When Social Inequalities Produce “Difficult Patients”: A Qualitative Exploration of Physicians’ Views

“Difficult patients” are most often characterized by their personality, behavior, or pathology. Little research has been done to understand how some patients, because of their disadvantaged social position, are perceived as “difficult” by their physicians. Our qualitative, exploratory study was conducted to understand how social inequalities contribute to this perception of a “difficult patient.” It was based on 12 semi-structured interviews with physicians, in Montréal, Canada. Participants identified three main challenging factors: a perception of excessive time required to manage these patients; a feeling of ineffectiveness, despite this additional time spent; and the pressure to “do something” about needs they perceive as serious, despite this feeling of ineffectiveness. To adjust their practice to the specific circumstances of these disadvantaged patients, they feel it is important to establish good relationships with them, to set realistic objectives, and to increase interprofessional interactions. We discuss these findings in relation to three issues that contribute to this sense of difficulty: the social distance between physicians and patients, the social determinants affecting patients’ health, and certain aspects of the health care system that impede the above-mentioned adjustments in medical practice. By exploring, social inequalities in health care access not through the experience of patients, but rather through the perspectives of the physicians who feel unable to protect them, this analysis highlights that the scope of action required to address such inequalities far exceeds physicians’ practice.


Introduction
Representing around one in six patients (Hahn et al., 1996;Hinchey & Jackson, 2011;Jackson & Kroenke, 1999), difficult patients are most often characterized by their behaviors (complaining, aggressive, demanding, manipulative, overconsuming of care), medical problems (often multiple in one patient, responding poorly to treatment and causing pain), emotional distress, or mental illness (often not recognized as such; Elder et al., 2006;Haas et al., 2005;Hahn et al., 1996;Hinchey & Jackson, 2011;Jackson & Kroenke, 1999). They are much less often characterized by their social status (such as gender, race, or socioeconomic status [SES]), and when they are, this status is not necessarily one of underprivilege. Thus, with respect to socioeconomic stratification, patients with prestigious social status can be experienced as difficult because they make their physicians nervous (feeling a strong pressure to "perform well" with patients they perceive as possibly quick to criticize them) or overly empathic (identifying too closely with a patient from a social status close to their own) (Roberts & Dyer, 2003). Interaction with patients living at the other end of the socioeconomic spectrum-that is, in poverty-is, of course, well documented, but mainly from the standpoint of physician-patient communication. In particular, it has been shown that physicians' sense of not really understanding these patients can transform into irritation when, for example, they observe these patients' poor compliance with treatment (Bloch et al., 2011).
An original feature of our research problem is thus its sociological dimension. Our interest in "difficult patients" did not focus on challenging characteristics in the "absolute" (such as patients' personality or pathology), but rather on characteristics that are challenging because they are placed in a certain context, that is, relatively disadvantaged positions within a social hierarchy. In this, we fully agree with Levesque and colleagues (2013) who observe that access to 894280S GOXXX10.1177/2158244019894280SAGE OpenCarde research-article20192019 1 Université de Montréal, Quebec, Canada care cannot be broken down into factors that relate either to patients or to the system, as it is a product of their close entanglement. For example, low income ("patient" factor) is an obstacle to health care access only if services are not free ("system" factor). In our study, our aim was to understand how a certain combination of social position ("patient" factor) and social hierarchy ("system" factor) could trigger a physician's perception of difficulty and result in differentiated health care access for the patient. In this sense, our intention was to renew the exploration of social inequalities in health care access. Indeed, although such inequalities are usually identified through quantitative differences in the amount of care received or in qualitative differences in the experience of receiving care, we chose to analyze instead the experience of providing care. More specifically, we explored the experience of physicians who find some patients "difficult" because of the social inequalities that burden them.

Method
Our decision to use a qualitative methodology was based on the experiential nature of our data (the description of clinical situations experienced as difficult by physicians), our inductive approach, and our objective of understanding processes rather than measuring their frequency (Pope & Mays, 1995). We therefore conducted semi-structured, in-depth, face-toface, interviews with seven family physicians and five family medicine residents. 1 Three researchers led the interviews: a family resident and two senior researchers-a family doctor and a qualitative sociologist who is also a public health doctor. A pair of researchers conducted each interview, except for the last one. The interviews, lasting 45 min on average, were recorded and transcribed verbatim. The study was reviewed and received full ethical approval from the Research Ethics Board of the health care facility where the participants were recruited. 2 Our participants were assured of anonymity and confidentiality and signed an informed consent form before the beginning of their interview.
