A heterogeneous sample of 12 patients and 14 care providers was selected. The sample selection was based on non-probability judgment sampling, assuring both relevance to the subject and diversity of the members selected. The sample size was determined on the basis of the principle of saturation and maximum variety sampling 31 .

Saturation was only a guiding principle. This concept can provide a sample size only contingently (i.e., contingent on data collection and analysis, as per Glaser and Strauss³¹). Therefore, we relied on the professional experience and competency of the research team to determine sample size. With the expertise of the social science researchers, saturation occurs within the first 10 interviews depending on the subject studied and the heterogeneity of the group (about 5 to 15). Moreover, the main data codes were developed after about 5 or 6 interviews (in-depth interviews). We also relied on the expertise of the physician members of the research team to determine the criteria of diversification and, thus, the sample size.

Another guiding principle of our sampling procedure was to explore diversity, and we sought maximum variety in sampling as well. For care providers, we selected a group of 14 participants, with experience in acute or chronic pain management at various points of the therapeutic course, from 2 different categories: physicians (9) and physical therapists (5). We aimed for diversification of points of views and care practices by including a variety of professionals from 7 medical specialties and health care professions (4 general practitioners [GPs], 1 rheumatologist, 1 physical medicine and rehabilitation physician, 1 geriatrician, 2 orthopedic surgeons, 5 physical therapists) and from both private and public practice.

To be eligible for the study, patients had to be in physical therapy programs and experience pain during at least one session. We aimed for diversification of the patient sample according to the following criterion: age (< 40 years old, n = 4; between 40 and 60, n = 4; > 60, n = 4), gender (7 women, 5 men), and place where the physical therapy program was performed (public practice, n = 6; private practice, n = 4; both, n = 2).

The patients were selected from the files of care providers identifying patients who had experienced pain during a physical therapy program. The care providers were recruited from the files of the members of the study board (pain specialists, GPs, physical medicine and rehabilitation physicians, rheumatologists, sociologists, anthropologists).

After interviewing experts in the field of pain management and physical therapy and considering patient perspectives of pain induced by care management, we formulated a semi-structured interview protocol with open-ended questions. All interview protocols were as similar as possible so as to observe the range of variation of discourses and practices between medical experts and patients and to vary data sources to allow a deeper understanding of representations and practices linked to PIEM.

The interview protocols were structured by combining a "funnel-shaped" structure and the "itinerary method" 33 34 35 36 . The funnel-shaped structure was adopted to ensure that the interviews allowed for an inductive comprehension of the social reality at stake underlying the PIEM situation. The itinerary method of data collection was derived from anthropological data collection techniques and focused on objects, practices and the decision-making process. Applied to a therapeutic situation, it allows the researcher to follow the course of the patients, from the appearance of the pathologic condition, sometimes long before the physical therapy sessions, to the time of the interview, thus placing the PIEM in a broader context than the medical one. The postulate underlying this framework is that studying patient expectations of PIEM management cannot be limited to collecting explicit expectations that they might explicitly express: their expectations have to be identified throughout an analysis of the global social situation in which the PIEM occurred and was (or was not) managed, identifying contradictions, ambivalence, implicit expectations, and unanswered needs.

The interview protocols were planned as a loose list of themes and questions, interviewers continually adjusting their questions to the specific leads of the interview and pursuing unpredictable emergent data. They combined a thematic structure with chronological sequences to detail the therapeutic journey and the course of consultation. We tried to keep a similar path for all interviewees for comparison. Because these topic guides include open-ended questions, they could be presented to all interviewees, even if they relied on different individual experiences and decision processes. Of course, some chronological sequences could not be probed to the same extent with all interviewees. Example situations were for the patient journey sequence before the first physical therapy session, because physical therapists have limited insight into this part of the patient journey, and physical therapy sessions themselves, because physicians did not attend them. However, even if some interviewees had no experience with some topics, they usually had views, beliefs, or expectations about the topics, and we collected these data. To gain insights into the topic under study, we were careful to combine different types of data and collect not just opinions, beliefs, and expectations but also descriptions of effective personal practices and the contexts in which they took place.

