What price a welcome? Understanding structure­agency in the delivery of respectful midwifery care in Uganda

Purpose – The purpose of this paper is to present the findings of research on mothers and midwives’ understanding of the concept of respectful care in the Ugandan public health settings. It focusses on one aspect of respect; namely communication that is perhaps least resource-dependent. The research found endemic levels of disrespect and tries to understand the reasons behind these organisational cultures and the role that governance could play in improving respect. 
Design/methodology/approach – The study involved a combination of in-depth qualitative interviews with mothers and midwives together with focus groups with a cohort of midwives registered for a degree. 
Findings – The findings highlight an alarming level of verbal abuse and poor communication that both deter women from attending public health facilities and, when they have to attend, reduces their willingness to disclose information about their health status. Respect is a major factor reducing the engagement of those women unable to afford private care, with health facilities in Uganda. 
Research limitations/implications – Access to quality care provided by skilled birth attendants (midwives) is known to be the major factor preventing improvements in maternal mortality and morbidity in low income settings. Although communication lies at the agency end of the structure-agency continuum, important aspects of governance contribute to high levels of disrespect. 
Originality/value – Whilst there is a lot of research on the concept of respectful care in high income settings applying this to the care environment in low resource settings is highly problematic. The findings presented here generate a more contextualised analysis generating important new insights which we hope will improve the quality of care in Uganda health facilities. 
Keywords: Leadership, Governance structures, Patient perspectives, Communication, Midwifery, Determinants of health Paper type Research paper


Introduction
Maternal mortality in Uganda is very high and shows remarkable resistance to change. Recent research suggests that promoting access to quality, midwifery lead, care has greatest potential to improve maternal health in Low and Middle Income Settings (LMICs). Furthermore, the key to encouraging access lies in providing respectful care. But, what does respectful care mean to mothers in Uganda? This is one of two linked papers presenting the initial findings of research in Kabarole District, Uganda. The study is funded by the Wellbeing of Women 5 and operationalised through a British Charity, Knowledge for Change 6 . The first paper focuses on one dimension of respect; namely communication. Communication arguably falls at the 'agency' end of the 'structure-agency' continuum to the extent that, in principle at least, respectful communication is not resource-dependent. In trying to understand the reasons behind high levels of verbal abuse in public facilities the paper considers the role that governance could play in shaping health worker behaviour and improving communication.