To increase the likelihood that our participants would have a large number of cases to present to us, we recruited them in a public health care facility in a downtown Montréal (Canada) located in a neighborhood whose rate of poverty is considerably higher than that of Montréal as a whole-45% of inhabitants are low income versus 29% in Montréal (Direction de santé publique de l'Agence de la santé et des services sociaux de Montréal, 2008). Also, as it was important for us to interview physicians who were interested in reflecting on the issue of social inequalities, we recruited the physicians who volunteered following a presentation of our study protocol to the medical staff of that facility. As our study was an explanatory one about the ways in which social inequalities produce "difficult patients," its objective was to bring new insights to the topic rather than a comprehensive examination of the subject. Therefore, we did not try to reach saturation to define our sample size. Instead, we considered the "information power" of our sample. The more information power a sample holds, the fewer participants would be needed. Malterud and colleagues (2016) propose criteria to evaluate the information power, including the aim of the study, the sample specificity, and the quality of dialogue. We judged that our sample's information power was rather high because (a) the aim of our study was rather narrow, as it was set at the intersection of two research questions ("what is a difficult patient" and "how do social inequalities interfere with medical work"); (b) our sample specificity was dense (physicians working in the same clinic and willing to reflect about inequalities); and (c) the quality of dialogue between researchers and participants was high, thanks to the complementary characteristics of the pairs of researchers who led the interviews. We then decided to limit our sample size to 12 participants.
Three of our participants were men and nine were women. Physicians' work experience ranged from 1 to 26 years and residents were in their second year of residency.
The semi-structured interviews allowed our participants to freely elaborate on their experiences, in their own words, rather than according to our assumptions. The interview grid covered (a) the impact of social inequalities on medical practice and in what ways they make this practice difficult and (b) the physicians' management of these "difficult" patients. To ensure the responses would not be biased by medical considerations, we specified to our participants that our questions referred to patients for whom it was not their medical condition that made them difficult (e.g., chronic pain or terminal stage of a disease).
The thematic analysis was conducted by one of the researchers (the author), who first read the transcripts several times, considering both latent and manifest content (Bengtsson, 2016). She then generated codes, from which she identified themes, later refined into subthemes, which captured some insight to the research question (Braun & Clarke, 2006;Vaismoradi et al., 2013), that is, the social processes that "produce" patients whom physicians find difficult. This coding process included analysis of the data's experiential dimensions (such as physicians' emotional experience) which were carefully described and associated with the meaning participants gave to them. This analysis was contextualized on both a meso (the health care facility and its neighborhood) and a macro level (the provincial health care system and systematic dimension of inequalities). It was conducted both within each interview and across the 12 interviews. The rest of the research team carefully read the transcripts, discussed the results, and validated the research report (Carde et al., 2018).

Results
Questioned about the social inequalities that perturb their medical practice, the physicians spoke about poverty, low education, unemployment, and social isolation. Thus, they viewed inequalities solely through the prism of the socioeconomic hierarchy. Our follow-up questions on other types of inequalities (gender and race, in particular) elicited only fragmentary responses. Accordingly, our findings are focused on low-SES patients. However, the objective of the article is not to describe low-SES patients' access to health care. It is to identify the processes that might make such patients difficult patients.
The expression "difficult patient" has less to do with a patient's characteristics and more to do with the subjective experience of the patient's physician; it is from the physician's perspective that a patient can be "difficult." We will first present this subjective data (the different feelings encompassed within this general notion of "difficult experience"). In a second step, we will identify the strategies adopted by physicians to adjust their practice to the specific features they associate with these patients.
It is important, when reading these results, to take into account at least two of the specificities of the Canadian health care system: (a) this system provides residents access without paying out-of-pocket to medically necessary hospital and physician services but not (or only partially) to other health care services such as dental care, psychotherapy and prescription drugs; and (b) most physicians bill the government on a fee per service basis.

Why Some Patients Are Viewed by Their Physician as Difficult?
The irritants identified by our participants are often interconnected and mutually aggravating. Consequently, the three main ones presented below frequently overlap in interviews.
Time pressure was omnipresent in our participants' statements. If managing a low-SES patient may be experienced as difficult by the physician, it is primarily because it takes time. As we will see below, because physicians consider it important to establish a quality relationship with such patients, they will spend more time at each consultation, see the patient more often, and even, in some cases, call the patient regularly. Furthermore, given the diversity of the patient's social and medical problems, physicians often need to contact a whole series of partners (pharmacist, social worker, specialist colleague, etc.) to obtain information or make appointments, which is also time consuming.
However, the "extra time" that physicians find difficult is not just the time they must devote to their patients but also the wait times that inexorably slow down the progress of treatment. Unable to afford private care, patients must endure long wait lists to get into the public health care system; then, once enrolled in the public system, their treatment may be blocked by their incapacity to pay for certain products-such as the solution to be taken before a colonoscopy, which is not covered, even though the procedure itself is covered-or difficulties related to transportation or getting time off work (which can lead to missing an appointment). Finally, a physician is sometimes unable to communicate a crucial piece of information because the patient has no personal telephone or stable postal address.
