescriptions or expert documents, are basic elements of the schemes of welfare and social justice.1 A specific genre of medical texts essential to the distribution of welfare resources is medical expert certificates. They function as social actors; they are documents that do things in modern welfare states: firstly, they release vast public costs—why society and lawmakers should be interested in them.2 Secondly, they may be decisive to individuals’ life quality, their access to healthcare and welfare. Finally, such documents contain data that are fed into the official statistics and used to elaborate future political actions.3

Issuing medical certificates is part of the physicians’ social contract. In Norway, most of them are issued by general practitioners (GPs) and legitimise governmental social costs amounting to about €15 billion.4 Despite their social impact, medical expert certificates have been largely overlooked by the medical academic critique. Direct studies of medical certificates by close reading, to our knowledge, have only been presented in our two previous articles, investigating how GPs represent the patient5 and how GPs seek to verify the connections between health, functionality and fitness for work.6

The requirement for disability benefit (DB) in Norway is that the patient has a recognised disease diagnosis that must have led to a major loss of function, which in turn must have resulted in a reduction of income level by at least 50% on a permanent basis. The patient’s work incapacity must apply to any-work. Before requesting DB, the patient must have undergone medical treatment (administered and provided by the GP) and work training administered by the National Insurance Administration (NAV) (Arbeids- og velferdsforvaltningen), all within a time frame of 3–4 years. This is in accordance with the ‘work-first’ policy of NAV, which prioritises employment to passive receiving of benefits. In principle, no application for DB will be processed if these measures are not tried out and completed. Only when the local NAV office supervisor accepts that there is real work incapacity, the patient may apply for DB, and medical documentation is requested. More than 90% of medical expert certificates are issued by GPs.

In the present article, we explore evaluative and rhetorical aspects of medical certificates in Norway. Our purpose is to direct attention toward the form and character of expert texts in medicine, a substantial part of the physician’s practice.

In Norway, nearly ten per cent of the population between the age of 18 and 67 receive permanent DB. As disease is a mandatory criterion for DB, GPs have a crucial role as experts within the Norwegian NAV. Officials within NAV report that the primary source of their decisions (in addition to the National Insurance Act) is the medical certificate written by the GPs.7 Because the final decision-makers do not meet the claimants in person, the information given in these and supplementary documents is paramount.

With a few exceptions,5 6 medical certificates have hardly been scrutinised, whereas GPs’ experiences with and attitudes towards issuing medical certificates have been paid broader attention within research. GPs have reported that they use a variety of ‘strategies’ to help their patients, or to cope with conflicting values in sickness and disability certifications.8–13 Illness is a vague concept, assessment of work is difficult and the intersection between medicine and law is not always understood.14–16

We aim at answering the following questions: (a) which value positions and attitudes do the GPs (termed ‘the textual voice’) communicate in their medical certificates; (b) how does the textual voice relate to alternative value positions and (c) looking at text as action: what does the text do to make readers feel or act in certain ways? We discuss the rhetorical and performative aspects of the expert texts and how they aim at creating meaning beyond the professional, medical description and evaluation of the patient’s health.

On a methodological level, our paper can be seen as a contribution to qualitative health research that combines the humanities and medical practice. It is an attempt to show how research may help doctors improve their task as certifiers, but that may also have an impact on medical writing in general.

Theoretical background

Document analysis takes into account that documents are significant social actors that, on one hand, are products of some reality; on the other hand, produce their own reality, working to achieve particular effects.17 This study of documents is concerned with how texts depict or construe ‘reality’ rather than whether their content is ‘true’ or ‘false’.17 18 The focus of our text analysis here is how the writer (interchangeably called the ‘textual voice’) addresses the reader. The study is inspired by the concept of dialogism usually associated with the Russian literary theorist, Michael Bakhtin,19 which claims that any utterance only has meaning in so far as it responds to a previous utterance or anticipates a response from others. As we cannot know anything about the real persons, we establish the writer as well as the reader as intrinsic voices emerging within the text. Thus, we use the concept of ‘the authorial voice’ or ‘the textual voice’, not to neutralise the historical GP as the source of the text, but to draw attention to the text as the object of analysis.

