Dealing with patient expectations regarding the prescription of antibiotics in ambulatory care in Germany: A qualitative analysis

Background Antibiotics prescription rates are relatively high in primary care in Germany. Patients’ expectations have frequently been mentioned as reason for high prescription rates. The extent to which patients’ expectations play a role and the strategies that physicians use to deal with these expectations and prevent non-indicated prescriptions in acute, uncomplicated infections are the subject of this paper. Methods In this qualitative study, twenty-seven semi-structured interviews with physicians were conducted in 2018. Data were audiotaped, pseudonymized and transcribed verbatim. The analysis was based on a framework analysis according to the Tailored Implementation in Chronic Diseases (TICD) checklist, and was expanded for deeper analysis and completed with the Behavior Change Techniques (BCT) taxonomy. Results The data revealed that patients’ expectations regarding the prescription of antibiotics can play an important role, although they may be ignored by physicians. If physicians respond to patients’ expectations, they use various strategies to deal and communicate with patients. Successful communication is more often achieved, if the physician takes sufficient time with the patient and if this results in a cooperative conversation. Thus, patients get informed about antibiotics and antimicrobial resistance, and it is possible to discuss alternative treatments. Seven strategies from the BCT taxonomy were mentioned by physicians in this study. Conclusion Constructive physician-patient communication was key to reducing antibiotic prescribing. limited the use the focus of implementation strategies. Trial

The prevention of antibiotic resistance has been assigned high priority on both national and international political agendas. Therefore, the World Health Organization (WHO) global action plan on antibiotic resistance was adopted in 2015. One of its aims is to improve understanding and awareness of the issue in order to reduce antibiotic prescription rates [1]. Germany developed the German Antibiotics Resistance Strategy 2020 (DART), which also pursues the goals of counteracting the development of antibiotic resistance, suggesting a comprehensive set of implementing strategies to address the issue [2]. In relation to the total population in Germany, the consumption of antibiotics (AB) is highest in ambulatory medical care, where 85% of the total amount of antibiotics consumed within the human medicine sector is administered [3]. The inadequate use of antibiotics is considered to be the main cause for the development of anti-microbial resistance [4]. In primary care in Germany, antibiotics are prescribed by primary care physicians in 41 % of patients of lower respiratory tract infections and only 52 % of these prescriptions comply with recommended guidelines [5].
The research project "Sustainable Reduction of antibiotic-induced antimicrobial resistance" (ARena) applied and evaluated a multifaceted implementation program, with the aim of achieving a sustainable reduction in antibiotic-induced antimicrobial resistance [6]. The effects were evaluated in a three-armed (non-blinded) cluster randomized trial, with an attached process evaluation. The latter involves interviews and surveys among the participating physicians, non-physician health professionals on the practice teams, and various stakeholders. Within the scope of the process evaluation, semi-structured interviews were conducted. One of the areas of exploration was the impact of patient expectations on antibiotics prescribing, which is the subject of this paper. More detailed information about the study can be found in the published study protocol [6].
Many patients with acute, uncomplicated infections expect their primary care physicians to prescribe antibiotics, albeit fewer than physicians may assume [7][8][9][10][11]. Patient-centered communication can help to address patient expectations and reduce inappropriate antibiotics prescribing. However, there is little research evidence on how physicians deal with patient expectations in their routine healthcare practice. For several decades, "patient-centeredness" and "patient-centered medicine" have been among the most discussed subjects in health care [12]. However, its implementation in routine healthcare practice has remained difficult. At the core of "patient-centeredness" is the communication between physicians and patients [13]. However, it can be challenging for physicians to act patient-centered, but refrain from prescribing antibiotics as desired by a number of patients [7][8][9][10][11]. This challenge is particularly large in children [14,15]. A study showed that in none of 42 patients with acute cough, physicians actually explored the expectations of their patients. This may imply that prescription frequency might be reduced if physicians knew about the needs and preferences underlying expectations of their patients. In addition, the study found that pressure exerted to obtain antibiotics is often overestimated [16]. This is confirmed by other studies, which showed that the interpretation of patient expectations is difficult and leaves room for uncertainty [17,18]. A Cochrane Review examined 10 studies on shared decision making (SDM) in antibiotic prescriptions for acute respiratory infections and found that it reduced the prescription of antibiotics by 40 % [19]. In 1997, Macfarlane et al. requested alternative strategies to support general practitioners in their decision to prescribe antibiotics. In this study, patients who demanded antibiotics received it three times as often as those who did not [20]. Overall, these studies support the assumption that communication with patients is crucial for reduction of antibiotics prescribing.
Currently, there is little research evidence on how attending physicians in routine practice deal with patients' expectations regarding antibiotic therapy in order to prevent unnecessary prescribing in Germany. The aim of this study was to explore the role of patient expectations with regards to antibiotic prescriptions for acute, non-complicated infections and what strategies primary care physicians use to deal with these and prevent unnecessary antibiotics prescribing.