Occasionally this "extra time" becomes "lost time." The second irritant is the feeling of ineffectiveness-not producing any significant or desired effect-when, despite the added time spent on the patient, the treatment remains unsatisfactory. According to one participant, this feeling is painful because doctors "all have in some way this God Syndrome, (we) want to help and be the one who can solve the problem." This sense of ineffectiveness is often associated with a sense of impotence-the inability to take an effective action-in the face of factors over which they have no control. Some of these factors relate to health literacy: Because of their low level of education, some patients would not understand what they are told and would be embarrassed to say so, or even to admit they do not have the means to follow the physician's advice or to buy the care prescribed. For their part, physicians admit they feel unable to find the "right words." But the main factor contributing to the physicians' sense of impotence is their patients' lack of income. They feel that their patients do not follow their recommendations because they do not understand them, but also because they have more short-term survival priorities. A particular sense of impotence is experienced when physicians resign themselves consciously to "not doing the best" for their patients. For example, a physician might not be able to prescribe physiotherapy for a patient in pain because it is not sufficiently available in the public system and too expensive in the private system. Faced with a patient in tears, he may decide to prescribe a medication with potentially serious side effects.
This feeling of impotence is heightened when physicians sense that poverty is not just an obstacle to patients' treatment but is also the source of their health problems. One physician gave the example of a patient who was anxious due to his economically precarious life situation. Unable to prescribe psychotherapy (same access problem as for physiotherapy), the physician prescribed an anxiolytic, even though he knew it could not be a long-term solution. Likewise, he prescribed an anxiolytic to a worker stressed by his job, even though he felt that improving the patient's working conditions would be more effective. He called this process "medicalizing social inequalities," because he felt that physicians use medical means to treat problems that really emanate from the social realm (such as the stress caused by living in poverty) and should be addressed through social means (such as social benefit allowances or housing). Physicians, knowing no other way of responding to their patients' distress than to offer medications, are compelled to contribute to this medicalization: We have to manage that anxiety . . . We doctors, what do you want us to do? We're powerless against this, so we use what we know, our tools. . . . A carpenter uses a hammer and nails, and we use pills. Like the other doctors, I try to do it as little as possible, but at a certain point, there's this pressure, this demand.
For their part, patients have no other choice but to see their physician; it is easier to go to a medical office than to find a job, and a physician can be consulted for free, whereas other health professionals, who would sometimes be more suitable, cannot (such as a physiotherapist for rehabilitation exercises, a psychologist, or group activities to counter social isolation).
Physicians' frustration is further amplified when, feeling impotent toward this social context that prevents them from doing what is needed, they have the impression that their patients expect them to resolve their social problems. Behind the explicit request ("heal me") would lie hidden the real request ("make life easier for me") before which physicians feel particularly helpless. One such case would be a man injured at work who asks his physician to relieve his pain, but whose real expectation is that the physician will help him to have his work accident recognized by his insurance.
Finally, several physicians said their frustration could be aggravated by their patients' behaviors. When patients present a long list of symptoms that are not easily associated with any specific diagnosis and are quick to criticize their physician for not taking them seriously or not knowing what to respond, they arouse in him feelings of incompetence and impotence. This is the case regardless of the patient's social status, but is accentuated in a context of poverty, because there is a higher risk that the treatment will fail (for the above-mentioned reasons), making the physician's stance even more uncomfortable in the face of the patient's reproaches.
After time pressure and the feeling of ineffectiveness, the third and final irritant mainly identified by our participants was the pressure to do something. Some physicians said that, even though they feel their efforts are not very effective, they would feel guilty if they did not do all they could to help these patients. Indeed, they considered that, as ineffective as it might be, their medical practice is the only possible way out of the vicious cycle in which their patients are caught; to get out of poverty they need to get well. The fact is that, while health problems are largely determined by social factors, the converse is also true: health problems influence social position. For example, the patient injured on the job must be treated so that he can get back to work and meet his needs. This pressure to do something is amplified when the patient is isolated, consults no other professionals, and lacks the resources (educational, economic) to find any. Some physicians thus feel overwhelmed by the responsibility they feel they bear alone but are unable to assume adequately because they lack the skills of other professionals that would be required by their patient's situation (e.g., psychologist, social worker). One physician described one of her patients, an adolescent "drop-out," anxious and isolated: "He puts me in a situation where I'm the only one who can see him and witness what's happening in his life, and where I feel rather powerless given the many problems he tells me about." Some physicians suggested that not all of their colleagues are equally susceptible to this pressure to do something. Their own decision to practice family medicine in a socially disadvantaged neighborhood might reflect an interest in psychosocial issues that would be less common among physicians practicing surgery or intensive care, who are more interested in the technical aspects of their profession.

