‘Oh, she is just a nurse’ Re-imagining the role of the nursing workforce in Uganda after a decade of ART scale-up (2004-2014)

Background The expanding roles and increasing importance of the nursing workforce in health services delivery in resource-limited settings is not adequately documented and sufficiently recognized in the current literature. Drawing upon the theme of 2020 as the international year of the nurse and midwife, we set out to describe how the role of nurses had expanded tremendously in health facilities in Uganda during the era of anti-retroviral therapy (ART) scale-up between 2004 and 2014. Methods A mixed-methods study was conducted in two phases. Phase One entailed a cross-sectional health facility survey (n=195) to assess the extent to which human resource management strategies (such as task shifting) were common. Phase Two entailed qualitative case-studies of 16 (of the 195) health facilities for an in-depth understanding of the strategies adopted (e.g. nurse-centred HIV care). We adopted a qualitatively-led mixed methods approach whereby core thematic analyses were supported by descriptive statistics. Results We found that nurses were the most represented cadre of health workers involved in the overall leadership of HIV clinics across Uganda. Most of nurse-led HIV clinics were based in rural settings although this trend was fairly even across all settings (rural/urban/peri-urban). A number of health facilities in our sample (n=36) deliberately adopted nurse-led HIV care models. Nurses were empowered to be multi-skilled with a wide range of competencies across the HIV care continuum right from HIV testing to mainstream clinical HIV disease management. In several facilities, nursing cadre were the backbone of ART service delivery. A select number of facilities devised differentiated models of task shifting from physicians to doctors to nurses in which the latter handled patients who were stable on ART. Conclusion Overall, our study reveals a wide expansion in the scope-of-practice of nurses

during the initial ART scale-up phase in Uganda. Nurses were thrust in roles of HIV disease management that were traditionally the preserve of medical doctors. Our study underscores the importance of reforming regulatory frameworks governing nursing workforce scope of practice in Uganda such as the need for evolving a policy on task shifting which is currently lacking in Uganda.

Background
Sub-Saharan Africa is faced with a severe human resource for health crisis despite an overwhelming disease burden [1], [2], [3]. The majority of the 57 countries listed by the World Health Organization as having a human resources for health crisis are from this region alone [4]. Part of this crisis is manifested in the acute shortage of health workers across multiple cadre and at all levels of the health-system [5]. The shortage of physicians is an especially pronounced constraint which is compounded by labour market dynamics that render many public facilities unable to attract and retain physicians [5]. As such, countries in SSA have grown to increasingly rely on nurses or other non-physician cadre especially those with shorter tertiary training cycles or those who are less-specialized [6].
In many countries in Sub Saharan Africa, such as Uganda, nurses constitute as high as 80% of the entire health workforce [7].
The HIV and AIDS epidemic exacerbated these staffing shortages and dramatically increased workloads due to escalating HIV client loads across Sub-Saharan Africa [3].
Nurses were thrust into roles they were not prepared for by virtue of their pre-service training considering that 'nursing education mainly focuses on clinical skills and theory related to patient care and management' [8]. Nurses were brought to the forefront of national HIV epidemic responses and drafted into a range of HIV clinical services which were traditionally the preserve of physicians based on western models of care that derive from better resourced health-systems [9]. However, the continuous increase in demand for HIV treatment brought on, partly, by changing thresholds of eligibility for treatment over the past two decades demanded pragmatic shifts [10]. The World Health Organization released policy guidelines recommending task shifting from physicians to non-physicians such as nurses and midwives 'the WHO recognizes that nurses and midwives can provide services in the clinical setting, such as HIV clinical staging and the management of opportunistic infections' [11]. Although several countries are yet to formally adopt task shifting of clinical HIV disease management from physicians to nurses in their regulatory frameworks, [12], [13], some countries have. In 2010, South Africa released policy guidelines providing for Nurse-Initiated and Managed ART (NIART) [14].This was one of the watershed moments in the re-definition of the role of the nursing workforce in health service delivery in general. There are several studies demonstrating the non-inferiority of nurse-managed HIV care and treatment [15]

Research design
We adopted a mixed-methods sequential explanatory research design [23] . This study was conducted as part of a four-year doctoral research project aimed at understanding the organizational strategies devised by health facilities in Uganda to promote the sustainability of anti-retroviral therapy (ART) scale-up programs from the perspective of WHO's building blocks of the health system framework [4]], [10][24] [25]. The study was conducted in two phases which were implemented sequentially [26]. Phase One entailed a cross-sectional health facility survey to assess the extent to which human resource management strategies (such as task shifting) were common. Phase Two entailed a casestudy of 16 (of the 195) health facilities for an in-depth understanding of the strategies adopted and the operational context(s) underpinning these strategies [27].