Furthermore, we use the linguistic framework of appraisal to identify evaluative language (feelings, judgements and appreciations) and dialogistic strategies in the text.20 In general, texts are different as to their dialogic style: they can be dialogically expansive in that they acknowledge and open to alternative voices and value positions, or dialogically contractive in that they curb, silence or deny alternative voices and their value positions.20 On the basis of texts as dialogic, we investigate how the authorial voice engages with external voices referred to in the text (as opposing them, aligning with them or as being undecided or neutral) and how the textual voice anticipates the value position of the reader (as whether the reader may take it for granted or find it questionable, problematic or contentious).


A qualitative document analysis of GPs medical certificates in Norway was carried out to explore a particular aspect of these expert documents: the values and attitudes promoted directly or emerging in the texts and how they are communicated to the readers. The study is explorative and interpretational, aiming at open a door to new knowledge about these documents not previously done.

Research material, sampling strategies and ethics

To achieve a typical variation of the material, we did a purposive (strategic) selection of medical certificates written by GPs.21 Four regions were selected according to the number of DB receivers: one region had a relatively high proportion of DB receivers (>12% of the population between 18 and 67), two were in the median (8%–12%) and one had a low proportion (<8%).22 The distribution of diagnosis in the strategic selection reflected the distribution of diagnosis among DB receivers on a national level: 1/3 musculoskeletal diseases, 1/3 psychiatric diseases and 1/3 all other diagnosis. The diagnoses had been coded according to the International Classification of Primary Care, second edition. The patients should be between 18 and 67 years and at least one-third belonging to one sex. The medical certificates had been written between 2007 and 2013 on a standard form.23 The patients’ applications for DB were closed by NAV, and decisions were made by the time of sampling. The certificates were collected during the first half of 2013. An initial request to access ‘their’ medical certificates was sent to 150 claimants of DB, attached a letter of informed consent. We expected to obtain consent by 1/3–1/4 of the respondents, a number of certificates considered sufficient, but not too large for an in-depth text analysis. The number who consented to the primary request was however too small and we added one more region with a median proportion of DB receivers. After one reminder was sent out by the local NAV offices (except one local office that had no capacity for the extra task) to the non-responders, the final number of certificates was 33, from 33 different GPs The Directorate of Labour and Welfare and local NAV officials performed the collection of medical certificates and anonymised the patients before they were released to the researchers. The project was approved by the Norwegian Data Protection Official for Research,24 by the Directorate of Labour and Welfare and by the Council of Secrecy and Research in the Ministry of Justice and Public Security.

Patients whom the certificates concerned gave their informed consent to publishing the results. They were anonymised to the authors and did not personally participate in the study.

Data collection and analysis

We collected and analysed the data between June 2016 and January 2017. We started with a ‘microanalysis’ of the texts,25 aiming at identifying and picking all textual elements that seemed to be relevant to concepts of attitude. Corbin and Strauss define microanalysis as ‘detailed coding around a concept. A form of open coding used to break data apart’ … enabling us to ‘think differently about things’.25 The analysed units were single words, grammatical devices, phrases and sentences that contained elements of appraisals; affect, judgements and appreciation. The sampling of these elements was guided by the linguistic framework of appraisal elaborated, among others, by Martin and White20 (table 1).

Table 1

Linguistic resources of appraisal (based on Martin and White20)

In addition, textual elements interpreted as intersubjective or dialogic actions 19 were selected. Following the initial microanalysis, the elements were further analysed, interpreted and grouped according to types of attitudes: those attributed to the patient by the writer (the textual voice), those value positions taken, implicitly or explicitly, by the writer and eventually, the dialogical or persuasive style by which he/she addresses the reader. The process from appraisal to ascribing values is shown in figure 1.

Figure 1

Diagram showing the analytical steps. (A) Detailed, descriptive and open coding focusing on textual elements ordered into three categories of attitude: feeling, judgement and appreciation. (B) The empirical categories of attitudes were further analysed, expanded and interpreted as certain values mediated to promote disability benefit (DB). (C) The values ascribed to the patient, the writer and the reader, respectively, were interpreted as values used as rhetorical means acting persuasively.

