Incremental infrastructure and the uncertainty of care

In recent years, Uganda’s boda bodas (Doherty Reference Doherty2017; Reference Doherty2022), as well as similar motorcycle transport in other African countries, such as Nigeria (Smith Reference Smith2022) and Ethiopia (Mains Reference Mains2019), have received increasing scholarly attention. Motorbikes have transformed urban and rural Africa, making getting from place to place faster, but also, especially from the point of view of city-dwellers, more chaotic and dangerous (Wanume et al. Reference Wanume, Nduhura, Mugerwa, Bagambe and Ninsiima2019). In rural Uganda – as in eastern Luuka district, where I studied maternal health – boda bodas are rarely criticized or viewed as unruly and dangerous as they often are in the capital, Kampala. Boda bodas provided much cheaper and easier transportation than was otherwise available in the district, a region where the use of cars was relatively rare, and where the conditions of roads often made movement by car or minibus more difficult than it was on a more nimble, smaller motorbike. Boda bodas could navigate rutted, narrow paths and carefully bypass the muddy potholes that plagued the region’s dirt paths and murram roads during the rainy season. They were used for everything from transporting children to school to moving supplies from town to a shopkeeper’s remote store. They were the primary way in which sick people were transported to clinics, even as some clinics, including the government’s level IV district referral clinic, kept ambulances parked near their gates. Ambulances were prohibitively expensive to fill with fuel and struggled on poorly maintained roads. Boda bodas could move people, and people could, mostly, afford to keep them moving.

If boda bodas were ubiquitous and essential, they also represented a particular approach to infrastructure expansion in Africa, one that has been defined less by top-down state-led programmes and more by the improvisation and individual will of residents and by investment by a patchwork of non-governmental and for-profit actors. Daniel Mains, writing about Bajaj motorbikes in Ethiopia, describes the proliferation of such bikes as a more flexible, entrepreneurial solution to transport that stands in contrast to major state projects such as tarmac road expansion and hydroelectric dams (Mains Reference Mains2019). Nonetheless, boda bodas, like other forms of ‘flexible’ infrastructure, are deeply intertwined in mechanisms of and for development and governance in Luuka. Motorbikes enter the rural community largely through community lending schemes, financed through community banks and NGOs that see boda bodas as good financial investments that have the dual benefit of profit making in tandem with developing a community’s infrastructure, potentially improving access to healthcare and other services.

The popularity of this approach to infrastructure has emerged hand in hand with a broader emphasis within global development structures on community-led development and community-based NGOs, which have been lauded as mechanisms for communities to advocate for infrastructure investment and respond in targeted ways to particular needs, though largely within a context of decreased state spending and an increasing emphasis on privatization (McKay Reference McKay2020; Pfeiffer and Chapman Reference Pfeiffer and Chapman2010). These grassroots initiatives potentially provide the means for communities to address the deficits of a global development industry that has been criticized for a hegemonic, one-size-fits-all approach to infrastructure expansion, and which has largely failed to fulfil its promise of infrastructure-for-all (Silver Reference Silver2023: 13).

The clinic transportation project was indicative of such a shift in infrastructure development. It had been conceived by a community NGO in Luuka, a locally founded and run organization devoted to improving the health of the surrounding community. A project manager at the NGO had applied and successfully raised a few thousand dollars from a foreign foundation to test the transport project, enabling the purchase of a new boda boda for the clinic, as well as supporting clinic midwives to do outreach visits to pregnant women. His proposal had emphasized the problem of transportation to health clinics as a key barrier to care for pregnant women, and he had proposed the boda boda programme to evaluate whether free transport would increase clinic-based births.