Study design
The ARena study was planned for 30 months and carried out in cooperation with 14 primary care networks in two German federal states (Bavaria (n = 12) and North Rhine-Westphalia (n = 2)) and 193 of their member practices. These networks are regional associations of physicians of different specializations, which can be registered by the National Association of Statutory Health Insurance Physicians according to § 87b Abs. 1 Nr 2 4 SGB V since 2013 [21]. Patients diagnosed with acute, non-complicated respiratory infections, non-complicated cystitis, or community-acquired pneumonia could be included in the study. They had to be insured with AOK Bavaria or AOK North Rhine-Westphalia and participate in a specific healthcare delivery program ( § 140a SGB V a.F. and § 140a Abs. 1. S. 2 Alt. 1 SGB V n.F.), jointly offered by AOK and the primary care networks [6].
A qualitative study, based on semi-structured interviews with physicians, was conducted. Qualitative research is essential for the study of clinical communication because it facilitates an exploration of complex and unknown domains [22]. The COnsolidated criteria for REporting Qualitative research (COREQ) checklist (Additional file 1) is a tool to support reporting on qualitative studies and therefore was applied in this study.

Study population
Between March and May 2018, study participants were purposively recruited through the ARena project leader (aQua Institute in Goettingen, Germany) and the research team at the Department of General Practice and Health Services Research, University Hospital Heidelberg. After potential interviewees sent back a signed letter of intent to participate in an interview, they received further material, and the study team contacted them by telephone. All interview partners were participants of the ARena study and members of one of the 14 participating primary care networks in Germany. The interviews were conducted with general practitioners, pediatricians, internists, and ear, nose and throat specialists. The participants were aware that the interviewers were members of the research team and responsible for the process evaluation. The research team had no contact with the participants before the study commenced.

Interview guide
The interview guide contained various topics: dealing with antibiotics, effects of the quality improvement program (ARena) on patient care and primary care networks. In this study, the focus was placed on the part of the interview guideline "Dealing with antibiotics". The guideline was developed and pilot tested by the research team. Furthermore, before the interviews were conducted, all interview partners were required to complete a survey that asked for their sociodemographic data and information about their practices.

Data collection
Data were collected using individual telephone interviews with a semi-structured interview guide. The interviews were conducted from April until June 2018 by researchers of the interprofessional research team who had all conducted interviews in previous studies. All scheduled interviews were carried out; furthermore, no interview was repeated, and no participants dropped out. Participants were interviewed via telephone at their practices during or after their consulting hours. No other people were present during the interviews. Interviewers took additional notes during the interviews. All interviews were audiotaped, pseudonymized and transcribed verbatim with the transcription software f4. Transcripts were not returned to participants for comment or correction. However, participants will be informed about the findings of this study through interim and final reports of the ARena study.
Finally, the research team concluded that data saturation was reached and no further participant recruitment was necessary.

Data analysis
Determinants of the implementation of recommended use of antibiotics were qualitatively analyzed using the TICD (Tailored Implementation in Chronic Diseases) checklist by Flottorp et al. [23] as a framework to identify determinants of practice regarding the relevance of patient expectations for the prescribing behavior of physicians. The TICD checklist consists of seven domains and 57 determinants. To ensure that all collected data were reflected by the analysis, the framework was complemented with new, inductive categories.
During the analysis, it became apparent that physicians used different behavioral strategies when dealing with patient expectations. Since the TICD framework is not specifically designed for analysis of these strategies, the Behavior Change Technique taxonomy [24] was used as a supplementary framework. It lists 93 techniques in 16 categories. The inductively categorized potential behavioral strategies were carefully compared with the BCT definitions of the techniques to determine whether they corresponded with BCT techniques. Figure 1 illustrates the final category system for the analysis focusing on communication strategies in connection with the identified BCT techniques (Additional file 2, Figure S1 provides the complete category system). The method of analysis as well as findings were regularly discussed in the study team. For this qualitative study, the focus of the data analysis was on the results of the inductively formed categories as well as the BCT strategies, not the results of the  The results of the study indicate that most physicians were rarely asked directly by their patients for antibiotics. A few physicians were frequently asked for an antibiotic prescription. A predominantly perceived reason for this demand was that patients would want to recover quickly and think that antibiotics could help them best. On closer examination, the data shows that physicians realized that their patients do not all demand antibiotics from them and that they may wrongly suspect this.