Study sites and sampling
With regard to the health facility survey (Phase One), we secured the published list of health facilities accredited to provide ART by the Ministry of Health of Uganda [28]. This published list indicates that 394 were accredited to provide ART as at March 2010 and the sub-regions of Uganda where they were located. This list served as a sampling frame for our study [3]. The 394 health facilities were placed in 10 cluster based on their location in the 10 geographic sub-regions of Uganda as designated by the Uganda Bureau of Statistics [29]. Using proportionate size-to-sample technique [3] we randomly selected 195 (out of 394) health facilities [3].
In Phase Two, we purposively selected sixteen health facilities to achieve diversity with regard to ownership-type (public/private), setting (rural/urban/peri-urban), level of care in the health system (tertiary/secondary) and sub-region of Uganda (e.g. Northern/Southern). Table 1 shows the characteristics of case-study facilities.

Data collection
Phase One: A 35-item questionnaire was self-administered by the ART clinic in-charge at each of the 195 health facilities. This questionnaire is described elsewhere [3], [10], [24].
A filled hard-copy questionnaire was picked on-site by a research assistant a week it was delivered to the ART clinic in-charge. To mitigate potential non-response bias, phone call reminders were made to respondents after a three-day interval [3]. The survey was conducted between January and April 2014.
Phase Two: A topic guide was constructed with the aim of understanding the operational contexts underpinning the human resource for health strategies adopted at participating facilities. We conducted 24 face-to-face in-depth interviews (IDIs) with ART clinic incharges and staff with personnel management responsibilities (such as the head of clinical services or facility in-charge) in their offices within the facilities. On average, these interviews lasted between 45 to 60 minutes. The qualitative interviews were conducted in August of 2014.

Data analysis
Phase One: Questionnaire data were initially entered into EpiData software (version 3.1).
Data were later exported into STATA (version 12). Descriptive statistics were generated relating to the demographics of participating facilities, the extent to which varied human resources for health strategies (such as task shifting from physicians to nurses) were common.
Phase Two: We broadly followed the procedures for qualitative data analysis recommended by Mile & Huberman [30].The audio recordings of semi-structured interviews were transcribed verbatim into text transcripts. For purposes of data familiarization [31] the transcribed transcripts were read multiple times in order to generate an initial coding scheme. The coding scheme was applied to the interview transcripts. Emergent codes were then abstracted into three broad thematic categories [30]. A data validation workshop was conducted with five ART clinic in-charges to assess agreement in interpretation of the study findings [32]. Their in-put and feedback informed the final analyses [32].

Mixed-methods integration
The quantitative and qualitative data were merged together [33], [38]. A qualitatively-led mixed methods approach [34], [38] was adopted where the emergent thematic categories arising from the qualitative analyses were expanded with supporting descriptive statistics [26], [38].

Introduction
In this section we begin by presenting the characteristics of the study population or the health facilities which participated in the study. We then present the study findings based on three three broad themes and the sub-themes therein. The three broad themes include a) Nursing cadre involvement in leadership of HIV clinics, b) The adoption of nursecentrered HIV care models and c) Differentiated models of task shifting to nurses.

Characteristics of participating health facilities
Overall, 195 facilities across Uganda participated in the study. With regard to ownershiptype, 121 (61%) were public facilities, while 35 (18%) were private-not-for-profit and 33 (16%) were private-for-profit. HIV research clinics in our sample were 6 (3%). Table 2 shows that by level of care in the Ugandan health system, the majority of participating facilities were Health Centre IVs (sub-district facilities) at 72 (36.9%). This was followed by district hospitals at 58 (29.7%).
In terms of setting, almost half of the health facilities 88 (45%) were based in peri-urban areas (urbanized parts of rural areas). This was followed by 78 (40%) health facilities which were located in urban settings while 29 (15%) of the facilities were based in rural settings.

The nurse in-charge: leadership roles in HIV clinics
Our cross-sectional survey reveals that across the 195 health facilities participating in the study, nurses were the most represented cadre of health workers reporting a role as ' ART clinic in-charge' or the overall leadership of HIV clinics. As Table 4  In the in-depth interviews with nurses, they described finding themselves in leadership positions in HIV clinics across Uganda. This was especially so in lower-level health facilities and was certainly the case in health facilities in rural settings. Rural health facilities are often shunned by more specialized health worker cadre. However, even in regional referral hospitals (such as PUB-02), the head of the stand-alone HIV clinic was a nurse.