In Luuka, fewer than half of women give birth with the assistance of a so-called ‘skilled birth attendant’ or biomedically trained clinician.Footnote ⁴ Since the early 2000s, when the World Health Organization (WHO) highlighted skilled birth attendance as a key determinant of maternal health, state and global health policymakers have sought to increase women’s use of biomedical clinics and have sought to marginalize (and even criminalize) the practice of lay midwives or traditional birth attendants (TBAs) (Wendland Reference Wendland2022: 246). In Luuka, despite Uganda’s efforts to increase clinical attendance, women still seek care from a wide array of providers during pregnancy and labour. Government clinics, private licensed clinics, drug shops (small kiosks that sell pharmaceuticals by the dose, but which also provide unlicensed medical care) and TBAs were all common choices for reproductive healthcare. Many women also gave birth at home, either out of necessity or by choice. Care by all kinds of providers could be unpredictable in quality. Government clinics struggled with shortages of both staff and essential medicines. TBAs had a wealth of practical experience but limited biomedical training and few resources to manage complications. These conditions created a ‘pragmatics of uncertainty’ among women and their families, meaning that people often chose care that seemed most practical for them, from within a landscape of care defined by uncertain quality, pervasive shortages in healthcare resources, and unpredictable health outcomes (Whyte Reference Whyte1997).

If shortages of drugs, staff and technology created one form of uncertainty within the healthcare landscape, I found that the role of targeted healthcare interventions – ostensibly designed to ameliorate such gaps in care – could exacerbate the sense of unevenness and unpredictability in care provision. Such interventions provided a patchwork approach to health investment. An NGO might intercede to address one part of a complex problem (transportation), for one segment of the community (pregnant women), for a limited duration. Over time, as I discuss below, a history of interventions like this had an uneven effect on communities. A family’s children could be saved by a programme that registered and fed malnourished children. But once a programme ended, people often had little sense of where the funding had gone or why the withdrawal had happened, or whether a new programme that might appear in its place would be durable or similarly transient. This is a different kind of unpredictability in care provision compared with the gaps in technology and expertise that plagued under-resourced providers, but this kind of uncertainty also critically shaped community attitudes towards healthcare. Because of this, a programme’s success depended not only on providing an infusion of material – free transport, maize for a feeding programme – but also on building a sense of trust within the community about the resource: its durability, its accessibility, its reliability.

The boda boda programme I describe here ran for slightly less than a year, from August 2023 to May 2024, officially enrolling 200 women from twenty villages that were arrayed around the clinic. Some villages were near the clinic and some over a twenty-minute boda boda ride away. VHTs, who were paid by the NGO to perform a variety of outreach and education tasks in their respective neighbourhoods, were charged with selecting and enrolling pregnant women in the programme. In 2023–24, I followed twenty-two women who were enrolled in the transport programme during and after their pregnancies, interviewing them at least once both before and after giving birth. I also conducted interviews following the death of an additional woman who had been enrolled in the programme, but who I had not been following during her pregnancy.Footnote ⁵ (Between June 2022 and June 2024 I interviewed an additional eighty-four women in Luuka during and after pregnancy who were not enrolled in this programme, as well as over a dozen different care providers in the community, from biomedical midwives to TBAs.) I also observed training and planning meetings for the project organized by the sponsoring NGO and travelled with midwives to do outreach visits associated with the programme. Only six of the twenty-three women I followed were successfully transported by the programme’s boda boda to the clinic to give birth. Eight more made it to the clinic using their own modes of transport (some walking, some using private boda bodas). Nine others, including Esther, whose birth I described in the introduction, gave birth at home or with the help of lay providers.

While the boda boda was a seemingly straightforward and stable material good that sought to connect people to healthcare, below I examine it as what Andrew Barry, writing about pharmaceuticals, has called an ‘informed material’, emphasizing how a standardized intervention (such as a medicine or motorized clinic transport) takes on different properties as it circulates in and among communities and is invested with varying levels of trust that inform its use and impact (Barry Reference Barry2005: 57). Importantly, the boda boda project depended on creating a sense of reliability and connection between patients and the resource offered. Notably, this connection was created not only from the top down, in how the programme viewed the women (Should she qualify for this resource?), but also in how women perceived the boda boda (Could it be trusted to come? Is this my most reliable form of care?). In examining these variable meanings, I reveal how the same attributes of community-led, small-scale infrastructures that can be so transformative – their emphasis on community collaboration, their flexibility – can also come to mean something different within communities themselves, where people seeking care increasingly navigate a healthcare landscape defined by partial, often unpredictable, and temporary solutions.