The Management of Difficult Patients by Their Physician
When asked about how they adjust their practice to the specific characteristics of difficult patients, given the social inequalities in which the latter are caught, the physicians responded with the following three main strategies.
First, they stressed the importance of building a good relationship. Often this task is arduous due to the patient's complex relationship history, but also because forging this relationship requires precisely what physicians lack-time. Furthermore, it involves perseverance with a patient they find difficult for the above-mentioned reasons but also sometimes because of the missed appointments, the displayed indifference, or the unexplained poor compliance. This bond will often remain fragile, at risk of being shattered if the physician antagonizes the patient by not being able to find tactful and supportive ways of admonishing him or her for a risky behavior. One physician, for example, took great care to avoid any paternalism or moralism when trying to warn a young female patient about the risks-infectious and relational-which her new sexual partner appeared to be exposing her to. However, this bond is precious, especially if the person is isolated and has few other contacts aside from the physician. Being as close as possible to their patients' daily life allows physicians to adjust, day by day, not only their communications but also their medical interventions, by seizing opportunities as they arise. For example, a physician might take advantage of a female patient's separation from a "controlling" partner to encourage that patient to take charge of her own health. The effort to create this quality relationship is therefore intended to counter the last two irritants mentioned above (the feeling of ineffectiveness and the feeling of guilt for not doing everything possible), but it contributes, paradoxically, to the first of them, the excess time devoted to the patient.
The second strategy is to be realistic. Physicians feel they must be careful to prescribe only tests or treatments covered by public health insurance, and, more broadly, to set objectives that are within the patient's capacity, and in limited numbers. Faced with a patient who presents a cascade of problems, this means adopting a "small steps" strategy that involves abandoning an ideal management and focusing instead on modest objectives that, being easily attainable, protect both the patient and the physician from becoming discouraged. For instance, the physician might focus on encouraging a small change in eating habits and postpone giving advice on smoking and exercise. A participant recounts how she adapted her intervention with a diabetic patient: At first, you want to improve his diabetes and improve his life, you give him lots of advice, you give him lots of papers, you spend forty-five minutes explaining to him and you expect there will be a change. Then there is none! Then you have the impression that you want more than the patient. At the beginning, it's a lot of helplessness-"What can I do more? Who else can I refer him to?" . . . Then I realize that this is not the right strategy . . . Now, our appointments are much shorter, I work on one issue at a time. I know there are ten issues that I should tell him, but I'm trying to get to one or two and no more. It doesn't matter, it'll go to next time, it's more bearable this way.
This strategy requires to know the patient's living conditions, which might seem difficult to some, such as this participant who alluded to an "impostor syndrome, even if I really try to empathize and imagine what their daily life is like," but the first strategy (building a good relationship) is precisely supposed to help the physician in getting this knowledge. Also, this second strategy contributes to the second irritant, the feeling of not doing what is best, but this is only in the short term, as it is expected to be effective over the long term.
Finally, the physicians stressed the importance of interactions with colleagues. Talking with colleagues allows them to release some of the tension associated with following these patients and obtain advice. At the same time, it makes them more sympathetic toward those patients because, when recounting all their problems, "you realize that, after all, he doesn't have an easy life." Collegial interactions also occur when the physician can refer a patient to other professionals (social worker, nurse, psychologist, etc.) whose specific skills are complementary to those of the physician; this helps to alleviate the physicians' impression that the patient's fate rests solely on their shoulders. When referring a low-SES patient, some physicians feel it is important to accompany the patient-make the appointment, and then introduce the other professional to the patient in person. Although this strategy can help relieve the physicians' feeling of ineffectiveness, it can also rapidly intensify the sense of time pressure, as finding out about possible resources, making the referral, and even accompanying the patient physically are all processes that take physicians' time.

Discussion
Social inequalities in health care are differences in health care access observed in groups situated at different positions within a social hierarchy (e.g., socioeconomic or racial; Aïach & Fassin, 2004). They are usually captured by readily quantifiable indicators, such as frequency of medical consultations and of preventive or curative care (Hutchison, 2007).