Nurse-led HIV clinics in rural settings
Our survey findings reveal a rural-urban dichotomy with regard to the phenomenon of nurse-led HIV clinics. As Table 5 show, most of the nurse-led HIV clinics were based in rural settings. However, nurse-led HIV clinics were fairly even across setting.
Our qualitative interviews with facility in-charges shed more light on why nurse-led HIV clinics were mostly in rural settings. This was explained by a facility in-charge of one of the case-study facilities selected for in-depth understanding of the human resources for health strategies adopted by providers. Although nurse-led HIV clinics were more common in rural settings, during on-site visits to facilities participating in the national survey, it was not uncommon to find that nurses were ART clinic in-charges even in tertiary facilities which are predominantly in urban settings. This was certainly the case in several district hospitals we visited and even in a number of regional referral hospitals across the 10 sub-regions of Uganda we visited.

The management skill sets demanded by HIV clinic leadership
As overall leaders of HIV clinics, nurses reported that they were responsible for the dayto-day operational aspects of running these busy stand-alone HIV clinics. As such, they

Nurse-centred HIV care models
Findings from the national health facility survey show that in 93% (181) of the 195 health facilities, non-physicians were engaged in the clinical management of ART. Facility in-charges explained that they had difficulty retaining high cadre such as medical officers.
We found that several health facilities had deliberately adopted nurse-centred HIV care models as a mitigation strategy. Nurses were relied upon as the primary cadre in providing and managing anti-retroviral therapy services in a number of facilities. In a large mission hospital (PNFP-01), nurses were trained to be versatile in a wide ranges of tasks within the HIV clinic. This was described by the ART program manager of the mission hospital.
'I think my nurses are the best asset I have here. HIV has brought out the potential of nurses. We used to think 'oh, she is just a nurse'. But they can do a lot! And I think that's our strength here. They will retrieve records and patient files. They will manage the reception. They will dispense. They can do a lot. And I think that's our strength here'. ART 'We do a lot of in-house staff development. We know that every Wednesday is a CME day (continuing medical education). We conduct on-site training for our nurses… many never did presentations but now our nurses do can stand up and do presentations, I mean power-point. So we empower them because we don't pay them that much and when there is an off-site workshop we try to make sure that everyone gets a chance to go to'. ART clinic in-charge, PNFP-02 Across our interviews with facility in-charges, there was a widely-held perception that nurses were more stable at health facilities compared to cadre with longer post-secondary school training. Medical officers were perceived to be mobile on account of their affinity for accepting additional training opportunities and better job offers. As such, nurses were perceived as a cadre that were dependable and could be relied on for long-term planning due to their tendency to remain in the same employment and work stations compared to more specialized cadre.

Head of clinical services, PNFP-01
Due to the low pay wages common in public facilities in Uganda, nurses were perceived to be more resilient. 'We have some cadres who are not easy to work with due to our low wages. For instance pharmacists because they have so many pharmacies around which they are supervising and they are crying that the pay is low. So, they always want to get something from outside .So, those ones have been a problem for us but not nurses and midwives. Another cadre is the medical doctors also. They have been a problem to us because most of them were crying of the low salary' Hospital administrator, PUB-01.

Nurse-led HIV care
Although a number of health facilities indicated that their HIV service delivery was nursecentred, 38

Differentiated task-shifting to nurses
In a section of health facilities, we found a more differentiated approach to task shifting to nurses. This was in contrast to the practice on some health facilities of devolving clinical management of ART from physicians to nurses.

Stable patients managed by nurses
Patients with advanced HIV disease were reported to be managed by clinician cadre such as clinical officers and medical officers (or assistants) while patients who were deemed stable on ART were handled by nurses.

Nurse-led community-based outreaches
Nurses were often assigned to lead HIV care outreach teams into communities. A mission hospital (PNFP-01) within our case-study facilities run a home-based HIV care program for stable patients who live within a 5-kilometre radius of the hospital. Nurses often led typically four-members teams that also comprised 'expert patients'. They conducted reviews during these home visits and distributed ART refills. The nurse-led outreach programs were aimed at decongesting overcrowded HIV clinics and easing transport burdens on patients. The nursing workforce was said to critical in these endeavors.