Relations of access: infrastructure is personal, resources are relational

An important aspect of a transitory programme like the boda boda project was that it was, by its nature, selective in its community impact. Women had to be chosen for the programme by their local VHT, who had been tasked by the sponsoring NGO to find pregnant women ‘in need’ of the extra support the programme would provide. Negotiations over inclusion were also often a part of outreach visits conducted by the midwives. During such visits, a midwife and midwife assistant, accompanied by the neighbourhood VHT member, travelled to meet each woman included in the programme to conduct a home-based antenatal examination. This was typically the first moment when community representatives (VHT members) and clinic representatives (midwives) consulted each other about which specific women would be included in the programme from a given neighbourhood, and the first time women interfaced with representatives of the programme.

‘The list’, or the group of ten women selected from each neighbourhood, was a point of discussion at nearly every outreach session I participated in. The midwives regularly consulted with the VHTs about the background of each woman on the list, inquiring about the location of family members, number of children, or other factors that contributed to a demonstration of need. During one memorable interaction, a midwife became upset by a confrontation with the husband at one woman’s house when he refused to fetch his wife for the visit. This midwife declared that the woman was not deserving of the programme because of the way the husband had mistreated his visitors. Deeply upset by the interaction, she demanded that the wife be replaced on the list with another woman ‘who is very bad off’ whom the midwife had encountered earlier that morning. Her assistant responded that an obstinate husband was not a reason for disqualification from the programme, and that such a difficult husband may even be evidence of her need for extra support. After considering this point, the midwife backed down from her demands. Nonetheless, neediness, and by extension the qualities that made a woman deserving of care, was a constant point of negotiation in a programme such as this. Midwives regularly remarked on the condition of homes or the health of the women or the families they visited – ‘That child had no proper clothes’, ‘She has had nothing to eat today’ – to reaffirm a woman’s appropriateness for the programme.

While the effects of how determinations of need help to cultivate certain kinds of patients have been a focus of studies of global health programmes elsewhere (Nguyen Reference Nguyen2010; Biruk Reference Biruk2022; Benton Reference Benton2015), in Luuka these practices seemed to be less about shaping patient behaviour (for instance, the compliance of patients to follow a treatment schedule, or their ability to provide a public face for HIV treatment) and more about an acknowledgement of the relational and variable nature of access to care. This variability in access to care was not due to the malfeasance of programme or clinical officials. This concern with selectivity was driven by conditions of extreme scarcity alongside extreme need. There was never enough fuel to transport everyone, never enough medical supplies to provide free care to all women in labour. Under these conditions, care became intensely personal and variable. Who is deserving? And how do I evaluate who needs support most?

Care was often determined by what I call ‘relations of access’: that is, who you were – needy or not – and how you were positioned in a community – known to a clinic or clinician or not – could be key determinants in access to resources. The collaboration and relational nature of such a programme could benefit many community members, but it also added to a sense of unpredictability in care. What could one expect from this programme? Would one be included or not?

Rebecca Namulondo was pregnant with her third baby and had struggled during this pregnancy with variable levels of financial support from her husband, who travelled far from Luuka for work. Her first child had been born at the clinic sponsoring the transportation programme, but her second had been born at home. She explained that she thought highly of the clinic, but at the time of her first birth she had struggled to pay even the modest fees she had been charged. Her first delivery was during the time of the ‘cards’ – a different, time-limited programme, sponsored by the reproductive health NGO Marie Stopes, which had provided completely free antenatal care and delivery for women who were registered – but she had not been able to register for that programme. Nonetheless she had been brought to the clinic during that earlier labour in the hope of benefiting from the programme, and had delivered her child there. Afterwards, her family could not pay her bill.

Here, Rebecca describes her relationship to the clinic in terms of both the variability of programme access – her inability to access the free care of an earlier intervention – and the ways in which fees themselves could vary based on evaluations of need. Notably, she never referred to this clinic by its actual name in our conversations. She always used the name of the clinic’s founder (‘Musa’s place’) to describe it. Access to resources in this context was deeply relational, and her views of the care she received emphasized it as such. In this case, Musa heard about her predicament and intervened on her behalf, having clinic staff refund the charges after her delivery.