Thirty-two certificates were written on a standard form (see online supplementary appendix), and only one was written by free-hand. The certificates’ word count (excluding the form texts) varied from 112 to 452; mean word count was 264. When the GPs’ texts quoted statements from specialists (not expert certificates), they were analysed as part of the GPs’ texts.

Supplemental material

The certificate texts were repeatedly read until we found no new elements related to value/attitude and types of intersubjective actions.

In the next and final stage of analysis, we interpreted the linguistic elements of attitudes and values into broader semantic categories. Of the 33 certificates, three illustrative certificates were selected to present and expand on our findings, based on the principles of ‘intensity sampling’.26 As stated by Patton, ‘An intensity sample consists of information-rich cases that manifest the phenomenon of interest intensely (but not extremely)'. At the same time, it requires systematic knowledge about the data material as a whole in order to be able to identify the intense examples. Our three cases were selected because they contain the attitudinal and rhetorical elements that were found throughout the whole sample. They can be said to stand in a part-for-the-whole (synecdoche) relation to the material, so as to represent the whole richness of the texts in a condensed form.26


Attitudes of judgement, affect and appraisal were found throughout the whole material (see online supplementary appendix). We grouped the results into five main value themes (see online supplementary appendix: results of value analysis grouped into five main themes). All the main categories of attitudes are present in each of the five themes of value.

Themes related to attitudes

Theme 1: the DB processing within NAV deteriorates the claimant’s health.

Theme 2: the patient ‘deserves’ DB for possessing values that are ‘NAV-specific’, and for having values accepted as universally positive.

Theme 3: patient described by low social esteem; DB will make life better for the claimant.

Theme 4: granting DB argued on basis of the patient’s wish or for being a pragmatic solution.

Theme 5: persuasive action: addressing the reader through authoritative statements and negotiations.

Our text analysis of expert certificates shows that the texts mainly combine an expansive (open) dialogic style with promoting positive human values attributed to the patient or anticipated to be held by the reader. The authorial voices draw on a variety of values related to social esteem and social sanction, some of which can be inferred as ‘NAV-specific’ values, attributing attitudes to the patients that evidently cohere with the values of the politics of ‘work-first’ (seeing work as preferred to benefits), such as willingness to contribute to one’s recovery, participate in work training or education and minimise the economic burden of benefits to society. Other certifiers pointed out the patients’ universally positive human values, such as sincerity, perseverance, tenacity and sacrifice, which seemed an important part of the rhetoric. Positive universal values are also attributed to the anticipated readers, such as compassion, understanding, helpfulness and willingness to be ‘realistic’ and pragmatic in their decisions. It should be emphasised, though, that the boundary between these inferred value categories—NAV’s versus universal ones—is not absolute or concise.

Affects attributed to the patients were almost only described in relation to the DB processing, either for taking too long time, or because the patients experienced misbelief by NAV.

In the following, we expand on our interpretations of the themes listed in the online supplementary appendix; we place our findings into a textual context to illustrate the diversity of rhetorical strategies.

The excerpt (A) refers to a long-lasting rehabilitation of a patient with arthrosis, dyspepsia and spondylolisthesis. The text in bold is the requests forwarded in the NAV’s standard form, medical certificate for work disability24:

(A)  Current clinical status (specify date of examination) and results of relevant investigations:

He has paid with his own money for physiotherapy, acupuncture and several MRIs to prove that he has physical, medical problems that cause pain […]. He is motivated for his own health, but after more than 10 years of treatment without significant improvement of his health, I see no reason why he should continue with rehabilitation/courses. The patient must get disability benefit.