"'Oh, thank God, I don't really want to.' This is what the majority say, whereas before, you had the feeling that a certain attitude of expectation was [...] there." (PH26)
When asked whether there are specific groups of patients with an increased demand for antibiotics, most of the participants felt that this occurred mainly among young people who were professionally committed, experienced stress and wanted to return to work as soon as possible. In addition, demands were reported to be more common in patients from countries where antibiotics are sold Some participants mentioned certain cases where they met patients requests for an antibiotics prescription even if they considered it non-indicated. This was explained with the assumption that patients otherwise would consult another physician who would prescribe antibiotics anyway. To prevent this, they rather prescribed it themselves. They also reported about patients who explicitly do not want antibiotics and want to be treated with alternative medications. Some physicians also met this type of patient expectation.
The described physicians differed from the group of physicians, who explained in the interviews that patient expectations had no influence on the prescription of antibiotics. If they considered an antibiotic to be unwarranted, then they would not prescribe it and accepted the risk of losing the patient to another physician over this.
Most physicians explained that they took patient expectations seriously and into consideration. There were different ways in which to proceed with those expectations. One of these approaches was based on the principle of delayed prescribing (DP), which almost all respondents applied in their practice.
Delayed prescribing means that the physician issues an antibiotics prescription to the patient to take home with the instruction to use it only in case of a worsening state of health [15]. The most commonly reported reason for applying this strategy was an exceptional situation; for example, shortly before the weekend or when the patient wants to go on vacation. Further reasons given were diagnostic uncertainty and that the prescription provided the patient with a sense of security. An argument against the use of DP was the shift of the treatment decision to the patient, when it is actually the physician's job.

Communication
The results demonstrate that communication with the patient was the most important approach for addressing patient expectations in the decision-making process concerning prescription of antibiotics, as almost all strategies used were related to physician-patient communication. This is illustrated by the following statement.  (Table S1).

Taking time and taking patients seriously
To be able to discuss with patients their expectations and the adequacy of an antibiotic prescription, there were communication components the attending physicians had to meet. The majority of the physicians agreed that they had to take sufficient time to talk and listen to patients and to take them seriously.
"It is very important that the patient is also heard; i.e., not immediately blocking [...] but you also have to understand the patient, perceive his worries, perceive his complaints." (PH01) Taking time for communication and taking patients seriously also meant to gain a deeper understanding of the individual patient and to question behaviour. This approach reflected the BCT strategy "re-attribution". Some physicians used this strategy when they felt uncertain about reasons for a patient's desire for an antibiotic treatment other than the feeling of being ill. Potential reasons stated were that patients wanted to limit being absent from work to a short time or avoid that at all, because they absolutely had to go to work and therefore wanted to recover quickly.

"[…] maybe I have to take a look -is there a bit more behind it or is it a repression of the patient? You
have to see that a little bit." (PH10)

Trusting conversation
If the patients felt accepted and met with understanding, they were more willing to listen to and be informed by physicians. If both aspects had been fulfilled, then a participative discussion, which the participants described as essential, could develop.
Further in-depth strategies of this approach are described by the BCT strategies "pros and cons", in which advantages and disadvantages of patient behavior are discussed, and "comparative imagining of future outcomes", in which it is discussed how the current behavior (here: expectation of an antibiotic) can affect the future of the patient.
Such a discussion could arise if, for example, the physician conveyed information to the patient necessary to weigh up the pros and cons of an antibiotic prescription.
"He may have a certain lack of information; I must provide him with the information to enable him to make a decision here, and this is usually the case when the patient understands that an antibiotic has more disadvantages than advantages." (PH04) The BCT strategy 'Comparative imagining of future outcomes' aims to clarify how patients' current behavior will affect the future if not changed. A physician illustrated this strategy by talking about efforts of trying to show patients that it was not important to be fit again the next day, but to be healthy in the future, when they would be older and, therefore, an antibiotic should not always be used, but rather, for example, rest should be preferred to medication.