More rarely, they are captured by indicators that involve a subjective assessment by the beneficiaries of health care access, such as their perception of the degree of empathy shown by their health care professional (Bedos et al., 2003;Mercer et al., 2016). Our study delved more deeply into this second angle, that of subjectivity, but with an original approach: We explored the experience not of beneficiaries of access, but rather of their physicians. More specifically, we asked them how the socially disadvantaged position of their patients could perturb their medical practice and make it difficult. In tracing this guiding thread of subjective notion of "difficult patients," we uncovered social processes that emanate from the unequal structure of society and thereby differentiate individuals' access to care according to their more or less privileged position within this structure. As such, our approach brings new insights in social inequalities in health care access, but also in "difficult patients" as the literature usually characterizes them by their personality or their medical problems rather than their social attributes. Therefore, it does not pretend to give a comprehensive overview of low-SES patients' health care access, nor one of difficult patients, but rather focuses on the intersection between these two issues: how some socially disadvantaged patients can be difficult in their physicians' views.
Our participants identified three main irritants related to disadvantaged patients-time pressure, the feeling of ineffectiveness and the pressure to do something-and three ways in which they adjusted their practice to the specific characteristics of those patients-building good relationships, setting realistic objectives, and interacting with colleagues. Obviously, none of these irritants (nor of these practices) are specific to socially disadvantaged patients. Our point here is to understand how they can be specifically triggered by social disadvantages. To do so, we will now discuss three main sources for the sense of impotence that underlies this perception of difficulty: the physicians' perception of (a) the social distance separating them from their patients; (b) the social determinants that influence their patients' health; and (c) the rigidity of the health system.

Social Distance: Evidence to Be Examined
Difficulties in communication between physicians and their low-SES patients have been widely studied. In particular, much attention has been focused on the bidirectional difficulty described by our participants: patients do not understand physicians' recommendations, and physicians do not know enough about the conditions of living in poverty to understand their patients' behaviors.
Some authors, when analyzing these communication difficulties, use the concept of social distance, to which our participants made reference without naming it, that is, that the reason physicians and patients do not understand each other is that they are at opposite ends of the social spectrum. The former is unable to comprehend the day-to-day experience of living in poverty, and all the more so because the latter is embarrassed to talk about it and thus is unable to convey all his needs (Bloch et al., 2011;Loignon et al., 2010). Patients, on the contrary, are not comfortable enough with medical jargon to understand their physicians, and they feel poorly understood, and even judged, by them (Ventres & Gordon, 1990;Verlinde et al., 2012). We believe it could be helpful to bring three nuances to this notion of social distance.
The first is that, because communication is a dynamic process, it is really the interaction between the physician's attitude and that of the patient, as modeled by their respective SESs, that affects physician-patient communication (Verlinde et al., 2012). In this respect, Willems, De Maesschalck, et al. (2005) speak of a vicious cycle in which: (a) the behavior of the low-SES patient (asking few questions, expressing less affect and less interest in taking decisions) makes the physician less inclined to provide information in which the patient is apparently not interested and seems incapable of understanding; (b) the patient, noting this reaction, becomes more entrenched in his passive attitude; and (c) consequently discourages the physician from any attempt to develop a partnership with him, such that the physician will explain even less and become increasingly directive. In the current climate of promoting physicianpatient partnership, negotiation, and cooperation, it is easy to understand our participants' frustration about patients who appear reluctant to assume any active collaboration. This frustration was even tinged, in some of our participants, by perplexity. Faced with patients whom they described as fatalistic, focused on their short-term relief and not very interested in long-term issues, nor keen to decide for themselves on their care, the physicians felt confronted with a dilemma: should these patients be compelled to choose the modalities of their care (at the risk of increasing the anxiety, already considerable, they feel about their living conditions), or should their physicians choose for them, in a more paternalistic way, and thereby maintain them in their minority position?
The second nuance is that each of the protagonists in the physician-patient relationship is characterized not only by a given SES, but also by a multitude of other social statuses (age, gender, ethnicity, etc.), and it is actually the combination of all these statuses that intervenes in the physicianpatient relationship. Thus, according to one study, patients' satisfaction with their relationship with their physician diminishes as the degree of similarity between the two individuals decreases on a series of social statuses (race, gender, age, and education; Thornton et al., 2011). The study concluded that perceived satisfaction is the product of the aggregated degrees of similarity for each of the social axes studied. In our opinion, it would be interesting to see whether the process is less cumulative than interactive, from an intersectional perspective (Hankivsky, 2012), in which each of the social statuses (of patient and physician) influences the physician-patient interaction differently depending on all the other statuses involved.