For the NGO’s transport programme, such negotiations over care happened frequently, not only in evaluating who was on or off the list of women, but also in how to respond to emergencies, and what extra costs of care might be covered for a specific woman. Women were supposed to be charged 20,000 USh for all care during labour and delivery, but other forms of treatment (such as malaria treatment, a common complication) were typically not covered and could vary in cost. Emergency transfers to the hospital in the adjoining district were covered for some women in the programme requiring referral care, while for others such transfers were not covered. Determinations were made on a case-by-case basis, taking into account several factors including need, availability of resources, and the personnel involved in the decision making.

Relations of access provided clinic staff with the ability to respond to and adapt to needs in the community, but the consequences were that transient programmes such as this became unpredictable resources for women and their families. Here, a sense of relationality is threaded through care. But such ties were not durable or predictable. Some people were included in programmes, some were not. Some were known to a programme director or deemed in need of extra care, others were not.

Relations of trust: community evaluations of programme reliability

If one effect of the scarcity of a limited resource was that access to a programme could be variable and dependent on personal evaluations by clinicians, the flip side was that community members were conditioned to assess programmes and clinics in return. Was a resource worthwhile? And could it be trusted to provide care? If ‘relations of access’ were about evaluations of patient need by clinicians, ‘relations of trust’ refers to the ways patients considered programmes in turn. The variability inherent in programmes – Where would scarce resources be directed? – created conditions whereby community members were forced to consider this resource in context. Was this variability going to work in their favour? Could Rebecca trust that Musa would intervene on her behalf again? (She considered it unlikely, and she did not pursue use of the transport programme during her third pregnancy.)

When women knew little about the clinic or programme, they tended to react with reservation about the promises being made regarding free resources. Nakayaga Sarah had never heard of the clinic until the midwives came to see her for a home-based antenatal visit. Sarah had used only the services at a nearer government clinic in Busalamu. The midwife began the visit by explaining what services her clinic offered, and the scope of the transport programme. ‘We are from [this clinic]. We treat pregnant women, and all kinds of diseases, even old people, children. We deliver babies. We do immunizations.’ Sarah, who lived about a fifteen-minute boda boda ride from the clinic, responded by asking about its location. ‘But where is Budondo [village]?’ The midwife responded, ‘You will get there because we will handle you during delivery. Transport is arranged. Whatever we use to care for you [treatments], it is only 20,000 shillings.’ Sarah remained sceptical during the visit, explaining that her first birth was at her parents’ home. She added: ‘I don’t know this area.’ The midwife continued to try to reassure her by drawing on the relationship with the local VHT. ‘Your VHT will tell you about the services at [the clinic]. We even have an ambulance. If you need a transfer, it will be arranged.’ Sarah’s hesitancy revealed how a resource alone, such as free transport, was evaluated in and through the degrees of trust and social embeddedness that brought the resource to a woman. In the absence of more social and relational context, the promise of a free motorbike ride was not a certain and reliable good for many women.

Women’s use of and trust in the programme – and thus the likelihood they would eventually use it and deliver in the sponsoring clinic – were shaped in part by how embedded they were in other relationships that gave shape to the programme. Women who were enrolled in the programme and who lived in neighbourhoods closer to the clinic – where they were likely to have visited the clinic in the past for health needs, knew neighbours who had used the clinic, and were likely to know clinic personnel – were much more likely to successfully deliver at the clinic during their current pregnancy than those who lived farther away. When I asked Salima, a mother in the programme who had successfully used the boda boda transport during labour and delivered in the clinic, about her experience, she described her personal ties to the clinic’s neighbourhood:

Salima was a neighbour of Esther’s, who I described in the introduction, and she faced similar barriers to care as Esther did. Her labour was at night, when travel is considered more difficult, and occurred during the same rainy season that stymied Esther’s efforts to get to the clinic. And yet, for Salima, the programme was desirable because of her ties to the clinic (and its surrounding neighbourhood). The programme, by extension, was a product of her ‘place’, adjacent to the home of her natal relatives.