The GP’s answer to the medical requests seems irrelevant, but we will derive the meaning of the seemingly irrelevant information. The information that the patient has spent ‘his own money’ indicates that the patient has had extraordinary costs of healthcare. As a rule, patients’ payment for health services in Norway is limited to a maximal annual sum, after which all healthcare is free for the year. It is therefore not obvious why the textual voice emphasises that the patient has spent his own money, unless he has been using private—and more costly—commercial healthcare (which we do not know). We take this to be a connotation of personal sacrifice, a special effort to ‘prove’ or objectivise illness (vis-à-vis NAV), indicating that there has been a mistrust from NAV towards the patient’s unfitness or illness. Finally, the claimant is ‘motivated for his health’, a proposition implicitly countering a possible assumption that the patient is a malingerer. The text indicates that the patient has complied with ‘more than 10 years of treatment’, which implies that he cannot be blamed for not having recovered to be fit for work.

According to the Norwegian National Insurance Act, only disease/illness-related loss of functional and working abilities is valid as a basis of DB. Moreover, appropriate treatments and reasonable measures must have been tried out before he/she is declared eligible for DB. How are these legal demands defended in the text? Being motivated and participating properly are associated with ‘NAV-specific’ values, and the patient’s will to make sacrifices, to be honest and to behave appropriately are associated with good values in general. Likewise, the long course of treatment shows the patient’s cooperative abilities, displaying patience, perseverance and tenacity, all elements of personal character. Altogether, the texts connote the patient’s character and behaviour and show good moraleand positive social esteem. By taking this value position, the textual voice establishes an image of a reader who is anticipated to share such values. However, there are more issues at stake in the relationship with the reader. The previous certificates written for the claimant (usually one is demanded for each year of receiving preliminary benefits), we may assume, have not convinced the reader that the claimant is eligible for DB. Instead, the text mediates that he is worthy of DB. Following the ‘motivation’ statement, the textual voice entertains the following propositions: "I see no reason why he should continue with rehabilitation/courses. The patient must get DB". The subjective contingency of the statement implies: "I see it this way, but others may see it differently". The authorial voice, thus, is dialogically expansive in that it recognises the possibility of alternative viewpoints, lowering the threshold for others who might wish to oppose the statement.27 The subjective stand of the authorial voice, however, puts the alignment with the anticipated reader at stake: the arguments put in favour of the patient’s DB may be contentious and risk being questioned or rejected. However, the ‘I’ is not just any ‘I’, but a voice that positions itself as authoritative towards a lay person (the NAV official) with the trustworthiness of the speaker as a physician. Thus, at the same time, it also works as dialogically mildly contractive (increasing the threshold to opposing voices). This shows the dynamics of the textual dialogue: it appears monological, it asserts and states, but its aim is nevertheless that of addressing alternative voices, although signalling that some voices—the physician’s—have greater authority than others. If previous medical information did fail in convincing the reader(s) of the patient’s eligibility for DB, the present information does not obviously add any new medical information. On this backdrop, we find that the textual strategy is composite: first, it states that DB is a necessary outcome; second, it construes an image of a reader with whom it negotiates values, a rhetorical act construing an addressee who is reluctant to the conclusion and who needs to be persuaded and third, the textual voice acknowledges other possible views, but, in addition marks itself as the more authoritative one.

In the next text (B), we investigate how the textual voice positions itself vis-à-vis the patient and the issues in question. There are similarities between the texts (A) and (B) consisting in that both patients have been on rehabilitation for many years. In (B), however, the textual voice displays disagreement with the patient’s reported speech(in italics):

(B)  Diagnosis: myalgia

History of the disease, symptoms and treatment:

Has been on sick leave and work training for 13 years. All previous rehabilitation has failed. The patient claims to have extensive and intensive pain making him unable to return to any work. (He) asserts that he is tired and worn-out and cannot do any kind of work. He feels powerless […]. Now he wants disability benefit.

Describe how the functioning is generally reduced due to disease:

The patient himself claims to have a general physical weakness with myalgia, etc., but the clinic rather indicates psychosomatic distress and, consequently, there is no physically explained dysfunction. One has not managed to get the patient to resume work during more than ten years, and therefore disability benefit should be considered.

Other information:

The patient claims to be burned out. The patient is convinced that he cannot get back to work.