Finally, from a constructivist standpoint, social distance is less an objective reality than the product of a social construct of otherness. Placing individuals at the other end of the social spectrum allows us to designate them as Others, and therefore "necessarily" different from Us. Stereotypes-abusive generalizations about a social group-are then readily used to consolidate this construct. Numerous studies on the social distance between physicians and patients living in poverty mention the stereotypes used by the former to "explain" certain behaviors of the latter, such as their poor compliance with treatments (Bloch et al., 2011;Woo et al., 2004). These stereotypes consist, for example, in assigning the responsibility for their poverty to the individuals themselves, who are seen as lazy (Willems, Swinnen, et al., 2005), and in suspecting patients of manipulation and abusive behaviors, such as overconsumption of health care and welfare fraud (Desprès, 2017;Loignon et al., 2018). If we did not encounter any such stereotypes in our study, it is probably because we were questioning physicians from the angle of social inequalities, and this focused their responses on the responsibility of society rather than on that of patients. Also, our participants were physicians who volunteered for the study, an undertaking that likely reflected their sensitivity toward social inequalities. It was to this particular sensitivity that several of them alluded when referring to the "pressure to do something."

Social Determinants of Health: From Awareness to Practice
Through establishing a good relationship with their patients, physicians hope to understand their patients' living conditions so that they can adapt their interventions in a realistic way. In other words, they try to be aware of some of the social determinants that affect their patients' health and health care. We have seen how the sense of impotence appears when the considered social determinants (such as poverty) seem out of physician's perimeter of action. Expedient practices such as easing their impact on health can be more or less satisfying. For example, physicians might feel uncomfortable trying to relieve social issues with medication and one participant talked about "medicalization of social problems" when he resigns himself to prescribing an anxiolytic to treat anxiety induced by working conditions. Although the expression "medicalization of social inequalities" strongly expresses the frustration of physicians in being forced to behave as doctors even though it does not seem appropriate to them, it deserves to be discussed. If we stick to Conrad's classic definition-"Medicalization consists of defining a problem in medical terms, using medical language to describe a problem, adopting a medical framework to understand a problem, or using a medical intervention to 'treat' it" (Conrad, 1992, p. 211)-then the above case is not a matter of medicalization of inequalities or of poverty. The medical intervention (a prescription of an anxiolytic) is not supposed to treat the "problem" (poverty or inequality): it targets one of its consequences (its impact on psychological well-being). In addition, this "problem" is not defined in medical terms. It is precisely because the physician does not reduce it to a phenomenon that could be entirely objectified by the medical gaze (as menopause is or the turbulent behavior of children since they have been medicalized) that he or she feels helpless. On the contrary, the term "medicalization of poverty" seems more justified-even though still debatable-when it refers to the trend of public authorities to manage poverty through a medical lens. This trend has been increasing in recent years, for example, in the United States (Hansen et al., 2014), in Quebec (Barbeau et al., 2008), and in France (Thelen, 2008): Social assistance benefits are becoming more difficult to obtain for those of the poor who are not declared sick (or unable to work) by a doctor. Poverty, a socially situated process, becomes an individual state whose objectification requires a medical look. Significantly, one of the consequences of this policy is that the persons concerned, to "authenticate" their diagnosis, are often required to take a medication: to take a medication, then, is part of a medicalization process.
The participants' claim to try to be aware of their patients' living condition to adapt their intervention does not seem widely corroborated in the literature. This is likely because our sampling selected socially aware physicians. Although it appears that physicians caring for low-SES patients are generally vigilant about the "gaps" in public health coverage (avoiding prescribing drugs that patients cannot get because they are not covered by public insurance and are too expensive for them to buy; Willems, Swinnen, et al., 2005), it is much less common for physicians to ensure that other aspects of their patients' living conditions do not prevent them from taking the drugs prescribed. In this regard, Bloch and colleagues (2011) present the example of a patient whose medications should be taken 3 times a day after meals, but are not, because he does not eat three meals a day. This is also the case for a patient who is unable to take his antidiabetic medications because they give him diarrhea, which is problematic, as he has difficulty walking and lives in a shelter with only two toilets for 60 residents. By paying attention to these patients' living conditions, the physician could have avoided these unsuitable prescriptions. Moreover, these living conditions can make the prescriptions useless even when the patient actually takes them. It is the case of a patient living in an apartment covered with mold that causes chronic respiratory failure, for which his physician prescribed drugs that were bound to be ineffective until such time as these housing conditions could be rectified (Bloch et al., 2011).
Thus, generally speaking, physicians tend to focus on medical aspects to avoid becoming involved in the more social aspects (income conditions, housing, etc.) that are considered not to be within their remit (Bloch et al., 2011) nor relevant (in the name of the principle of the universality of the biological condition; Desprès, 2017). That being said, when physicians do attempt to take these social determinants into account, they run the risk of doing so clumsily, based on a mistaken interpretation of the implications associated with their patients' SES. Desprès reports the case of physicians who do not offer advice for smoking, eating, or drinking habits because they anticipate a reluctance of the patient toward such advice or aggravating the patient's stress. According to Desprès, these well-intentioned physicians unduly lower their standards of care. This interpretation sheds a different light on the "small steps" strategy described by our participants, that is, is there not a risk that adjusting their practice to modest but so-called "realistic" objectives will deprive the patient of certain options and thus ultimately be detrimental to him or her?