Behaviour change and Governance
The dynamics of health worker behaviour forms a recurrent theme in research across all health systems. Indeed much of the early research on behaviour change took place in high resource settings grappling to understand why training and education failed to deliver targeted and well-resourced changes in the behaviour of health workers and patients. More recent work has applied theories of behaviour change developed in these settings to an understanding of the impact of aid interventions (and professional voluntarism in particular) in LMICs (Ackers and Ackers-Johnson, 2016). Although this work comes from very different disciplinary and theoretical positions there is a broad consensus that individual behaviour is constrained, to varying extents, by the environment within which it takes place. Put simply, individual 'agency' or choice is structured by context. This does not imply that health workers have no choice about how they behave; quite the opposite. Whilst some theories emphasise local resource factors (immediate access to equipment, medicines or workload), others emphasise organisational dynamics (management of health facilities) and still others, wider socio-legal, political and cultural landscapes shaping reproductive rights and social inequalities. Governance features at all these levels be they local, organisational or societal combining to create complex webs of contingencies which together shape health worker behaviour and intervention success. Not only is behaviour change shaped by these external opportunity structures (Michie et al, 2011) it also varies by the type of behaviour we are seeking to achieve (the change objectives). An individual may have more latitude (agency) in relation to some aspects of their work than others. Tasks as simple as hand washing or as complex as performing a caesarean section may be confounded by the lack of basic resources such as water and soap or access to more complex and specialist resources needed to functionalise an operating theatre. But why, when soap, water, hand towels and training are provided do health workers fail to comply with basic hand hygiene protocols? Why, when theatres are new and staff are employed and trained, do they fail to perform surgery? And why, when a task as simple as welcoming a mother in a kind and friendly way when she arrives at a health facility, does that welcome 2 fail to take place? Or, put another way, why is it that some health workers are able, in similar contexts, to behave with optimal professionalism whilst others fail at the first hurdle? The Importance of Respectful Care According to published data Uganda has one of the highest levels of maternal mortality in the world. The Ugandan Ministry of Health's Strategic Plan suggests that little, if any, progress has been made to reduce maternal mortality (MOH, 2010:43). A United Nations report on the Millenium Development Goals describes Uganda's progress as 'stagnant' (UNDP, 2013: iii). Figures on maternal mortality in Uganda vary considerably depending on the source. The World Health Organisation suggests maternal mortality ratios (MMRs) of 404 per 100,000 live births (WHO, 2012). 7 However, the benchmarking exercise undertaken as part of the Sustainable Volunteering Project (McKay and Ackers, 2013) indicated levels of over 1000 (more than double reported levels). 8 The quality of care at government health facilities is critical to encouraging mothers to access facilities and ensuring that they receive timely and effective support. Research increasingly points to the role that respect plays in this process (Renfrew, 2014: Sharma et al, 2015Hulton et al, 2000;Srivastava, 2015). In recent years, the relationship between quality care and maternal outcomes has become more apparent with the World Health Organisation issuing a statement on the prevention and elimination of disrespect and abuse during facility-based care (WHO, 2015). Disrespect and abuse in facility based care remains difficult to quantify and poorly defined but includes; physical abuse, non-consented care, nonconfidential care, non-dignified care, discrimination, abandonment of care and detention in facilities (Bowser and Hill, 2010). Respectful, quality, maternity care is a fundamental human right that encompasses the right of every childbearing woman to receive care where there is choice, companionship, dignity, privacy and safety (UN, 1948, UN 1993, WHO, 1999, White Ribbon Alliance, 2011. Fear of disrespect and abuse by health care workers can deter women from attending services on time, attending services enough or attending at all (Atekyereza, 2014, Kruk, 2009, Hulton et al 2007. It is unsurprising then that disrespectful and abusive care is linked to multiple health outcomes, effecting maternal morbidity and mortality worldwide (Amroussia et al, 2017). Furthermore, women's experiences of their care can have lasting psychological and emotional effects, impacting their utilisation of health services in the future. Increases in population coverage, antenatal care attendance or rates of deliveries with skilled birth attendants, is not synonymous with care that is respectful or free from abuse. Maternal mortality will only be reduced once essential interventions, structural improvements and access to resources goes hand in hand with maternity care that is respectful and empowering for both women and their families (Warren et al, 2013).

The Study and Methods
The WeCare study builds on 10 years' ethnographic action-research in Uganda. This research has confirmed out belief that, in comparative research, 'context is everything ' (ref). In 2017, the authors conducted 64 in-depth interviews and 3 focus groups with mothers and midwives working in public health facilities in Kabarole District, Uganda. The objective has been to understand what respectful care means to mothers and midwives in Uganda and what hinders them from providing it. Ethical approval for the study was obtained from the University of Salford and Mountains of the Moon University, Uganda.