The patient interview guide was designed to collect data on the following:

- The therapeutic journey, from the initial health problem to the physical therapy prescription and its organization. This gave the interviewers the opportunity to gather views and beliefs on pain and to identify and characterize previous pain experiences.

- Physical therapy sessions including exercise and mobilization; their description; PIEM appearance and patients' reactions and ways of expressing this pain; effects on the patient and their interpretation of this pain; and adjustments made in response to pain. This section allowed us to probe the patient-care provider relationship and communication and their views on physical therapy.

- Patient expectations related to the management of PIEM, particularly the assessment of PIEM, its prevention and treatment.

The physician interview guide was designed to collect information on the following:

- The physical therapy prescription process, and knowledge and views on physical therapy, specifically mobilization.

- Pain induced by the exercise and mobilization prescribed. Physicians were asked to express their views on PIEM, allowing us to probe their definition of this phenomenon and the process of interpretation they relied on, including the assessment process, if one existed; the effects on the patient and care providers; how physicians were engaged in the management of this pain; prevention and treatment practices; physician views about the treatment options available; and the description of the patient-physician relationship and the physician-physical therapist relationship and process of communication.

- Physician expectations and views of patient expectations.

The physical therapist interview guide was designed to study the following:

- The physical therapy session, in particular how prescriptions are filled and what indications are given by physicians; the decision process that governs the choice of physical therapy techniques; therapists' definitions of "mobilization techniques," how these techniques are used during the session, for what purposes, and how they are performed; the patient-therapist relationship, topics discussed, information given, and patient behaviours and demands.

- PIEM and more specifically the therapist definition of PIEM and comparison with other types of pain; PIEM identification and how therapists decipher patient behaviors; PIEM assessment; PIEM effects on treatment, on patients' life, on disease; the adjustments required and the therapists' practice to relieve and prevent PIEM; and views on treatment options available. This section allowed us to probe the patient-therapist and physician-therapist relationships and procedures of communication.

- Physical therapist expectations and views of patient expectations.

The mean duration of these interviews was 1½ hours for patients and 1 hour for care providers.

Patients considered that PIEM during physical therapy programs is only one facet of a more general problem of pain induced by care (medical examination, administered personal care, treatments or investigations). They differentiated pain induced intentionally by care providers for diagnosis and that induced by care:

"The physician examined me from all angles. He pressed hard where it hurt a lot.... I don't blame him, it's his job." (Female patient, 24 years old, Houilles, received physiotherapy in secondary care hospital)

"One night, nurses came because I wet my bed. They tossed me just like a pancake to change the sheets. They hurt me so much that I screamed." (Female patient, 68 years old, Paris suburb, hip replacement, physiotherapy during hospital stay)

The description of PIEM was polymorphic and experienced as a degradation of HRQoL. The global therapeutic context affects this experience. PIEM could be sensed as "an extra pain" or as "too much pain" depending on the context, the impact of the patients' health condition in their everyday life, and their views on physical therapy. These factors may amplify or minimize the perception of PIEM. The occurrence of pain during a therapeutic program was cognitive dissonance for patients and a paradox, opposing the hope of getting better and the feeling of pain.

However, the objective of the care-induced pain was a pivotal element in how patients sensed pain and experienced it. It determined the interpretation, significance, and acceptance of the situation. Therefore, PIEM can be considered to be more a social construction than a simple painful sensation or a physiological response. It is indeed a combination of different components: sensorial, emotional, and social. The social dimension involves the meaning given to pain and the contexts of interaction in which pain is expressed.