It is particularly illuminating here to reverse the focus and examine the expectations of low-SES patients as described in various studies. These patients say they attach great importance to the bond with their general practitioner, to relational continuity (knowing one general practitioner in particular and being able to consult him or her as needed), to sufficiently long consultations during which they can talk to their physician, and to the latter's demonstration of empathy and understanding of their life context . This echoes our participants' efforts to establish a quality bond. However, such efforts might not be seen as sufficient by the patients. Indeed, according to various studies, the reason why these patients expect their doctor to understand their life context is that they feel that a general practitioner is all the more competent when he or she knows their personal life well enough to be able to do more than provide care. When asked why they consult their physician, although they might first give a medical reason, they go on to describe the social context within which this medical reason is embedded and which is part of the reason for their medical visit (Popay et al., 2007b). They expect the physician, who is in a position of authority, to help them navigate a system in which they feel powerless. This might mean, for example, making their case to an administration, such as writing a letter to ask for a change in housing for health reasons. In this respect, we recall our participants' unease when they feel their patients expect them to become explicitly involved in this social realm (e.g., not only relieve the pain of an injury at work but have the work accident recognized by their insurance provider): Even though our particularly socially aware participants try to adapt their practice to their patients' living conditions, their efforts probably remain below the expectations of their patients as they do not intend to exceed the realm of care. It is indeed precisely their sense of impotence toward such conditions that feeds their sense of difficulty with these patients. However, according to Popay et al. (2007b), the general practitioner is only one level among others that patients are likely to activate, although one that has the advantage of not being stigmatized. This observation is in contrast to the statements of our participants, whose difficulty arose precisely from their feeling of being their patients' only advocate, and thus feeling the "pressure to do something."

When the Health and Social Services System Exacerbates Tensions
Within these combinations of social position ("patient" factor) and social hierarchy ("system" factor) that trigger their perception of difficulty, physicians principally identify the "patient" factors, and among them, poverty, as most significant. More subtly, some allusions are made to "system" factors, and especially the health and social services, for example, to deplore that public medical insurance does not cover all of their prescriptions. Throughout the interviews there is also a sense of the rigidity of a system which restricts the flexibility that physicians need to adapt their practice to low-SES patients.
As a first strategy, our participants stressed the importance of getting past their feeling of arduousness to take time with their difficult patients, in more frequent and longer consultations, to build a quality relationship. Their reports are corroborated in the literature. The first adjustment physicians make with their low-SES patients is to devote more time to them, providing consultations that are longer (Gulbrandsen et al., 1998) and more numerous (Asada & Kephart, 2007). However, physicians practice within a health care system that is more or less amenable to such adjustments. Thus, feefor-service remuneration discourages physicians from offering longer consultations (Mercer et al., 2016). And, paradoxically, measures to reduce the fee-for-service portion of physician remuneration may discourage physicians from following low-SES patients. For instance, France uses a payfor-performance (remuneration based on public health objectives) system in which physicians' remuneration is modulated according to their success in managing their diabetic patients. Given that economic insecurity is a factor associated with diabetic imbalance, physicians may be reluctant to follow diabetic patients living in precarious situations (Desprès, 2017). In Quebec (Canada), the imposition of a minimum number of patients to be enrolled with family physicians under penalty of financial sanctions could also, according to our participants, discourage physicians from taking on low-SES patients, precisely because of their time-consuming nature. Asada and colleague observe that, in Canada, low-SES individuals are on average less often followed by a general practitioner than the rest of the population, but when followed, they have more frequent consultations. Their hypothesis is that while general practitioners cannot influence a patient's "first arrival," they can influence the frequency of the following appointments-in this case, increasing frequency when they feel it is necessary (Asada et al., 2007). Our study confirms this hypothesis, reveals the sense of difficulty associated with this practice, and suggests that the health care system contributes to this sense of difficulty.
The third strategy cited by our participants is also corroborated in the literature on low-SES patients. It has to do with the intensity of interprofessional interactions, that is, discussing with peers to obtain support and referring patients to professionals with complementary skills, particularly social professionals (Norbury et al., 2011), and ensuring patients are accompanied in such referrals, at least by telephoning the professional (Willems, Swinnen, et al., 2005). These approaches can be geared toward meeting several needs at the same time, as in group dining activities that are beneficial to both physical and psychosocial health because they overcome social isolation (Rankin et al., 2009). However, the configuration of the health and social system can also hinder this strategy. General practitioners tend to counsel these patients themselves or to refer them to psychologists much more often than to social workers, thus responding to the psychosocial consequences of social problems rather than tackling them head-on (Gulbrandsen et al., 1998;Popay et al., 2007a). This tendency, observed even among physicians who wish to take social issues into account in their practice, is likely due to the fact that it is difficult for them to keep their knowledge of community resources up-to-date, as these resources are subject to numerous changes (closures, mergers, name changes) due to the instability of the public subsidies allocated to them (Popay et al., 2007a). This observation is entirely in line with our own findings and reinforces our participants' insistence on the importance of working within a multidisciplinary team, which obviates the need to be familiar with external resources.