Conceptualising Respectful care in Ugandan public services
The interviews enabled us to gauge how women perceived respect and what priority they attached to different elements. Very many mothers spontaneously linked respect to communication and spoke openly of their experiences or the experiences of their family and friends. Verbal abuse was expected and normalised. Most mothers also immediately connected notions of respect with the quality of care they received referring to neglect (especially in the regional referral hospital), extensive delays and 7 The newly published Sustainable Development Goals (that replace the MDGs) cite a target MMR of below 70 per 100,000 births (UN, 2015:13) 8 These figures only capture recorded deaths in the facility and thus miss cases where mothers die in the community or where record are unavailable. abandonment often connected to endemic corruption. There was a strong sense amongst mothers that patients should be treated according to their clinical needs and not the ability to pay bribes. They also referred to the lack of access to medication, basic consumables, equipment and the quality and cleanliness of facilities. Together these issues emerged as the most important elements of respectful care. This paper focuses on the issue of communication and, the following paper, on quality of care. Quality of care issues are very explicitly related to governance, fall largely outwith the control of an individual midwife and lie towards the 'structure' end of the structure-agency continuum. But what about communication?
It is absolutely clear from the interviews that basic communication -or lack of it -plays a very critical role in shaping women's attitudes towards using public services particularly in lower level health facilities and in relation to non-emergency primary care. If women feel unwelcome or 'under-looked,' 9 either as result of a health worker's body language, facial expressions or verbal abuse they will shun services preferring to manage without any support or to access other forms of support. Depending on the mother's educational, residential and financial status (and her husband's attitude) these may include unqualified traditional birth attendants who are effectively self-employed and charge a fee or a range of private and not-for-profit facilities. Where women make this decision not to access public facilities this will have multiplier effects beyond delivery (birth) extending to all aspects of preventive care for themselves and their families. Once in facilities the way women are spoken to plays a major role in shaping relationships with health workers; if communication is poor mothers make it clear that they will withhold vital information about their health. On the other hand, if communication is positive, women are not only more likely to access that facility themselves and talk more openly and honestly about their health and their concerns; they also quickly transmit this message to family, friends and communities. Word of mouth is by far the most important mode of communication about health facilities in this environment with often quite immediate impacts.

What Price a Welcome? The Importance of Communication
Unsurprisingly, when asked to talk about their experiences of care most mothers identified forms of abuse. Although published research continues to identify incidences of physical abuse (including slapping or pinching of mothers) our interviews 10 found little direct evidence of this behaviour. Whilst some respondents spoke of the experiences of their friends or witnessing such practices during their own training none of the respondents reported experiencing physical abuse. However, 20 of the 52 mothers interviewed referred to some experience of verbal abuse. In most cases they spoke of midwives shouting or 'barking' at them or speaking in a 'rude' manner. The following examples are typical: When you ask a question where you have not understood they just shout at you (M01) They said, 'you're not the only one to take care of' in a rude way (M02)

Midwives from [the public] hospital used to shout at me. I delivered my first baby at night and I would make a lot of noise so that's why nurses shouted at me but I could tell them, 'it's not me it's the pain' (M36) The midwives from [public] hospital talk in a rude way abusing pregnant mothers in labour. Whenever I called a midwife to examine me she would say, 'I'm tired of your noise' (M48)
Examples of respectful care were also cited with an emphasis on the importance of being 'welcomed.' Welcomes in Uganda are very important. It should come as no surprise then that mothers are highly influenced by their first impressions of a facility and the welcome they receive. Many mothers responded by simply stating, 'Midwives should behave well by welcoming patients on arrival' (M4) Others used words like 'calm' and 'friendly' to capture their understanding of respect: The nurses used to talk calmly. It was not me alone but to all mothers because I used to hear them talking to other mothers from other rooms (M2).
The following mother explains how the initial welcome at her first ante natal visit encouraged her to continue to access that facility for the duration of her care. It is interesting to note the sense of surprise at the quality of her care which she likens to a private facility and also to the question about whether she knew staff there. This reflects the interviewer's awareness of (quite common) incidences of patronage where staff show greater respect to people they know and/or other health workers: The midwives treated me in a good way from the time I started attending antenatal visits up to the time of delivery. They used to talk in a friendly manner […] I was treated as if I was in a private hospital. Q: Did you know any midwives there? I didn't know anybody from that facility and it was my first time to be in that facility (M45) In the following cases, the mothers explain how reports of respectful care encouraged them to access specific facilities: The midwives from this facility are good. They welcome mothers who have come for antenatal, treat pregnant mothers and babies well. Even mothers who have delivered from here talk good about this facility and that's what forced me to come (M42).
There was a friend of mine who was pregnant and requested me to escort her during the time of delivery. She delivered from [here]. The way staff treated her attracted me to deliver from there (M50).
In the final case reported here the respondent is delighted at the welcome and also being congratulated on her delivery: When the labour pains started, I went to deliver. The midwife whom I found there was so good to me. She used to encourage me that I'm going to deliver normally. She delivered me without abusing me. What made me happy was after delivering the baby, she congratulated me (M32).