The significance that patients attributed to PIEM was ambivalent; PIEM could be both positive and negative for them. Among positive representations of PIEM, patients cited interpretations such as "part of a scarring or healing process", "a necessity in order to achieve recovery", "a useful alert to reconsider diagnosis or treatment", or a protective signal, an indication of the "limits" beyond which one must not go. PIEM was not always seen as constructive. Although some individuals interpreted it positively, it still remained an undesired experience:

"I don't think it is a punishment, and I don't think it's a way to transcend oneself. Pain is pain! (...) I don't need pain to get mature or to grow internally.... If it can be avoided, so much the better." (Male patient, 25 years old, Paris, hemarthrosis, physiotherapy during hospital stay)

PIEM could also be interpreted negatively, as care providers' lack of technical competency or lack of attention and thoughtfulness:

"The day after my surgery, a physical therapist came. She was quite inhuman. I have a terrible memory of her. I had surgery on my right side and she grabbed my legs on the left side. Of course, I screamed!" (Female patient, 68 years old, Paris suburb, hip replacement, physiotherapy during hospital stay)

PIEM was referred to as control: when interpreted positively, it was pain "under control," and when viewed negatively, it was an unsought experience, associated with fear and helplessness:

"This pain (PIEM) is mostly coupled with a feeling of helplessness. It's frustrating because you don't know what ultimately comes from the pain and what comes from your feeling diminished." (Female patient, 26 years old, Châteauroux, surgery for hip fracture, physiotherapy at an outpatient clinic)

The impact of PIEM was complex, ranging from denial of symptoms to disruption of the therapeutic contract. When confronted with PIEM, patients did not systematically express their pain. Showing and verbalizing pain was not a neutral decision, and it could be viewed as a social and strategic decision. Strategies for emphasizing, softening, minimizing, or "silencing" were observed, depending on the sense given to the physical therapy, the pain felt, and the social context of occurrence:

"Maybe I should have accepted more [than I did], or perhaps I accepted too much. (Is it difficult to judge?) Yes. It's like for syrups: before they were bitter but (to get better you had to take them anyway. Physical therapy is the same! It's not because it hurts you that you have to stop. Because that's what makes you better, so I do accept a little bit. But when it hurts too much, I stop.... (Male patient, 36 years old, Ocquerre, patella fracture, physiotherapy at an outpatient clinic)

PIEM expression appeared to be a real issue for patients who tried to control it. Shouting might be felt as a loss of control, an inadequate verbalization, or a worthless option:

"I wasn't screaming. A kid might have screamed reacting to this kind of pain but not an adult: I just gritted my teeth." (Male patient, 25 years old)

"There are people who scream right away, but you shouldn't -- otherwise you can't progress.... At physiotherapy, they told me that I no longer feel the pain.... I do feel it, but I never say anything because I know that it [the physical therapy] must be done.." (Female patient, 61 years old, Boulogne-Billancourt, trauma during car crash, physiotherapy during hospital stay and at an outpatient clinic)

Expressing one's pain could be felt as socially incongruous, and patients adjusted their expression according to how they defined the PIEM and how they interpreted their social situation:

"Last week, I had a crying fit at the hospital. I couldn't stop crying. That happens regularly. And yet, I wasn't off worse than before, and the worst thing is that I would hide the crying because I was in the physiotherapy room and I was afraid of being watched." (Female patient, 61 years old, Boulogne-Billancourt, trauma during car crash, physiotherapy during hospital stay and at an outpatient clinic)

The "presentation" of PIEM also depended on bystanders. For instance, one patient was allowed to express her pain by singing, because shouting was not an option for her in a care context; others limited themselves to non-verbal indicators:

"During the [physical therapy] session it hurt. I was sweating [with pain]." (Male patient, 77 years old, Villiers-le-Bâcle knee osteoarthtritis and low back pain, physiotherapy during hospital stay)

Interactions with the health care providers affected patients' response to PIEM, sometimes encouraging patients to bear the pain and the physical therapy at the same time:

"They [physical therapists] are so nice. They're working for us but they are so involved that you also want to do it for them." (Female patient, 68 years old, Meudon, hip replacement, physiotherapy during hospital stay)

In some aspects, the acceptance or minimization of the expression of PIEM may be part of a specific exchange between patients and physical therapists, denoting reciprocity in the care process. The information delivered to patients was also an active part of this relationship and could have an influence:

"I remember a physical therapist who told me: "we're not going to relieve you but we are going to teach you how to live with your pain". It made me think..... If they hadn't told me at the rehabilitation center to do things but to do them reasonably, I think I would have stopped immediately, saying that this pain was abnormal." (Female patient, 43 years old, disk herniation, physiotherapy during hospital stay and at an outpatient clinic)