Conversely, in the neighbourhood that was farthest from the clinic (about a twenty-minute ride on a motorbike), none of the four women I interviewed was ultimately transported on the programme’s boda boda to deliver. None of these women had attended antenatal appointments at the clinic prior to programme enrolment, and none had delivered prior pregnancies at the clinic. Two delivered their current pregnancies at the clinic but arrived only after paying for a private boda boda to transport them. One of these mothers explained to me that she had thought the programme’s driver was unlikely to make the trip to reach her because she had gone into labour at the weekend:

Because the distance to the clinic was significant, and the reliability of the programme driver was in question, her husband decided to pay out of pocket for a local driver. Her family paid a significant amount for transport (6,000 USh or US$1.75) because the local driver’s presence was assured. A programme that was located farther away, with employees less embedded in one’s daily life, introduced questions about reliability that shaped community members’ actions and use of the free transportation.

Another father, this one the husband of the most tragic delivery associated with the transport programme, also resisted calling the clinic on the night of his wife’s labour. Agatha was a mother of three children and was pregnant with twins when she went into labour near her home, in the Kiroba trading centre. Early the next morning she died soon after giving birth to the twins, only one of whom survived. Her home was located on the district’s main road, about two kilometres from the programme’s clinic, yet her husband never called the clinic’s number to arrange for free transport. Instead, he was advised by a local boda boda driver to take Agatha to a TBA who was located on the same main road where Agatha lived. I interviewed Agatha’s neighbour, who cared for her during her labour, about what transpired that night:

It is difficult to say what exactly led to Agatha’s death – Agatha’s neighbour declined to assign blame – but certainly a major contributing factor was her family’s decision to take her to a TBA, who was experienced in managing women in labour but had almost no biomedical supplies, such as sterile sutures or misoprostol to control post-partum haemorrhage. From the family’s perspective, however, the risks associated with the TBA were considered alongside the risks of travelling at night, and on poor roads, to a better equipped biomedical clinic. A neighbouring boda boda driver, a man the family knew, told them such a journey was risky and would take them farther from the direct routes leading to the nearest referral hospital, a necessary journey (as it turned out to be) in case of complications. In the crisis following the twins’ delivery, when Agatha was haemorrhaging, the family turned not to the programme that had promised free transport, but to other known relational resources, friends associated with the nearby police station, who might be able to mobilize ambulance transport to the nearest hospital, twenty kilometres away in Iganga. Agatha died before the ambulance could be arranged and have sufficient fuel. Here, as with other decisions to delay or defer the programme’s care, women and their families considered programme resources alongside other kinds of resources and risks, and the relationships that gave shape to different kinds of care.

Like all infrastructures, the boda boda is both a thing and a ‘relation between things’ (Larkin Reference Larkin2013: 329), and use of it is animated by the variable relationships that brought women to the project. These are practices of both inclusion and, often, self-exclusion that are driven less by explicit forms of local and state power, and more by messier experiences of social belonging and relational embeddedness that foster different perspectives on the ‘fit’ between a programme resource and a specific woman. In the absence of permanent solutions to clinical access and transportation, in the immediate moment of a health crisis people relied on each other, and on their assessments of who and what they could most reliably count on. In a broader infrastructural landscape that was filled with projects that were temporary, community trust in a project and a sense of belonging to it – a belief that a driver will show up, that an ambulance has fuel – become central to use, access and the success of a given health innovation.

The failures of flexibility: the partial success of innovative interventions

If women and their families viewed programmes in terms of the relationships that gave them shape and built trust, a programme’s strengths and weaknesses were also conditioned by another defining construct of short-term interventions: flexibility. In its design, the boda boda programme was always intended to be temporary, with funding for fuel and staff lasting less than a year. The grant was small by the standards of large international programmes – US$12,000 – but the clinic aimed to make the most of the investment. Immediately, clinic staff raised questions about the best use of the funds, and the shape and scope of the programme was up for debate by clinic employees and managers. From the outset, this was a programme driven by community ideals – designed by a local NGO manager born in Luuka and adapted and modified by clinic staff. At the first planning meeting, key employees discussed whether home visits by midwives were a good use of the funds. How much should mothers be charged for delivery, if anything? Should free transport be provided only during labour or also during antenatal visits? Should transfers to the hospital be covered in case of emergency referrals? The employees were open to debating a number of scenarios.

