‘He should feel your pain’: Patient insights on patient–provider communication in Rwanda

Background Patient–provider communication is an interpersonal interaction between a patient and a health care provider. Objective This study explored patients’ communication preferences and perceptions on what factors influence the patient–provider communication in primary health care settings in Rwanda. Methods In-depth semi-structured interviews with 15 individuals including 8 with limited literacy. A thematic inductive analysis was used. Results Patients valued communication with providers and expressed the need for interacting with caring, empathic providers who can share all the information they want and involve them in their own care. Health literacy and power issues were factors that may influence patient–provider communication. Patients with limited literacy appeared to rely highly on health care providers for making decisions about and managing their health care. Conclusion The expressed preferences, including those of patients with limited literacy, aligned well with the patient-centred care model. There were indications of a power imbalance weighing on the provider’s side. Although patients with limited literacy were reliant on providers for decision-making, they were ready to be more involved in the care, suggesting a potential for improved patient involvement even for patients with paternalistic care preferences. These patients’ insights can impact policies and curricula to optimise clinical practice. Generated knowledge will contribute to the indispensable yet underdeveloped field of health communication in sub-Saharan Africa. Practice implications Findings call for more inclusion of patient perspectives in the patient–provider encounter. This could require more training of professionals and research on the topic, both in Rwanda and in other regions.


Introduction
This study aimed at gaining more insights into patients' perceptions of their interactions with nurses in PHC settings in Rwanda. The following questions guided the exploration: • What are patients' communication preferences?
• What perceived factors may influence communication?

Methods Design
This is an explorative qualitative study using in-depth, semistructured individual interviews of patients.

Setting and participants
People above 21 years old who had two or more interactions in the past 12 months with nurses in the OPD of health centres were included, regardless of the health problem. People considered too weak or sick to go through the interview were excluded.

Sampling
We strategically aimed for around half of informants to be patients who reported themselves as unable to read. We used the term 'patient with limited literacy' for this group. This ensured the richness of the data from this potentially vulnerable and often unheard group. During the selection process, we also aimed for gender balance to capture the perspective of both male and female patients. We applied a purposive sampling strategy. 36 The purpose was to get deep insight into views of patients who had experienced PPC in a consultation room at a health centre in Rwanda. Most informants were approached in the waiting room of two urban and two rural health centres. The interviews were conducted on another agreed day and location to allow time to 'digest' the experience of the recent encounter, and for patient convenience. We suggested to meet at our office and gave the alternative to choose another convenient place. Six people were approached in busy public places within 2 km to 5 km of the health centre and interviewed immediately to also include patients without recent contact with a provider. In these cases, the interviews took place in the researchers' car parked in a quiet place on the side of the road.
We approached 27 people in total, of whom 16 were eligible and willing to participate, and 15 kept the appointment.

Data collection
The interviewer (J.B.K.), trained in journalism and communication, conducted all interviews and was accompanied by another author (V.K.C.). Interviews were carried out from April to September 2016. Study questions were reviewed by all co-authors, who also contributed to the interview guide, which was translated into Kinyarwanda by the interviewer.
In two pilot interviews with patients with limited literacy, several sentences were incomprehensible or full of truisms without clear relation to the questions asked. This helped develop a simplified version of the interview guide for patients with limited literacy.
Interviews were conducted in Kinyarwanda and were audio recorded, lasting on average 58 min (shortest 24 min, longest 76 min). Interviews were transcribed verbatim in Kinyarwanda, anonymised and translated into English by a professional translator. The interviewer (J.B.K.) controlchecked a random page of each English transcript against the Kinyarwanda transcript and the corresponding audio recording and approved the translation.

Data analysis
A thematic inductive analysis guided by the framework approach was carried out. 37 MaxQDA 11, a qualitative data analysis software, was used.
Two researchers (V.K.C. and M.S.) familiarised themselves with the 15 transcripts and identified potentially relevant codes. The two researchers together selected, defined and refined emerging key themes to develop a thematic framework. One researcher (V.K.C.) systematically indexed (coded) the text from the 15 transcripts using the agreed thematic framework. Concurrently, data were charted by

Ethical consideration
Institutional Review Board of the College of Medicine and Health Sciences, University of Rwanda (CHMS/ IRB/216/2015 and 324/2016), approved the study. All participants signed an informed consent form prior to their participation. Participants with limited literacy were asked to mark their fingerprints on the informed consent form.