Nevertheless, the occurrence of pain during a physical therapy session had negative consequences for the patient and the therapeutic contract. Patients expressed feelings of loss of confidence with the medical management of the situation and fear of reactivation of the disease:

"For a while it did not hurt anymore. I began to recover. But the physical therapy, the manipulations, it woke up the pain. What bothered me most was that it (my ankle) had begun not to hurt, but the physiotherapy sessions brought back the pain. That's all I noticed." (Male patient, 46 years old, ankle strain, physiotherapy at an outpatient clinic)

Care providers agreed that PIEM is generally not a part of their management strategies and that a generally accepted definition is lacking. When analysing the meanings of PIEM, care providers emphasized the positive dimensions of PIEM, such as pain as a "red flag" allowing for treatment adaptations or as a signal of evolution of the underlying pathology:

"Pain is a limiting factor; it is an alarm system that tells you not to go further. For instance, if you consider hand surgery, it is useless to prescribe painkillers to relieve the pain, because there are things you mustn't do (and pain will stop you doing them)." (GP, private practice)

"(I view PIEM differently according to the situation). It is not the same the week after the surgery as it is six weeks afterwards, when normally the pain is decreasing [and] the patient regains mobility. If the pain or the level is reversed at this time, it means that there is a problem, a complication. It can be a hematoma, an infection, or dislocation of the orthopedic material. (Orthopedic surgeon, public practice)

These "virtues" seemed to have more weight than the negative dimensions of the situation. Care providers' downplaying the negative dimensions of PIEM involved a double process: the purpose of PIEM (patients suffer to get better) and the reminder of medical deontology and the absence of intentionality ("we do not do it on purpose"):

"When you examine a patient, you may cause pain. There is no malice intended. It is because it helps us make a diagnosis. It's the same for physical therapy: we might actually have to cause pain. We are not trying to hurt the patient; it's to make them better."(Rheumatologist, public and private practice)

"In my mind, pain is impossible in physical therapy. It may hurt at the time but afterwards, it is so much better that it is a 'blessing in disguise'. That's why it does not even occur to me to anticipate [the pain]. I am probably wrong, but that's the way I think automatically." (GP, private practice)

PIEM was identified mainly by physical therapists and less by physicians. Different modalities of pain expression were reported: clinical signs, attitudes, and verbalization. The identification of PIEM did not imply management of PIEM; the patient and physical therapist discussed the level of pain and what was acceptable to the patient before treatment might be modified.

Although several care providers tended to minimize the consequences of PIEM, consequences cited were a worsening of the patient's clinical situation, fears about physical therapy, rejection of the physical therapist, and refusal of care:

"I think if it's too painful, it will have psychological consequences in following up the physical therapy program. Patients will be afraid to come. They will look on physical therapy as something horrible instead of thinking it will improve (their condition)." (GP, private practice)

"If you exceed the limit once, the patient may no longer want to see you. It's rare but it happens. If you simply exceed the limit once and the pain is unbearable, it's all over with the patient." (Physical therapist, public practice)

Care providers admitted the existence of PIEM, but stated that it had little or no impact on their practice for managing patients' condition as compared with side effects observed with other treatments.

Patient expectations about PIEM management included more information and transparency from care providers about what factors induced pain, how to assess pain, and whether it requires a specific treatment:

"There is no debriefing on this.... The first day and the last day, we fill in a little questionnaire for the physician to assess the impact of pain on our everyday activities. We fill it in, but it's the same, we don't get any feedback. I would have liked it, for example, if someone had told us about our progress after the sessions" (Female patient, 43 years old, Paris, disk herniation, physiotherapy during hospital stay and at an outpatient clinic)

They also expected education on how to better prevent stress and anxiety generated by pain, counseling on how to anticipate PIEM and on mobilization, and accurate adaptations of physical therapy programs according to pain intensity. These expectations were modulated by patients' own interpretations and behaviors concerning pain in general and PIEM. Some patients were fatalistic about PIEM:

"I think it's better to do the physiotherapy session without medication. That way if there is a problem you can feel it. If it hurts, it is better to take analgesics after pain occurs, because if you take it before and there is a problem, you won't feel it." (Male patient, 46 years old, Paris, ankle strain, physiotherapy at an outpatient clinic)

"In any case, even at home, I can't control pain, and I don't see how I could do so during physiotherapy sessions." (Female patient, 43 years old, Paris, low back pain, physiotherapy at an outpatient clinic).