Limitations of This Study
The limitations of this study are set by its methodological design. This was a qualitative study that sought to uncover as precisely as possible a series of processes leading from social inequalities to a physician's difficult experience, but without seeking to identify their frequency. Moreover, because our focus was on physicians' feelings of difficulty, we did not explore the characteristics of providing care for disadvantaged patients not experienced as difficult by physicians. If we had done so, we would probably have referred to the satisfying aspects associated with this care, such as the feeling of "making a difference" in the lives of the poor. Another limitation is that our sample was recruited from a strongly disadvantaged Montréal neighborhood. This increased the probability that our participants would have knowledge of numerous clinical cases corresponding to our topic, but it is possible that disadvantaged patients living in a more affluent, or less urban, neighborhood would present different problems for their physicians that were not raised by our participants. Moreover, while this neighborhood is disadvantaged in terms of its SES (high rate of poverty), the percentage of recent immigrants is lower than in Montréal as a whole (Direction de santé publique de l'Agence de la santé et des services sociaux de Montréal, 2008), its residents are mostly White and French-speaking, and thus belong to the majority population in terms of the racial and linguistic hierarchies spanning Montréal's society. If we had conducted the study in a Montréal neighborhood with more immigrants, these other hierarchies might have assumed greater prominence in our participants' statements. This might have provided more insight into the extent to which the links between SES and other statuses (such as gender or race) "produce" a difficult patient. This avenue would certainly merit further exploration in a future study.
Another limitation of our study is that we did not attempt to determine whether the fact that the participants were residents or physicians, men or women, might have influenced their difficult experiences with low-SES patients. Various studies have shown that physicians who are more stressed, less experienced, but also less interested in psychosocial issues and communication are more likely to perceive some of their patients as difficult (Haas et al., 2005;Hinchey & Jackson, 2011;Jackson & Kroenke, 1999;Krebs et al., 2006;Steinmetz & Tabenkin, 2001). The latter contrasts with statements by some of our participants indicating that it is because they are sensitive to social justice issues, as family physicians who have chosen to practice in a disadvantaged neighborhood, that they are particularly at risk of experiencing difficulty in their relationships with these low-SES patients. This contrast suggests that it would be relevant to further this explanatory study with a sample consisting of less socially conscious physicians, so as to understand how this would translate into the perception of "difficult patients." Finally, the "difficult patient" concept emerges from a relationship between two parties, physicians, and patients. Some studies have interviewed both parties and have found a concordance between their perceptions, that is, patients considered as difficult by their doctors are less satisfied with their clinical encounters (Hinchey & Jackson, 2011;Jackson & Kroenke, 1999). It would be relevant to replicate these studies while considering difficulties specifically originating from social inequalities.

Conclusion
In 1971, Tudor Hart described the inverse care law, according to which access to care decreases as the need for care increases. This phenomenon, which is observed even in universal health systems such as Canada's, exacerbates social inequalities in health, as the most disadvantaged are most at risk for their health and thus most in need of care (Hutchison, 2007;Mercer & Watt, 2007;Stewart et al., 2005). Our study corroborates this law by demonstrating how the most disadvantaged are exposed to particular difficulties when accessing care. By following their physicians' sense of frustration as a guiding thread, we identified a series of obstacles to their optimal medical management.
Many publications dealing with social inequalities in health thus stress the importance of the role of general practitioners in reversing the inverse care law (Casanova et al., 2015;Norbury et al., 2011;Starfield et al., 2005). Our study presents some of these avenues, such as taking the time to get to know the patient's daily life to create a bond of trust and set realistic objectives, and encouraging interprofessional interactions. However, we feel that it is crucial to conclude by highlighting that the difficulties envisaged in our study manifest themselves at the time of the medical consultation but originate from inequalities in exposure to a wide range of social determinants of health, which very often stand outside the range of the physician's scope of action. Physicians can mitigate the consequences, and it is very important that they do so, but actions aimed at the organization of the health care system and, beyond that, the "fundamental" causes of social inequalities in health (unequal distribution of money, knowledge, power, prestige, and social connections) would be more effective (Link & Phelan, 1995). This fact is precisely what gives rise to this feeling of impotence, which appears to be the basis of physicians' experiences with their "difficult patients."

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.