This period of debate early in the programme revealed the fluid and flexible nature of the programme’s design: funds were granted to the community organization with the guideline that they be used for maternal health transport, but the application and scope of the programme were in the hands of the local organization itself. There were clear benefits to this approach, as people within the community played active roles in determining how the programme should be designed. (All of these employees had a deep understanding of the community and the health barriers women faced.) These early discussions revealed a very adaptable approach, which defined the programme’s implementation. The programme director’s initial desire to fund antenatal transport was scrapped in the early stages of implementation because of the cost of fuel. As I noted earlier, decisions about whether emergency referral transportation to a hospital would be covered were made on a case-by-case basis, and often after an assessment of need was made by clinic staff.

Flexibility and innovation are hallmarks of neoliberal humanitarian interventions, where adaptable technologies and the collection of better data regarding efficacy and efficiency are framed as solutions to faltering health infrastructures. Flexibility has also been analysed as a form of African adaptation and resilience in the face of retreating state investment in social services (Simone Reference Simone2004; Ibrahim and Bize Reference Ibrahim and Bize2018). Here, flexibility – as a product of both top-down mandates and bottom-up innovation – is the result of certain constraints on aid and investment, a necessary approach in response to the absence of durable and predictable infrastructures: reliable roads, phones accessible to all, cheap fuel. Flexibility is also the result of the temporary nature of short-term interventions like this. This impermanence meant that the resources needed to run the programme – especially human resources – were difficult to account for and secure. While programme grants try to predict every cost associated with implementation – the midwives needed to be fed lunch during their long days in the field, a blood pressure cuff needed to be purchased for use in the field because the only other one was needed in the clinic – other needs, especially human infrastructure, were harder to account for. One of the biggest constraints of a temporary programme was the inability to adequately staff the ‘back office’ – the people who would answer the phones and field pregnant women’s calls. The short-term nature of this project meant that the programme would be staffed on an ad hoc basis, by people already doing other jobs at the clinic.

It should be noted that other formats for the transportation programme – for instance, reimbursing women for private boda boda rides instead of dispatching the clinic’s own boda boda – were never seriously considered. Part of the reason for this was the logistics of such reimbursements. Would women be paid whether or not they used a boda boda? Would that incentivize them to walk instead of hiring transport? Would drivers artificially inflate their prices? But the primary reason was a desire to acquire the material resource of the boda boda itself, which the clinic has kept after the terms of the grant ended. Here, flexibility of funds is considered strategically by the NGO, in terms of what kinds of investments might have the most durability and impact for the organization, as well as for its clients. But the singularity of this resource also created constraints on access to it.

Flexibility ultimately created conditions that eroded the programme’s predictability. Women who tried but failed to access programme transport usually did so either because they were unable to reach someone on the programme phone, or because the weather was determined to make travel too difficult for the programme’s sole driver. Faith, who experienced false labour and later delivered at a lay provider, initially travelled to the clinic but did not do so on the programme’s boda boda, because she could not reach anyone by phone:

Barbra, the mother of a woman who was successfully transported to the clinic, noted that even though her family received free transportation, she too faced issues accessing the programme. First, no one answered her first call to the clinic. (She persisted in calling and soon reached the operator.) Then the programme delayed, and ultimately failed, in carrying her and her daughter back home (transport home after delivery was also promised to mothers):

Esther, whose failure to deliver at the clinic was the case with which I began this article, recounted her own difficulty securing access to the programme’s boda boda:

All three of these women encountered the limitations of a programme that sought to directly address a key community problem, but whose limited resources – both internal (drivers, staff) and external (road conditions) – resulted in an almost unavoidable unpredictability.