Results
We present the findings using exemplary quotes, as these elaborate the points we make while conveying a sense of the study field we could not ourselves paraphrase. The purpose was to give direct voice to an often unheard group: patients.
We use the terms 'she/he' and 'her/him' within quotes to represent a translation of the genderless personal pronouns in the Kinyarwanda language. Gender is specified only when the context allows it. A bracketed with ellipsis […] is used when we omit a portion of the text to simplify the quote, without altering its original meaning as expressed by the patient. Each quote is followed by a person identifier (P1 to P15) in Table 1.
Fifteen informants were included in the analysis, 7 females (4 self-reported as unable to read) and 8 males (4 self-reported as unable to read). The age range of the informants was 22-64 years (Table 1).
Six main themes emerged from the data: • patient preferences • confidentiality and trust • power issues • health literacy • patient satisfaction • provider lottery and patient choice.

Patient preferences
Patients brought up a number of wishes and expectations that they believed would improve the interaction with providers. Four subthemes were associated with patient preferences: (1) Welcoming, (2) Attention, (3) Empathy and friendliness and (4) the Right to know.
The first three subthemes can be embedded into the concept of 'caring behaviour'.
It was highly important for patients that providers receive them in an acceptable way. This is considered the starting point in establishing an instant rapport. A proper greeting may simply be friendly eye contact. Several patients reported that the way they are welcomed may determine the success of the rest of the encounter and a comfortable welcome enables them to share concerns.
'It doesn't require me a number of minutes to know whether or not the health care provider is a bad one. It shows through the way he welcomes you in when you enter into the room.' (P2, male, 35 years old) Patients requested more attention from providers. This included listening and moving away from distractions and interruptions. The most commonly expressed wish was that informants wanted providers to really listen to them. Patients recognised that this depends on providers: 'There may be a clinician to whom you start telling an illness but she/he does not understand, cannot list them the way you said. A good clinician is the one [...] to whom I explain and she/he listens to me attentively.' (P1, female, 38 years old) Patients, including patients with limited literacy, spontaneously requested providers to repeat or rephrase their problems so they felt sure to be understood: 'When I hear that he talks about everything that I have told him, I understand that he understood me well.' (P7, female, 29 years old) Several complained about providers being too hurried to actively listen: 'There are health care providers who even finish prescribing medications for you before you finish talking and just give you the paper and you go.' (P3, male, 55 years old) Some patients suggested providers could ask more about their feelings. Providers should also listen to patients' feedback. Patients believed their feedback is a way to improve the interactions. 'The problem that we have is that no meeting is organized for both the health facility staff and the people who live in the health facility catchment area so that we would be able to report the problems that we encounter with. If there was such a meeting, they would solve our problems.' (P12, male, 40 years old) Distraction during a consultation was a severe problem to patients: 'The first thing that he should do is to avoid any distraction while you are conversing.' (P3, male, 55 years old) Patients need more than good listeners. They requested empathetic providers who are able to sincerely put themselves in the position of the patient: 'He should also feel your pain as if the pain was on him.' (P12, male, 40 years old) Patients linked empathy with friendliness and ultimately considered it an expression of love: 'He has to talk to me in a friendly way and I should feel that he has a love for me as a person whom I have come to see.' (P7, female, 29 years old) Empathy includes the doctrine of treating patients as one would like to be treated: 'We would love to see that he is behaving like he is in our shoes and says [...] "let me treat this patient as I would like to be treated".' (P12, male, 40 years old) Being friendly was found in almost all interviews and several complained about experiences of unfriendly behaviour: 'They do not receive you empathetically as someone who is ill, like a suffering person who has come to see a health care provider. They do not care for you as a person who needs them to save your life. There is carelessness.' (P12, male, 40 years old) Some thought training could stimulate empathy: 'The health care providers must be trained so that they know that patients are like friends, close relatives or neighbours.' (P12, male, 40 years old) Others seemed less convinced that training was the only answer: 'He should listen to the patient's problem whether he was trained or not. But he has to use his heart as a human being in order to interact with the patient.' (P8, female, 64 years old) Patients expressed an uncontested desire for clear and precise information from providers about their health. This includes providers' thoughts, findings and writings and for some also the purpose of examinations: 'Sometimes, you see her/him writing clinical notes after examining but without explaining to you. Some do not even tell you what you are suffering from, and you go without any explanation, and I think this is a problem!' (P1, female, 38 years old) A patient had experienced that requesting further information was a problem: 'When I asked her "what do the results say? What have they diagnosed?" she rudely asked me "Are you in a position to ask me that? Are you supposed to know it?" She didn't tell me the illness they had found.' (P3, male, 55 years old) Patients also expected guidance, advice and answers. This would require a comprehensive and holistic approach:

Confidentiality and trust
This section includes issues around confidentiality and privacy, as well as disclosure.
Patients were concerned about confidentiality and privacy of the encounter with providers. Several complained about providers in that regard: 'It might happen that [...] the health care provider is not happy or he is lamenting like "The previous patient was weird" and so forth. He can also for instance warn the next patient, "Don't behave like the previous patient". I

Power issues
This section includes paternalism, power of knowledge, shared decision and patient involvement.
Issues related to paternalism were already mentioned under the theme 'patient preferences', which also gave examples of paternalistic practice. One patient compared providers' power to that of parents and gods, expecting them to act accordingly: 'Health care providers are like our parents, and they are also like our gods. They should therefore behave like a parent who is conversing with their child.' (P3, male, 55 years old) In return, this patient was ready to disclose everything, even follow orders: 'Because I personally cannot have any problem with the health care provider because I take him like my parent -I cannot hide anything from them. When they tell you "take off your clothes" and you undress, "do like this'" and you do so, there is nothing I can hide from him. I do not have any problem in front of the health care provider, I am ready to do everything that they order me to do without any problem.' (P3, male, 55 years old) Many informants felt they had to be obedient towards the provider: 'I should not have a disagreement with the health care provider because whatever he tells me is what I have to abide by.' (P10, male, 32 years old) Most informants regarded themselves below or under providers: 'My gosh! I have to be below her/him as we are not at the same position. Anyway, it is the reason I come to a health facility so that the clinician -who is above me, who has knowledge, who has learnt/studied something -gives me a piece of advice on how I should do/behave.' (P6, male, 58 years old) Patients expressed the view that the quality of care was justly conditioned by the respect patients show providers: 'Usually, he listens to you; he finds that you have given him the respect and he therefore helps you as somebody who is superior to you, yeah.' (P5, female, 30 years old) Some patients, particularly with higher level of education, would see themselves as equal to the provider. But most would highlight the provider as a superior person, who is busy, whose time is precious and who can't be wrong. Still, many other patients suggested the provider to be humble, http://www.phcfm.org Open Access and lessen herself or himself to correct the power imbalance for better communication: 'What should be done to make it better is that the clinician should simplify her/himself, not showing she/he is at the higher position. For me, I am still below but she/he should bear in mind that I'm at the lower position and talk to me humbly without making me worry.' (P6, male, 58 years old) Providers were considered to have the answers to the patients' problems. Informants' views of providers as superior to them were explained by comparing providers' knowledge and expertise with their own, and considering their dependency on providers. It was considered natural and unproblematic to be inferior to a provider: Some seemed to have a blind trust in providers' ability to make decision: 'She/he cannot actually prescribe medications which would kill me. That's why I abide by the decisions she/he takes… because she/he knows and she/he has also studied, she/he knows it very well enough, she/he doesn't learn anything bad that would harm a patient.' (P2, male, 35 years old) The concept of making a decision together with the provider was not easily understood: 'We would probably make bad decisions for ourselves. We could say for example they would give me two pills when they should give me four pills; and that would not be good for me.' (P3, male, 55 years old) A middle ground was to have a discussion before the provider decides: 'She/he can make the final decision, but this should happen after having had a discussion and this happens only when you are comfortable to talk to her/him.' (P9, male, 22 years old) Some patients did not experience an opportunity for discussions and hence no opportunity for shared decision: 'When you are conversing with them, most of the time they ask you something and they write. There is nothing that you discuss, so you cannot say for example "Don't write that". '  More knowledge was seen to allow patients to better interact with providers and engage in their own care. However, a patient claimed that explanation was sometimes denied because providers did not believe patients could understand it: 'If you ask him "What is this device used for?" he tells you "You cannot understand it". That is difficult […] He could simply tell me "this is used to monitor lungs, this is used to measure blood pressure".' (P9, male, 22 years old) Most informants would describe that their level of education has an impact on their ability to interact with health providers, for instance, the ability to follow an instruction:

Patient satisfaction
It is important that several patients perceived their interaction with providers as generally quite satisfactory. This satisfaction was commonly linked to a request for medication and lab tests: 'You cannot be pleased by a health care provider who listens to you, but doesn't give you medications that can help you, even if he has understood your illness.' (P8, female, 64 years old) Medication and laboratory tests were the main expectation of patients when seeking care, sometimes valued over good communication: 'We think that if we have been examined, we must be given medications. If you examine me and explain to me the problem that I have but you don't give me medications, I will ask myself, "Glad that you have explained to me about my illness; however, am I going to stay like that? Do I have to go home and keep the child there until when?" So, you understand that I will remain with dissatisfaction.' (P9, male, 22 years old) While most literate informants requested explanation over medication, several patients with limited literacy had difficulty accepting that medication might not be necessary for a sick child. They seemed to express a complete trust in medication: 'So, if the health care provider examines the child and tells you the illness that he has found but adds that it is not necessary for the child to take medication [...] you wouldn't understand that explanation as a person who brought the child for medical care. For me, I like the person who immediately gives me the medication that cures my child.' (P7, female, 29 years old)

Provider lottery and patient choice
Patients reported differences in the way they are cared for, particularly in the way providers communicate with them. Often, an idea of 'good' and 'bad' providers was expressed, as well as a perception that one meets 'a good provider' by luck: 'It's only by luck that you meet with a good health care provider. It is a chance that the Lord gives to you. Then you meet with the one who knows her/his profession and who cares for the human being.' (P12, male, 40 years old) We apply the term 'provider lottery' to describe this perceived aspect of luck and uncertainty in meeting 'a good' provider. A patient expressed how a bad experience could influence one's trust in other providers too: 'Yes, you cannot come back there. From that you make a general assumption that even all other staff are also like him.' (P8, female, 64 years old) Several explanations emerged for the differences in providers' ability to communicate, such as the provider's personality or mood. Also, work overload would influence care quality on any particular day. Others explained that the lack of communication training would cause differences in care quality: 'You don't feel that they were trained in the same way. When you talk to one now and next time talk to the other one you feel that there is a difference.' (P15, female, 45 years old) The inability to choose a provider was often described as problematic. One patient directly said: 'We do not have the right to say, "We want to be seen by such and such health care provider". If they told patients, "Once a patient arrives at the health facility, they are allowed to choose who will examine them", this would help patients much more.' (P3, male, 55 years old)