In this text, there are the authorial (textual) voice and the external voice (the patient’s), and what is at stake is the alignment of the reader’s value position with that of the authorial voice. By attributing speech to another person,20 the textual voice acknowledges another voice in another language (heteroglossia): ‘the patient claims', ‘he asserts’, ‘he feels’, ‘he wants’, ‘he says’ and ‘he is convinced’. Attributions have different meanings: ‘says’, ‘feels’, ‘wants’ and ‘is convinced’ indicate acknowledging (and perhaps showing empathy for) the external voice, and they are treated as unproblematic (one can hardly oppose that the patient feels powerless), and the attributions are therefore dialogically contractive (curbing alternative voices). Conversely, by using attribution by ‘claim’ and ‘assert’—also acknowledging the patient’s voice—the authorial voice distances itself from the patient’s statements; it takes no responsibility for their reliability. Reported speech by ‘claim’ and ‘assert’ is dialogically expansive and signals that the patient’s claims may be questioned or doubted,28 which is done immediately: ‘…but the clinic rather indicates psychosomatic distress and, consequently, there is no physically explained dysfunction’, which curbs further dialogue on the issue.

The textual voice does not argue that the patient is incapable of working. What is in play here is how this can be compatible with granting DB. The textual voice, in the words of the Russian linguist Bakhtin, ‘is attempting to arrive at the coming answer while at the same time anticipating it’.29 The anticipated answer influences the text’s rhetorical strategy: it acknowledges the patient’s sayings, but sets them aside and suggests a solution that may be more pragmatic than biomedically legitimate (‘there is no physically explained dysfunction’), although reasonable (‘one has not managed to get the patient to resume work during more than 10 years’). However, the addressee is construed as likely to find the proposal problematic. The textual voice, rather than authorising itself to take ‘command’, opens to dialogue (‘DB should be considered’) and invites the addressee to speak out and make a decision. It directs itself to a construed reader assumed to have certain beliefs and values and whose anticipated or desired answer already has influenced the text.29

Some texts directly recommending the granting of DB for an individual argue that DB—in addition to be a right, given the criteria—is also morally right. In the excerpt (C), the physician’s value position is to defend DB as a way of making life better to a woman addicted to alcohol.

(C)  Diagnosis: alcoholism, lower back pain, panic anxiety disorder

Describe how the functioning is generally reduced because of disease:

Because of alcoholism, her abilities are generally reduced.

Other information:

The patient should get a 100% disability benefit. It is possible for her to get an ‘engagement’ by the local AA and this will be very good for her. Not least to her self-esteem. It could also, to some extent, act as a deterrent to her alcoholism.

Evidently, alcoholism is seen as the main cause of the patient’s low functionality. ‘Should get a 100% DB’, is a deontic modality expressing how things should be, showing the authorial voice’s stance and intersubjective positioning: a call for action. It is a contingent, subjective assessment in favour of DB, taking a dialogically expansive position: "I think that the patient should get DB (but others may think differently)". The first two sentences under ‘Other information’ are put in an additive, neutral (not causal) relationship to one another. It is unclear whether the textual voice means that 100% DB will facilitate the patient’s engagement in the local AA, and such an inference is an option. With this as a starting point, the textual voice suggests some positive consequences of DB: increased self-esteem and deterrence of alcohol abuse. These proposals, however, are put with reservations and open to alternative voices: It is possible that she will get an "engagement"’ The quotation marks connote irony, suggesting that "engagement’ is elusive, and that the textual voice does not commit itself to the validity of the proposition.20 Furthermore, the hedging of the proposal "It could also, to some extent, act as a deterrence to her alcoholism" suggests that others may find the content doubtful or unreliable, that the authorial voice’s commitment to  the proposition is low, and the outcome uncertain. The strong appreciations in ’this will be very good for her… not least [t]o her self-esteem' is not in line with the language that vaguely speaks of ending alcohol abuse, but marks a value position which—on moral grounds—is difficult to oppose by the reader.