The same flexibility that allowed for community-led innovation and local programme design also created limitations in the programme’s reach. The programme’s small-scale and time-limited nature meant that it was unable to invest in perhaps the scarcest resource for clinical programmes, especially in rural areas: workers. The transport programme, especially for the woman who had to answer the phone at all hours, created a new and unforgiving burden that sapped time and energy. She was sometimes unable to manage the calls and complained of women who misused what was supposed to be an emergency-only system by calling simply to ask questions or register health concerns. Also, the programme could not overcome the limitations of the broader network of infrastructures in which it was embedded, especially the poor state of the local roads, which could mean that even if calls were answered and a driver was available, the programme might be unable to reach a woman in labour in time. For pregnant mothers who were well served by the programme – they called and a boda boda showed up – the project was a minor miracle, solving an established barrier to care in the region. But the baked-in constraints also revealed broader structural limitations to a piecemeal approach to health infrastructure, even in the case of innovative, community-led endeavours.

Conclusion: tracing a project’s success through affective ties and intermittent flows

The boda boda programme was an example of an approach to addressing infrastructure deficits that has been idealized in many sectors of the global health and humanitarian world. It was a targeted solution, focusing on the seemingly tangible, visible and measurable problem of clinic transport. It was cheap and flexible, funded using a small grant and conceived and designed by community members from the ‘bottom up’ rather than dictated by the preferences and designs of funders worlds away. And yet, for all the possibility imbued in a programme like this, it was utilized by only a small subset of enrolled women. The limitations of the programme were not only from the outside in, in the ways that an anti-political approach has hindered global health programmes by designing only ‘good enough’ solutions to healthcare problems. They were also found in the ways in which a programme like this plays out and is shaped by local relationships and perceptions of use and value within the community. These limitations were driven by the ways in which targeted programmes are by design shaped by the broader social context and existing matrix of limited health and other infrastructures. When infrastructures are already restricted – poor roads, unpredictable quality or accessibility of clinical care – community members weigh even free programmes in terms of their relative strengths and inherent risks.

A surprisingly high number of women failed to use the free transport. Only a quarter of the women in my sample called for and received a free ride to the clinic. There were two points at which the programme failed to reach women: first, when women chose not to call for transport once their labour began; and second, when women tried but failed to get the free boda boda to come for them. Both of these points of failure revealed the limitations of targeted, adaptive programmes. All resources and forms of care were understood on personal, relational terms; not as objective goods, but as extensions of other kinds of social relationships that could be either strong (close by, well embedded in one’s world) or weak. This made some women delay care or defer to other kinds of care, rejecting the free transport. Women also tried to reach the free transport and failed to do so, in most cases because of the shortcomings of larger-scale infrastructures on which the programme depended for success, especially well-maintained roads, and of ‘soft’ resources (workers). Conditions of uneven and unpredictable resource allocation meant that people were left to evaluate risks and benefits themselves and judged programmes in terms of the kinds of resources they viewed as most reliable: friends, family and past experiences.

This article has shed new light on key attributes of what have been described as ‘incremental’ and community-led approaches to infrastructure development in rural Africa (Silver Reference Silver2023). Much of this work has highlighted how heterogeneous, ad hoc and community-led infrastructures are inherently flexible and adaptable, built on modes of social collaboration and reciprocity (Simone Reference Simone2004: 409). My focus here has been to better understand the unevenness and unpredictability that may follow from an emphasis on adaptability, and that may impact access to new resources. This focus has revealed how even a free resource is engaged in complex ways within a community and is judged against a landscape of care that is inherently unpredictable. In this context, women considered the programme relative to other kinds of care that existed in the community and that could be deemed more reliable, closer at hand, or better integrated into their social network.

The boda boda itself is imbued with possibility in Luuka, representing both accessibility and modern movement and connectivity, a kind of mobility that is both future-oriented and within reach. In this sense, the free transport programme, despite its only partial success, stood for something greater within the community; it was representative of change, possibility and bottom-up solutions. But more than anything else, perhaps, the boda boda programme revealed the ways in which the meanings imbued in infrastructures – especially the ways they play out through cycles of promise and failure, through variable experiences of trust and perceived risk – shape experiences of use and access that extend beyond any simple calculation of the material value of a healthcare resource.