Not all patients suggested that care providers should be more involved in PIEM recognition and management.

Expectations about pain killers were heterogeneous. The criteria mentioned were efficacy, rapidity of action, absence of side effects, and methods of administration. For over-the-counter drugs and acetaminophen, expectations were mainly for more efficacy, whereas for opioids, they were mainly for fewer side effects:

"Paracetamol. It might be worse if I didn't take it, but it is totally inefficient" ( Male patient, 57 years old, Paris, hip osteoarthritis, physiotherapy during hospital stay in a tertiary care teaching hospital).

"I had a lot of pain killers. Paracetamol and codeine didn't do any good. When I was on these prescriptions, sometimes I didn't take anything because I figured 'if that's what a pain reliever is...' I don't take medicine if it does nothing for me." (Female patient, 43 years old, Paris, disk herniation, physiotherapy during hospital stay and at an outpatient clinic)

"I couldn't stand tramadol. I was dizzy, I threw up, and I didn't even know my name anymore." (Female patient, 24 years old, Houilles, physiotherapy in secondary care hospital)

"I felt nauseous, I even threw up, and the last time I took tramadol, I took 2 tablets and I was completely drowsy, really, really sleepy but when I was taking 6 tablets a day, I didn't experience those side effects." (Female patient, 43 years old, Paris, disk herniation, physiotherapy in an outpatient clinic and in a tertiary care hospital)

Care providers had few expectations about PIEM management or specific continuing education programs on PIEM, either because they were satisfied with actual practice or not sensitized to PIEM. They did express a need for tools or procedures to better identify, assess, and follow PIEM:

"I don't really know how to measure the pain... It is just like these super PAS scales from 0 to 10, but you'll never manage to get an orthopaedic surgeon to keep this thing in his pocket. Even my goniometer, that I use every day, I can't keep it in my pocket." (Orthopaedic surgeon, Paris, tertiary care hospital)

"For pain assessment, there have been policies that were widely circulated that might give you a false good conscience, saying 'I've used my little EVA scale, so there, there's no problem'". (Geriatrician, Issy Les Moulineaux, Paris suburb, secondary care hospital)

Care providers' opinions about pain killers strongly depended on the images reflected and opinions expressed by patients, and the criteria that they mentioned were similar to those defined by patients.

As for continuing medical education programs, care providers mentioned the already large number of educational programs and their limited time to participate in these programs:

"We are so overwhelmed. We choose one or two and that's it. It is impossible to manage if you get in any deeper." (Orthopaedic surgeon, tertiary care hospital, Paris)

Several physicians pointed out the need for a large sensitization campaign on pain before proposing specific educational programs on PIEM. Physicians expressed a need for better education on the pharmacokinetics of analgesics and on prescriptions for physiotherapy:

"My expectation would be to have explanations on drug therapies for pain, their side effects, those I can associate. I'm missing that. I receive physician visitors from the industry, but that's not enough." (Physical medicine and rehabilitation physician, Issy les Moulineaux, Paris suburb, secondary care hospital)

Education on the patient-physician relationship concerning pain management was also mentioned:

"(If you should have an education program on 'pain and mobilization', what format would be most appropriate?) I think the psycho-behavioural approach is important. It's important for almost everyone. The relationship you have with the patient has a great influence on pain. A relationship of trust can help them forget about pain and concentrate on movement." (Physiotherapist, outpatient clinic, Villiers le Bel, Paris suburb)

Physicians insisted on the need for educational programs aimed at increasing competency rather than knowledge and on interdisciplinary programs because PIEM management implies different and complementary skills.