The Multiplier Effects of Disrespectful Communication
Disrespectful communication discourages mothers from attending facilities and from disclosing important information. Asked if she felt able to ask questions of the midwife a very typical response was, 'I was not free with them because of their facial expression' (M30) or, 'I was not free with them because of the way they treated me' (M45): The interviews reported some very positive experiences, predominantly at midwifery led Heath Centres, showing how good communication improves a willingness to share information:

I have been free with them even I'm planning to ask them the cause of abortions [miscarriages] I'm getting. The midwives have not yet abused me and everything is going well. (M15)
The phrasing of the last comment is interesting as the mother is clearly surprised by the respect she has been shown and anticipates disrespect in future. The consequence of disrespect in terms of patient trust and information-sharing were commonly referred to by midwife respondents: When a mother is communicated to well she can disclose her problems and a midwife will be able to handle. When you communicate to her in a wrong way she will hide yet she will get problems, maybe she has come with a problem, she wants to be helped she can't open up, she will go back and get issues or complications (MW10) One of the Focus Groups with midwives identified a very tangible example of how disrespect and poor communication can result in very serious conditions being concealed from midwives: P1: Good communication helps the mother to open up. If she is having a serious condition or maybe in the private parts when you are not a good communicator the mother may hide that information. P2: That is actually true, because I remember a scenario when a mother at the exit of the clinic was asking me about what she could do about a problem. She was experiencing some growth which was nearly covering the vaginal orifice and that growth was not established in the clinic by the midwife who was attending to the mother. So probably there was poor communication and poor communication can be brought about by first impressions, probably the way the mother was welcomed was not good enough, the verbal and non-verbal communication skills, probably the non-verbal showing a very bad facial expression and the mother could not give out all the full details of her health status. (Focus Group STM4) As indicated above a major concern is that disrespectful care discourages mothers from using public facilities. In the following case the mother informs us that the midwife told her to go to the private sector if she wanted respect: The nurse who on duty told us that if we want care, we should go to a private hospital (M27) The following case, where a mother decides to leave a public facility and access care at a local not-forprofit facility is common. Unfortunately this type of facility still requires fees to be paid and many women are unable to exercise this choice: The (midwives) used to act rudely like shouting at us. When the labour pains started, I went to (public) hospital to deliver but the way I heard the nurses talking to the mothers scared me so much I had to go to a private hospital. (M40)