Address the topic antibiotics openly
An open discussion could also result from the strategy of asking the patient about his expectations and thus openly addressing the subject of antibiotics. The physicians reported that through this strategy, they confronted patients with the topic, and it did not remain unspoken. In addition, they sometimes observed whether the patient actually wanted an antibiotic or whether they had the wrong impression of the patient.
A suitable BCT strategy follows the approach of not only openly addressing the topic of antibiotics, but also making patients feel a little insecure, so that they realize that the expectation of an antibiotic has no discernible reason.
The BCT 'Incompatible beliefs' is a technique for dealing with a patient's demand for antibiotics. It demonstrates that a demand is unwarranted and in doing so, a slight discomfort is created. One of the physicians used this technique by asking patients which exact antibiotic they thought they needed if they were certain about the prescription. Patients then felt a little discomfort as they realized they could not answer the question and that their current behavior made little sense and probably did not reflect their actual self-image.

"[The] best line is always 'I need an antibiotic', and I ask back, 'Which one?' (laughs). Then the answer is, 'You must know that', and then I say, 'I know that you don't need one, I can't think of one
that would help with your illness; you'll have to tell me, so I can write one down for you, because there isn't actually any.' [...] and then, 'yeah uhh' and then the thing is usually over." (PH17) Alternative treatment A further strategy to dissuade patients from their demand for antibiotics was to propose alternative treatments. If these did not cure the patient, then they could come back for a re-assessment and a new consideration regarding a potentially necessary antibiotic treatment.

Patient education
The most frequently cited strategy was patient education. All interviewed physicians made it clear that it was of fundamental importance to educate patients and explain why an antibiotic prescription was not indicated. If this information was provided in detail, then the probability of patients agreeing not to take an antibiotic was high. Using guidelines and theories as credible sources, physicians explained to patients that their diagnosis and treatment were correct so that patients did not feel they had been treated insufficiently in cases where no antibiotics were prescribed. In addition, physicians explained they demonstrated to patients that an antibiotic would cause more harm than good in their case. Pointing out the effects of non-indicated antibiotic prescriptions on the environment was identified as a further strategy used by the physicians.

"And when I have the feeling [...] that I don't want to prescribe an antibiotic at this point [...] I prepare a bit of the theory, so that he doesn't have the feeling that he is being treated wrongly, but that this
is correct according to probability." (PH16) "I also explain to the patient that an antibiotic given in the wrong place changes the microbiome in the intestine in any case and that this can then also have a negative influence on the immune system, on the intestinal flora. So, to describe it amateurishly, an antibiotic given in the wrong place can also cause damage to the body; I have always said so." (PH14).