In the following, we expand on how the text uses rhetorical strategies to make an alignment with the construed reader. Alcoholism is defined as a disease in the National Insurance Act,30 but is socially subject to disapproval and blame. The authorial voice fleshes out how the patient’s life might possibly look like if she is granted DB, but shows little belief in its proposal. While the future health and lifestyle of a DB recipient is irrelevant to NAVs politics, other things are at stake here: the alignment with the reader on certain value positions. The text appeals to an imagined reader who is familiar with the importance of good self-esteem, of the morally good in helping others (participation in AA) and of getting an end to alcohol abuse. Because the textual voice has no strong faith or commitment to what it suggests, it needs to make assumptions about, anticipate, the reader’s feelings of responsibility for the possible outcomes: if DB is granted, there is a chance that the patient’s alcoholism may end, and qualities of her life will improve; if not, she may go on miserably. The main rhetorical strategy, altogether, is an appeal to the addressee to share values that might justify—at least morally—the granting of DB.


All the 33 medical certificates, either explicitly or implicitly, argue in favour of granting the patient DB, but on various grounds. Argumentation based on factual medical information and how it affects the patient’s functionality is not foregrounded in the texts; rather, marking attitudes, sharing values and hence also responsibilities with the reader (the decision-maker) make up the essence of the texts’ intersubjective strategy. The textual voice construes a readership that is potentially questioning the issue at stake and that may not share its value position. Accordingly, the dialogic style is marked by various techniques: firstly, the textual voices may engage with external voices, particularly with the patient’s attributed voice, acknowledging it, justifying, countering or correcting it. Secondly, by hedging utterances, they mark their dependence on the reader’s acceptance,31 and showing uncertainty and sketching probabilities attached to the writers’ statements, they allow alternative voices and value positions to come up. Thirdly, there are direct appeals to the reader to grant DB.

The values they promote are partly those that are strongly relevant to the politics of the National Insurance scheme, such as the client’s cooperative abilities, motivation and will to resume work. Other values found in our material are universally accepted, such as the patients’ positive personal character (alluding to deservedness) or the possibility of getting a better life with DB (calling for the reader’s compassion); they are, however, irrelevant to NAV, which ranks disease as the key value, followed by activation and motivation for work.

This document study represents new and direct information about an aspect of GPs acting as experts for their patients, and about the performative role of medical texts. To our knowledge, this is the first textual analysis illuminating the value position of medical expert texts written by GPs. Documents are pre-established actions,3 and their making is, of course, not influenced by the researcher (reactivity). Biased data selection is, however, not completely avoidable, but we believe that theoretical criteria guiding all our selections have reduced the risk of picking data that ‘stand out’ to the researcher.32

Our reading of the documents is done with reference to two theoretical models. It must be noted, however, that the reading is a work of interpretation (as is all reading). Unlike descriptive validity, for interpreting validity there is no in-principal access to data material that would unequivocally address threats to validity.32 Document analysis, nonetheless, is ‘making the text speak’,17 and interpretation of meaning is a substantial part of reading text. The way we interpreted the texts is one way among possible others that might prefer other theoretical models for collecting ‘our’ data or other kinds of data.

Our results indicate that the texts of medical certificates mirror the real problems experienced by GPs acting as experts.9 15 33 34 Interviews with GPs attitudes find that ‘illness’ is complex,14 35 and the patient’s impact on the GP is strong.33 36 In addition, ‘unfitness for work’ may comprise wide grey areas of ill health, and of social, personal and life problems; "working capacity" is poorly defined and disease may not even be on the list of decisive factors as considered by GPs.37 This may in part explain why clear connections between disease, functionality and work incapacity are missing in many certificates.6 It is not conspicuous that GPs’ personal attitudes and values influence what they emphasise in their certificates.8 36 38 GPs acting as experts, however, are reported not to be aware of the distinction between their personal value positions and the medical evaluations they make in clinical practice.38

There is a probability, we suggest, that advocacy and subjective value positioning replace relevant medical information. The explicit call for granting DB on behalf of the patient counters the testifying GPs’ position as impartial experts and the purposeful argumentation of such texts can hardly be seen as objective. The rhetorical strategies in these texts seem to be the balancing of the physicians’ professional integrity on one hand, and their wish to comply with the patient’s desires (recommending DB) on the other. In addition, it seems important not to jeopardise the central issues at stake by pushing out the reader (the decision-maker), setting aside her legitimate discretionary evaluation or restraining her freedom to act.31 A reasonable degree of reservation, uncertainty and discretion displayed by the expert enhances the impression of professionalism and trustworthiness, and reduces the social distance between the authorial voice and the reader.