Bad People or Bad Behaviour?
The section above has evidenced the co-existence of disrespect and respect within the same public health system. The interviews identified some marked differences between facilities. Kagote Health Centre which has been supported by Knowledge For Change for the past 4 years emerges as an example of best practice. There were no negative comments made about the quality of communication and care received from this unit and most of the positive feedback was linked to that facility. On the other hand, the Regional Referral (Buhinga) Hospital received the most widespread criticism suggesting an almost endemic culture of disrespect. However, mothers did point to examples of good practice and respect even within this environment. In several cases respondents spoke of the positive attitudes of student midwives at Buhinga. Several mothers also spoke more positively of recent experiences of giving birth in Buhinga in comparison to previous deliveries: The midwives (at Buhinga) have been talking well and I have been free to ask them questions (M20).
The student midwives were so good to me. When I went to deliver my last child, I found midwives have changed their behaviours compared to my other pregnancies (M33) Mothers were also aware that different staff in the same facilities behaved very differently. And some took great care to avoid meeting certain individuals: Every time when we go to the facility we meet different midwives. Some respond well while others respond rudely (M14) Midwives are not the same; others are rude and others talk well (M21) In the following cases mothers reported improvements in their treatment at Buhinga hospital but in both cases they had taken husbands and attendants with them, as witnesses or advocates, and clearly felt that this improved their treatment: When I went to deliver my first child, I went at night but the nurse whom I found on duty was very arrogant. When I called her, she told me go find a bed and sleep. I will examine you later. It reached morning without her examining me. When I went with my husband, the staff were talking to me well unlike when he was not there (M29) The nurses would shout at me during antenatal visits. I was not free with them at all because of the way they talked. I went to Buhinga hospital for delivery but what surprised me the nurses shouted at me even before examining me because I didn't have a polythene paper to put on the delivery bed. And I knew that they are within the mama kit 11 . She was not abusing me alone but every pregnant mother and I decided to go to a private hospital because of my first impression. The third time I went back to Buhinga hospital with my husband and four people to take care of me thinking the midwives will treat me well. The midwife didn't abuse me the way the others had but the bad thing of her she used to sell gloves (M41) Sadly, the mother experienced corruption which may have influenced the quality of her care. Many examples suggest that women in Uganda expect to pay for respect. Put another way, they do not expect to be respected when receiving free public services unless they pay a fee (bribe) to the midwife: She didn't abuse or shout at me because I had given her the money and when her duty was over, she handed me to another midwife to keep checking on me (M12).
Forms of patronage or clientelism were common. Midwives who had experienced giving birth in public facilities were generally happy with their care knowing full well that they had been given preferential treatment by their colleagues:

Once you are a health worker the care is really respectful. If you don't say….you will be cared for like other mothers (MW1)
Others who had friends or connections with midwives were treated with greater respect:

Midwives in Mubende hospital knew my aunt so they feared to abuse me but they abused others (M13)
In the next case the mother was disappointed that the staff had 'forgotten' these connections during her second delivery: I was disappointed because the midwives had forgotten me. By the time I went for delivery, my sister was transferred to another hospital so the care I received was not good. The midwives didn't encourage me and would shout at me (M45)

Understanding Behaviour
The findings presented above reveal a culture of disrespect that permeates many public facilities. On the other hand, clear differences emerge both between and within different facilities and in their relationships with different patients or even the same patient at different times. Certainly there was a very clear sense amongst the midwifery respondents of what respectful care looks like, what it means to local women and why it is important. So, why do some health workers in some facilities respond differently to the same stimuli? How much of this is down to individual agency or choice and how much is a reflection of structural factors? Communication, as a key dimension of respect, is not simply determined in any direct way by resource constraints. But neither is it accurate or fair to assert that all health workers have the same opportunity to exercise excellent communication skills. Just as the familiar Western adage, 'it costs nothing to keep clean' falls down in many low resource settings, so too does the idea that being kind and calm is ubiquitous. The interviews with midwives identified some reasons for disrespectful communication.
These included reference to shortages in human resources (staffing); the effects of low and unreliable remuneration creating pressure to earn money through fees (bribes); possible gender differences in approach, the disrespect shown by patients towards midwives and a lack of training. Others suggested it was a reflection of personality.

Workload and Communication
One of the focus groups with midwives triggered a range of 'justifications' for poor communication. In some ways participants grasped the opportunity to share pressing concerns about their status, pay and workload emphasising the structuring determinants of individual agency. LMICS are undoubtedly associated with far lower staff-patient ratios. However these figures mask marked differences between facilities and locations. Inability to recruit staff compounded by absenteeism means that real staffing levels are often far below published figures. Compounded by dysfunctional referral systems and selfreferrals this leads to overwhelming congestion in hospitals. The situation in lower level facilities varies considerably. Facilities with a reputation for high quality care will experience times of very high workload and individual midwives will struggle to communicate as well as they would like to. Some of the midwives talked of the pressures of work and the impact on behaviour: Maybe because they are tired and it's overwhelming or maybe due to their personality …. the way they speak to these mothers, sometimes rude, being judgemental but I think it's because the work is overwhelming (MW2).
They shouldn't but some do shout. You may find the midwife has been on duty from morning to evening. The woman comes in the evening and she is very tired and the mother is uncooperative and so the midwife starts barking at the patient. She translates all her stress to this mother (MW4).
In the following case the respondent immediately connects time-keeping to respect: I make sure that I come on duty early.