Discussion
Through the analysis of the interviews with the TICD and BCT frameworks, determinants and strategies could be identified that ascertain how physicians deal with patient expectations with a focus on identifying behavioral strategies of physicians. This could be achieved by extending the TICD framework and involving the BCT framework.
The study showed that primary care physicians use different strategies to deal with patient expectations regarding antibiotics prescriptions for acute, non-complicated infections. Communication with patients was the most important approach for dealing with patient expectations. This category and several related subcategories were most prominent as it was the one most frequently addressed by respondents on the one hand, and on the other hand several important subcategories could be derived from it. Through a conversation in which the patient feels taken seriously and for which the physician takes time, patient education can be successful. The interviewed physicians' further BCT strategies for patient education were as follows: refer to scientific sources such as scientific guidelines, present information about health consequences or provide information about social and environmental consequences. They also used the strategy of slightly unsettling patients (cf. Incompatible belief) by critically questioning their demand for antibiotics. A total of 7 BCT strategies have been identified, which were already used by physicians: re-attribution, pros and cons, comparative imagining of future outcomes, information about health consequences, information about social and environmental consequences and credible source.
Interpretation and relevant literature When asked whether their patients demand antibiotics directly or indirectly from them, many physicians replied they were aware that patients did not actually expect antibiotics and they often misinterpreted. Such a misinterpretation of the demand for antibiotics has also been found in other studies 25, 26. In a German study by Altiner et al. 27, primary care physicians received communication training, and it became apparent that they interpreted expectations too early. An important reason to prescribe an antibiotic, although it is not indicated, was the fear that patients otherwise would switch to another physician. This was also mentioned as a reason for prescription in a study that dealt with physicians views of interventions to promote prudent antibiotics use 7, 28. Physicians used DP to ensure possible antibiotic treatment in exceptional situations, for example before the weekend or holidays. On the other hand, they use it to meet diagnostic uncertainty and patient dissatisfaction. A Cochrane Review of 2017 29 suggests that DP can be an appropriate strategy to reduce antibiotic consumption while not harming the physician-patient relationship. Nearly as many patients were satisfied with DP as they were with a direct antibiotic prescription and there was still a lower consumption of antibiotics. Hoye et al. explain that the use of DP as a strategy is only successful if patients are motivated for SDM and understand the effect of antibiotics 30. Other interviewed physicians refer to the strategy of "waiting and make a new appointment with the patient". This is only possible if there is a continuous relationship between the physician and the patient. If this is not the case, as in on-call duty, then physicians tend to prescribe antibiotics more frequently, which was also confirmed by O'Connor in 2018 31.
It has been shown that national and international studies from the field of "reducing antibiotic prescription for acute, uncomplicated infections" demand research on better physician-patient communication and strategies for better patient education 26, 31-39. A qualitative study conducted in nine European countries with 121 patients revealed that most are aware of the antibiotic resistance problem but cannot explain it in detail 35. The fact that patients are poorly informed and do not understand how an antibiotic works is also demonstrated by a study in Germany, in which more than one third of the interviewed participants answered "yes" to the question of whether an antibiotic works against the common flu 26. The communication misunderstandings between physicians and the parents of sick children are particularly strong. In an observational study from England, primary care consultations were recorded and analyzed. They discovered that parents believe that an antibiotic is only used for severe diseases, and they do not understand that while a viral infection can also be serious, it does not benefit from an antibiotic 33. An earlier study demonstrated similar results 25.
Parents want to talk to their physician about the severity of their children's illness, justify their visit and feel taken seriously by them. However, this should not be interpreted as a demand for antibiotics.
The present study demonstrates that the interviewed physicians partly share similar experiences and applied strategies, since many of them adopted strategies involving DP or taking the patient More importance should be placed on strategies that have been used only occasionally, such as unsettling the patient if he unnecessarily insists on an antibiotic, or on educative approaches physicians might use in individual situations. Furthermore, it would be possible to have a more detailed look at the BCT taxonomy. There are further strategies for behavioral change that could be used by physicians to make patients aware that antibiotic prescription is not necessary and to emphasize the risk of antibiotic resistance. This includes the BCT category "salience of consequences", which in this case clarifies the danger of resistances to the patients. In the strategy "behavioral experiments", physicians should give evidence-based advice instead of prescribing an antibiotic directly. They should examine how their patients react to it and whether they accept this approach well or do not understand it. The strategy "exposure" aims at confronting the patient with a feared stimulus, in this case the danger that the patient himself could become immune to various antibiotics. A final and simple BCT strategy is "identification of self as a role model", in which patients are educated about the topic in a way that they could act as role models other potential patients and for their social environment and know and share the correct use of antibiotics 24.
Strengths and limitations of the study One strength of the present study is that it was conducted as part of a large pragmatic randomized trial. 193 practices from 14 primary care networks in Germany participated in this study, 27 physicians of those practices were interviewed. Data analysis was conducted with the help of the established TICD framework and extended by the BCT taxonomy. The use of framework analysis enables the comparability of results from qualitative studies and increases the quality of the analysis 44-47. Data analysis and interpretation was discussed with the ARena research team and thus different opinions and perspectives could be included.
The analysis of the data with the help of the TICD framework formed a suitable framework. However, it did not specify all relevant issues so we added categories on the basis of inductive analysis of the interviews. It is possible that a selection bias is present because the participants of the interviews were not randomly selected; instead, interested physicians participated voluntarily. Potentially, they were physicians to whom the topic of antibiotic resistance was important already, and who generally had low prescription rates. In this study sample, no differences could be found between the various specialist groups involved or due to membership in the respective primary care network. This was to be assumed, since specialists are often visited only at an advanced stage of a disease, whereas