Discussion
Patients expressed their communication preferences clearly. They are summarised as a need for a caring provider, for access to more information about their health and health care, and for being involved in their care. These needs are embedded in the patient-centred care model that has impact on health outcomes. 16 Patients also revealed perceived factors that may affect PPC and thus require further consideration. Patient-centred care is a leading PPC model, yet there are indications that it is not fully conceptualised and applied in our context where the current practice of care may hamper its understanding and implementation.
Despite the relatively low levels of conventional schooling for the majority of our informants, they expressed critical thoughts and deep reflections on PPC. While our data represent narratives of patients, their preferences and needs did not appear as abstract or hypothetical. They rather seemed grounded in and backed by illustrations from their experiences. However, several patients would talk about the kind of behaviour providers should avoid, even if they themselves were reluctant to describe such behaviours as part of their own experiences.
It should be noted that literate patients easily acknowledged the lack of comprehension of a question and gladly asked for clarification before answering. This seldom happened with patients with limited literacy that would often give irrelevant or incomprehensible answers, or feel that everything was perfect. We interpreted this as attempts to hide ignorance or please the interviewer in some way. It may also reflect a Dunning-Kruger effect describing that persons of limited ability may mistakenly assess their ability as greater. 38 Another explanation could be termed as 'Nta Kibazo'-effect. Nta Kibazo means 'No problem' in Kinyarwanda, 'Hakuna Matata' in Swahili, and has equivalents in almost all other African countries. Our informants, particularly with limited literacy, have a background of poverty, being used to daily struggles for maintaining health and life in the family. 'Nta Kibazo' could reflect a cultural coping strategy for dealing with adversity -'accepting of life's difficulties'. 39 This might hide existing problems while generating an impression of 'comfort' among people who daily deal with adversity. 40 While these interpretations require further study, providers need at least to be aware of a potential tendency among patients particularly with limited literacy to underestimate their problems and overestimate their capabilities.
Furthermore, there was a tendency among patients to direct the conversation to satisfaction with tests and medications. This reveals a dilemma to providers: should they simply please patients with tests and treatments as requested, or engage patients through discussions about tests and treatment that may challenge patients' expectations? The former appears to be an easy and common approach, not least because of busy OPDs and difficulty of making followup appointments; however, it may also maintain or cultivate a paternalistic practice ill advised by evidence. 16

A caring behaviour
This is probably the most pressing need and is at the core of patient-centred care. Patients' descriptions of a caring provider follow Mosby's definition: 'Actions characteristic of concern for the well-being of a patient, such as sensitivity, comforting, attentive listening, honesty, and non-judgmental acceptance'. 41 It entails a provider who is friendly, listens, understands and shows interest and empathy. This is someone who 'feels your pain', as stressed by several informants. Active listening has been described as the key skill for effective patient-centred care, 16 and the feeling of not being listened to can damage trust in the provider. 4 To our informants, trust appeared as a key prerequisite not only for successful encounters but also for successful treatment and recovery. Other studies found trust in providers to impact adherence, patient satisfaction and reported symptom improvement. 42,43 A recent meta-analysis found that patients who trust their providers tend to report more beneficial health behaviours, less symptoms and higher quality of life. 44 Patients also wished providers to pay attention to their feelings and react with empathy. Empathy has been linked to better health outcome as well, 2,45,46 and providers' attention was a main focus for patients asked to share their views without being restricted by a questionnaire. 4

Patients' rights
Effective communication between patients and providers may be seen as a patient's right. 47 This includes the right to knowledge and the right to make choice. 48 Informants wanted to know the diagnosis and the treatment plan. They believed this would help them participate actively in their care and better take control of their health as also proven by a recent study. 49 Patients' perceptions indicate important differences and inconsistency in providers' communication styles. This inconsistency may pose challenges to effective health communication. It may reflect lack of standardised and effective communication training programmes. This issue triggered some patients to express the need to choose a preferred provider. They would like to interact with the health care providers they are most comfortable with. This is an important aspect of the continuity of care, fundamental in PHC, described as relational continuity. 43,50 While ideal, this may be challenging in resource-constrained settings, perhaps unfeasible. Contextual factors such as a high turnover and scarcity of providers, as well as inadequate organisation of care, often found in resource-constrained settings, are examples of systemic issues that may make it difficult to honour such requests. However, patients' rights call for providers' responsibilities. Providers must be aware of these key requests among patients and seek to accommodate them to the extent possible without generating harm.

Patient involvement
Prima facie, most patients were hesitant of being involved in their own care but would change their view after we shared examples of the value of patient involvement. Our study shows that patients, particularly with limited literacy, may have strong though modifiable preferences for paternalistic health care providers. 51 Perceived power imbalances may inhibit patients' involvement in their own care, for instance patients hiding their knowledge to avoid problems, and a preference for passivity and reluctance to actively engage with health care providers. Our study indicates that discussing such behaviour may be a means to challenge it, confirmed in other studies. 52,53,54 This requires a favourable environment for collaboration in which the provider actively invites the patient to participate in the care, while explaining the value of such initiative to gain the full collaboration of the patient. Responding negatively to patients' need for more participation may maintain patient passivity, hamper patient autonomy and contribute to unimproved outcomes and impaired patient safety. 51