If a mother comes she should find me around. That is very important. Am I on duty? If I am supposed to be on duty have I come on time? Now if a mother comes, what is my attitude towards her? Is the mother scared of me? (MW1)
On the other hand, the endemic practice of trying to 'clear the lines' as quickly as possible so that staff can leave work to go home or engage in other jobs (moonlighting) generates congestion and workload that would not exist if they were working a full shift. It is very important to point out here that all health workers engage in this behaviour and as a general rule midwives work longer hours than other health workers many of whom try to leave by lunchtime. Mothers were clearly aware of the ways in which staff 'managed' patients so that they could spend as little time at work as possible. Several mothers spoke of being shouted at either because they arrived at the clinic late, arrived late in labour or had not attended ante-natal clinics at the facility. Midwives were acutely aware of how verbal abuse is used as a way of managing patients: The particular midwife they say is bad. She does not even listen to why she has come late or why she delivered on the way. She starts there and then shouting at the mother (MW5) There are some midwives who are harsh that, 'you didn't attend antenatal from here why have you come to deliver from here?'' So when you talk to them you say, No a mother is free to attend antenatal from anywhere. She can deliver from anywhere'. You tell them, 'last time we had a customer care workshop, we were trained about it', so it changes midwives' attitudes (MW10).
Whilst workload is most evidently a problem in well-functioning lower level facilities and hospital settings it is important to note that in many lower level facilities midwives have very low patient volumes -often with no or very few deliveries at all. And any mothers who DO present in labour will have to move on to the hospital as no staff are present or will be referred (often inappropriately) to the hospital. This is most common in the afternoons and evening when many facilities are effectively closed and employed staff absent themselves from work. In reality mothers are fully aware of these practices and simply by-pass facilities altogether. By way of example, it would not be uncommon for a health centre III facility to employ between 3 and 5 midwives but deliver fewer than 8 babies a month. Kagote Health Centre had conducted no deliveries for 16 years when K4C became involved despite the presence of midwives. The point is not to suggest that some facilities are not desperately understaffed; they are. But there are major problems in the management of those staff who are employed which contribute to perceived workload and manifest themselves in verbal abuse.