Whereas medical expert documents are supposed to be impartial and objective, we have shown that values were ubiquitous in the texts we have studied. This is not surprising, given that values have been argued to pervade all texts.39 Descriptive texts also have a purpose,25 40 and even aiming at neutrality is in itself a value position.41 Medical expert texts, acting within the social scheme of justice, are subject to legal regulations which emphasise qualities such as relevance, accuracy and objectivity.42 Being guided by values is not in itself irrational or problematic. It is not even problematic that these values may not be immediately discerned by the reader or when they are only inferred.20 It can be a problem, however, when the rhetoric replaces professional information that may justify important decisions, such as the allocation of social goods.

Some GPs acting as experts may see their role as the patient’s advocate unproblematic and important. Sympathy for the patient is highly legitimate, and ‘putting the patient first’ is a traditional axiom which, however, may conflict with other commitments in a number of circumstances.43 Therefore, mastering expert texts is important to the authority and the trustworthiness of such documents and should be taught in medical schools.

The importance of seeing medical texts such as certificates and scientific articles as actions or strategies in themselves can hardly be overestimated.40 The present study, to our knowledge, is the first study describing the rhetoric of values and attitudes of medical certificates by detailed text analysis. Medical certificates influence what kind of decision-making is being performed. In the case of DB, it is the question of large sums of money and a fair, predictable distribution of welfare resources Medical certificates are part of the discoursal expertise of medicine. They need to cease being an ‘orphaned field’, so that future physicians can be conscious about how they write and to what effects. In a Swedish study,44 introducing new and more specific guidelines for GPs’ functional assessments in long time sick-leave showed a limited effect in the form of increased information, whereas there are no corresponding studies made in Norway. There is, however, no evidence that any single measures to be taken would essentially improve medical certificates. Writing medical certificates is an activity that is exerted within a complex field: the doctor-patient relationship, frequently vague concepts of disease and—in Norway—a strong GP autonomy. We believe that increased professional writing skill is not in opposition to any of these issues, but all the more calls for better learning in medical schools in managing the challenges of being an expert and a GP.

Comments to the analysis

The language of values is complex, can be subtle and ambiguous and must, of course, always be related to context. The present text analysis is a work of interpretation and inference based on a particular framework of evaluative language. This does not mean that an analysis of evaluative language is a straightforward work. Equal words denoting attitudes may have different meanings in different contexts, and sometimes the boundaries between categories of attitude can be debatable. This implies the possibility of alternative interpretations.


The authors would like to thank statisticians in the Directorate of Labour and Welfare and the local offices of NAV that helped select and collect the medical certificates used in this study.


View Abstract


  • Contributors GAA and AKL conceived of the project and AKL, BN and GAA elaborated the design of the work. GAA, AKL, EE and BN planned the sample method of the material. GAA collected the raw material. GAA, AKL, BN and EE participated in the collection of empirical data, data analysis and data interpretation. GAA elaborated the draft and GAA, AKL, BN and EE critically analysed and repeatedly revised it. All the authors had full access to the data and all finally approved of the present version to be published. GAA, AKL, BN and EE agree to be accountable for the accuracy and integrity of all parts of the work.

  • Funding The study was funded by the Norwegian Research Fund for General Practice (AMFF) administered by the Norwegian Medical Association

  • Disclaimer The funding body has not taken part in any stage of the study or in writing the manuscript.

  • Competing interests None declared.

  • Patient consent Not required.

  • Ethics approval All the patients (anonymised and deidentified to the authors) gave their written consent to using their medical certificates in research and publishing. The project was approved by the Norwegian Data Protection Official for Research, by the Directorate of Labor and Welfare (NAV) and by the Council of Secrecy and Research in the Ministry of Justice and Public Security.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement There are no available data.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.