Power issues
Patients expressed the asymmetric, hierarchical relationship with providers as naturally explained by the difference in knowledge. 55 The metaphor of providers as parents or gods captured this well. Such metaphors are not new and often related to paternalistic practice of care. 56 Rwanda, like most African countries, is often described as a traditional authoritarian society with strongly hierarchical structures, and a history of strict rules with legacies from the colonial period. 57,58 One could speculate if this oriented the health system towards more provider-centred care models. Limited literacy may amplify the power differential. 59 This may have patterned the interaction between patients and providers over time and helped condition the patient perceptions uncovered. 3 Informants perceived the power imbalance as legitimate and necessary, contrasting with common descriptions of the power imbalance reported in the literature, often presented as unwanted. 60 We found indications of strong dependency among our informants on health care providers, often bordering on blind trust, much like the child without power of choice. 56 Particularly among patients with limited literacy, trust appeared as a general trust in modern medicine as an ever-present source of cure and solutions to health problems. They also tended to be less critical towards provided health care, often with no interest to understand their own medical condition, left in the hands of the provider. This attitude may increase the risk of providers overlooking symptoms, not getting or sharing all necessary information, or suggesting an inappropriate plan. 61 Providers should be aware of this issue and would probably benefit from using plain language and asking plenty of openended questions, requiring patients to reflect, such as how do you feel about this plan? Patients may also benefit from effective teach-back techniques. 62,63 Providers need to be aware of the existing power imbalance in their interaction with patients and find ways to reduce its influence on the interaction and relationship. 55 Feedback was mentioned as a possible way to voice patients' preferences and concerns. Feedback may help providers to improve their communication skills, as long as they are prepared and ready to receive feedback. This might be challenging if the paternalistic paradigm dominates.

Implications of this study
This study has provided preliminary insights into PPC in Rwanda. The findings of our study demonstrate once again the importance of valuing patients' point of view, particularly in a context of limited literacy and health literacy. This calls for more consideration of patients' perspectives in future explorations of PPC in Rwanda and beyond.
This study also suggests that limited literacy may be linked to a number of health literacy problems and warrants further study on how best to approach this vulnerable group.
The study shows the need for well-designed communication training to improve communication knowledge, skills and attitude to ensure the genuinely caring behaviour expected by patients.
Also, a secure environment should be created for patients to freely express their preferences, feedback and complaints to reach the level of communication excellence they request. To allow patients to give feedback without fear, one approach may be to formalise the feedback, for instance, through an anonymised assessment tool.

Strengths and limitations
The strategic inclusion of patients with limited literacy gave voice to a marginalised portion of the population, often excluded in research, and allowed access to richer data. 8 Our findings are drawn from a relatively small sample of 15 patients. While further interviews might have generated more ideas, it was our interpretation that saturation was reached on matters related to our research questions.
This study focused on patients attending outpatient clinics at health centres, and these patients tended to share experiences within the health centre and beyond. We, therefore, believe findings are potentially transferable to PPC in other health services and other care providers in Rwanda.
Limited literacy may reduce informants' ability to actively engage in the type of open discussion necessary to generate rich qualitative data. This may have limited the quantity and quality of information generated from interviews with patients with limited literacy. We tried to accommodate this by simplifying the question guide and ensuring comfort and privacy during the interviews that were conducted far from the health facilities. Further studies may benefit from exploring the use of pictures and storytelling as means to generate information with this patient group.

Conclusion
To our knowledge this is the first study exploring patients' perceptions of their communication with PHC providers in OPD in Rwanda. The participants in this study were able to demonstrate shared and consistent views of the perceived factors that influence effective PPC. Patients' preferences reflect components of the patient-centred care model and include a need for meeting caring, empathic providers, who can inform them about their health and involve them in health care. Patients with limited literacy appeared to rely highly on health care providers for making decisions about and managing their health care on their behalf. They were, however, ready to be more involved, suggesting a potential for improved patient involvement even for patients with paternalistic care preferences. These findings should be considered while planning services, training health care providers, measuring quality and developing health education strategies to empower patients or simply while discussing issues around PPC. The generated knowledge may help advance PPC and health communication in Rwanda and in sub-Saharan Africa.