Organised in the Pocket 12 ?
Another major (and legitimate) concern of midwife respondents was the issue of pay both in terms of poor levels of pay, very limited opportunities for pay progression and poor administration of pay leaving staff often unpaid for several months. Subsistence wages create a compelling rationale for corruption. This manifests itself in disrespect as health workers conduct an intuitive 'means-test' assessing the potential to extract fees and shun those who look poor: Some midwives have a poor attitude to mothers… others need money. When she sees that this mother is poor she keeps shouting at the mother which doesn't bring respect (MW6).
Mothers referred to what they viewed as inequality or discrimination in the ways that health workers prioritised and managed patients suggesting that staff shunned mothers who were unclean or poorly dressed. In reality this may not be because they are unclean or badly dressed but rather that staff realise they will not be able to extract a bribe, or receive a 'blessing'. Whilst we have made a distinction between forms of behaviour that are not resource-dependent (such as being friendly and welcoming), in practice it is impossible to completely disentangle physical resource constraints from perceptions about attitudes. Critical shortages in some consumables or medications (arising from a combination of poor investment and management) combined with endemic corruption lead to situations where patients misinterpret the behaviour. The following midwife explains that, on some occasions, endemic resource challenges effecting workload and supplies of medicine and consumables, may be interpreted (wrongly) by patients as indicating a lack of respect: Sometimes I hear mothers talk. They find where things are out of stock like the polythene papers they deliver on, then they are told to buy. So she feels this midwife is rude and yet she is telling her the right thing. So she thinks; 'This midwife is rude, she asks us for dresses, for polythene papers and yet we don't have money for that'. Instead of seeing the issue at hand they say they will not deliver from there; 'I will deliver at home because I don't have a dress or gloves, so if I go to hospital the midwives will abuse me'. The midwife is not abusing but telling the truth (MW5).
This indicates the constraining impact of contextual issues on the ability of midwives to be seen to show respect. Having said that, other midwives argue that it is ultimately a personal decision -that agency is more prominent than constraint: It is more personal; some people have a good character, some don't. At times you see someone who just tells this mother, 'you just enter there'. You don't even ask, you don't greet, you take someone as if she is useless; as if someone is begging for your service and also shouting at the mothers (MW11).
It depends on their personality, it's their way of behaviour, it's how they are, it's how they talk. Some are very compassionate (MW4) One male respondent suggested that men showed greater respect than female midwives perhaps because women, many of whom had been mothers themselves, normalised pain: Q: I don't see many men working in midwifery in Uganda. How do mothers respond to you? Actually mothers don't have a problem with it and most of them they like it. Maybe because of the care; we don't shout at them, we talk to them in a good way, that's why some of them like men -most of them actually. Men are less likely to shout because they give a lot of respect to ladies. They look and think they have a lot of pain. Because [female midwives] have suffered the same, they don't take it so serious (MW11)

Understanding Disrespect
To the extent that communication is closely linked to culture -organisational culture to a greater extent than ethnicity -there is a possibility, as the behavioural science theorists indicate, that training incorporating behaviour change methods, could improve communication (Byrne-Davis et al, 2016). Put differently, that investing in capabilities (health worker education) could trigger change. However, we would contend that the majority of Ugandan health workers and citizens understand quite clearly what it means to show respect and to welcome people. Ugandan society and its cultures value respect greatly and routinely extend a very warm welcome. When Ugandan health workers show a lack of respect they are fully aware of that situation. And Ugandan mothers are also very aware that disrespect is associated with a kind of 'othering' or 'under-looking'; of viewing some mothers as less deserving or potentially less 'rewarding' than others. We would certainly agree with Cane et al that, 'for most health behaviours [..] knowledge is not an important source of variance ' (2012:15). Training in isolation will not generate behaviour change. And, further that lack of 'opportunities' (resources) will undoubtedly block skills utilisation. In the case of communication, although as we have noted, resources Do matter, especially human resources, this is the least resource-contingent aspect of 'respect' and the area where individual agency is least constrained.

Summary
This paper has focused on one dimension of respect; namely communication. This is an aspect very much at the forefront of those mothers' minds who depend on public services and these are the poorest women in Uganda. We have also chosen to explore communication because we feel it sits at the agency end of the structure-agency continuum. It is the dimension of respect that Ugandan midwives have the greatest autonomy to deliver on. The results suggest that a culture of disrespect pervades public services. There are some marked but rare exceptions to this. The findings suggest to us that capabilities (individual know-how) are not the problem. It is organisational cultures, not national cultures or lack of education that shapes behaviour. Ugandan health workers know what respectful communication means and how to deliver it. The failure to translate that knowledge into action and behave respectfully is a function of wider contextual factors affecting both opportunities and motivations. At the very personal level some individuals are simply behaving badly. There is no system of accountability; of reward or discipline in most Ugandan health facilities to identify these individuals and manage them effectively. Women -and patients -in Uganda have very few rights. Or, where rights exist on paper, there is no investment in implementing those rights. Put simply, there is no genuine interest at national level in how women are treated in public health facilities. At best (and very rarely) bad behaviour may result in staff being rotated to another facility. Implementing rights and training staff in the consequences of breaching those rights is essential. This is not an issue of national culture and we must take care not to essentialise these behaviours. Customer care in private facilities in Uganda matches that in the best facilities in the UK and the same staff often work in both sectors. The culture of disrespect reflects a kind of 'othering' associated with residualised health systems where those people managing and delivering public services cannot imagine a situation where they would be at the receiving end. They either use the private sector (in most cases) or, for less well paid cadres of staff, benefit from forms of patronage within public facilities. The WeCare project was stimulated by a UK intervention known as 'Whose Shoes?' This intervention was focused on encouraging health workers to put themselves in the shoes of service users. That requires us to imagine a reality in which we may indeed be in those shoes, giving birth in a public health facility. In a universal health care system such as the UK's National Health Service, the majority of us (including the authors) continue to rely on that service for our own healthcare. Imagining wearing those shoes is part of our lived reality; our citizenship. In important ways the behaviour of individual health workers reflects wider societal and political processes and requires action to reduce systemic inequalities. But not all health workers are simply behaving badly, 'othering' the recipients of public services. Much of what we perceive as discrimination or bad behaviour is intrinsically linked to personal income; it is strategic. It is not so much that a poor woman is 'underlooked' but that the health worker knows immediately that she will not be in a position to extract a fee from her; she is undertaking a means-test. Most midwife respondents drew immediate attention to the impoverishment of their profession and the inability to survive on subsistence wages. This is a fact. Current levels of pay cannot support basic subsistence. Moonlighting (doing a second or third job) is both technically illegal but endemic in practice; it is culturally expected. Most health workers spend as little time as possible in the public facility that employs them and manage their patients to facilitate that. They aim to arrive late and leave by lunchtime. Midwives are under huge pressure to deliver a 24/7 service and, if this breaks down, mothers will by-pass that facility. Low pay and a total lack of accountability and management in most facilities facilitates a workload management 'system' that compounds congestion at referral hospitals. Disrespectful communication is an intrinsic and strategic component of this system providing health workers the space to attend to their other paid work. Workload in facilities associated with respect escalates and the staff working there are impoverished as a result with less potential to augment their meagre salaries. This is one of the reasons that health workers can resent foreign intervention, especially when, as in all K4C interventions, co-presence is required and labour substitution and salary augmentation banned (Ackers and Ackers-Johnson, 2016;Osman, 2017). The solution to this demands both carrot and stick. In a system where over 60% of remunerated staff can be absent at any one time 13 it is challenging to communicate respectfully. Public sector pay and career progression must be addressed to begin to change this system. Once that is addressed staff need managing effectively, educating about the impacts of professional misconduct and disciplining accordingly. At the heart of this the daily reality of endemic corruption needs tackling head on so that mothers no longer expect to pay for respect. In a corruption free environment trust can begin to flourish. It is clear from the interviews that mothers understand the importance of triaging patients and are prepared to wait when urgent cases need attention. They also understand the pressures on public resources resulting in stock-outs of medicines and consumables. Corruption creates an environment of suspicion amongst patients that further damages the potential for respectful relationships. Extending a warm and professional welcome to mothers is essential to building relationships between public services and local communities. It is an essential precondition to any improvement in maternal health in Uganda. This paper concurs with recent evidence that women seek care givers that are skilled, knowledge and compassionate with an emphasis on 'good quality clinical care and improved communication' (Renfrew et al, 2014). Even in under-funded, badly managed, health systems health professionals have the agency and responsibility to behave respectfully. We applaud those who, in the face of such adversity, do exactly that. According to Pyone et al, 'governance is a practice, dependant on arrangements set at political or national level, but which needs to be operationalised by individuals at lower levels in the health system' (2016). The responsibility ultimately rests with the State to create an environment respectful of women's rights and willing to take action.