Discussion
The changing roles and increasing importance of the nursing workforce in health service delivery in resource-limited settings is not adequately documented and sufficiently recognized in the current literature [6], [20], [22]. We set out to describe how the role of nurses had expanded tremendously in health facilities in Uganda during the era of ART scale-up between 2004 and 2014 [10]. We found that nurses were the most represented cadre of health workers involved in the overall leadership of HIV clinics across Uganda.
Most of the nurse-led HIV clinics were based in rural settings although this trend was fairly even across the rural and urban divide. A number of health facilities in our sample deliberately adopted nurse-centred HIV care models owing to an inability to attract and retain cadre with longer training cycles. Several health facilities devised differentiated models of task shifting from physicians to doctors to nurses in which the latter handled patients who were stable on ART. Overall, our study reveals a wide expansion in the scope-of-practice of nurses. Nurses were thrust in roles of HIV disease management that were traditionally the preserve of medical doctors. In many health facilities nursing cadre were the backbone of HIV service delivery. These expanded roles include prescribing ART, Tuberculosis (TB) management and ART adherence counselling.
An important contribution of this paper is in providing quantitative data showing the extent of representation of nurses in the overall leadership of HIV clinics in a national sample of health facilities in Uganda. In this study we found that nurses were the most represented cadre of health workers indicating a role as 'ART clinic in-charge'. Due to the leadership and management roles associated with leading HIV clinics, management generalist skill sets became imperative. Nurses often have to grapple with the challenge of motivating personnel in the face of rapidly expanded workloads owing to escalating HIV client loads. Nurses have to ponder responses to frequent ART stock-outs and to manage the associated supply chains. Our study therefore suggests that leadership and management training needs to be strengthened in nurse pre-service training curricula in Uganda and other resource-limited settings to prepare them for the leadership roles that many nurses are frequently thrust into [6] , [37], [43]. Importantly, our study highlights the importance of reforming regulatory frameworks governing nursing workforce scope of practice in Uganda and other countries. Several studies have noted the absence of a task shifting policy in Uganda [12], [13] one that provides an enabling framework for the expanding roles of nurses [43], [44].
Our survey results show that most of the nurse-led HIV clinics were based in rural areas.
These data provide further empirical credence to notions of the mal-distribution of health workers in sub Saharan Africa [1], [5] which suggests that more specialized cadre such as doctors and pharmacists shun rural settings. Studies have highlighted how frequently nurses are the only cadre available to provide health care in rural, remote outposts or hard-to-reach area [1], [2], [6]. These findings suggest that in rural settings HIV care is largely nurse-led. There is therefore need to strengthen HIV disease management in nurse education curricula as the likelihood of nurses engaging in the clinical management of ART is high ,as our survey findings demonstrate. Given that nurses are the mainstay of health service delivery in rural settings, their shortage per capita is a fundamental constraint that merits governance reforms [36]. To address the underproduction of nurses, there have been calls for the promotion of science education in secondary schools in Uganda to boost the pool of prospective nurse recruits [5].
Beyond the global HIV response, Rabkin and colleagues [35], have noted the need to leverage HIV lessons such as task shifting to nurses in the response to the burgeoning non communicable diseases (NCDs) epidemic with the requisite training, guidelines standardization and mentorship support.
We found that several health facilities had deliberately adopted nurse-led HIV care models. Nurses were empowered to be multi-skilled with a wide range of competencies across the continuum of HIV care right from HIV testing to mainstream clinical HIV disease management. Our findings underscore the need for regular refresher trainings in HIV care and treatment for nurses and other less-specialized cadre who are the mainstay of ART service delivery across Uganda .This could be in the form of regular seminars and workshops whether these be in-house or off-site by the Ministry of Health and donors. Onsite support supervision and mentorships of nurses by specialized cadre have been acknowledged in the literature as supporting interventions [3], [12], [13], [38]. Although our study reveals that 93% of the 195 health facilities in Uganda allowed non-physician cadre to prescribe ART, Baine et al [12] and Dambisa et al [13] have observed the lack of a formal policy on task shifting in Uganda. A recent multi-country survey by Ford and colleagues [38] found that in many countries policies currently provide for ART initiation on first line ART by non-physicians although the same guidelines preclude them from managing children and patients on second line ART. This, despite an estimated need of 3 million people in Sub-Saharan Africa in need of second line ART 2020 [38]. Our study findings add a voice to calls for reforms in regulatory frameworks that acknowledge the proven competence of nurses and allow for advanced roles for nurses in their permitted scope of practice [20], [22], [42], [43], [44].
On the other hand, our study reveals that select health facilities adopted differentiated task shifting to nurses. Whereas patients with advanced HIV disease were managed by clinician cadre (especially medical doctors), patients who were stable on ART were managed by nurses. In a number of health facilities running community-based platforms of HIV care, home-based visits and the monitoring of patients within communities was nurse-led. Our findings find resonance with the currently topical differentiated service delivery (DSD) agenda [39], [40], [41]. DSD calls for tailoring HIV care to the needs of individual patients including differentiating the cadre assigned to manage patients based on clinical criteria such as those who are stable and those who are not [9], [39]. The role of nurses in managing less-intensive models of HIV care for stable patients especially those enrolled in community-based ones such as community outreach ART refill pick-up points and community-based patient adherence support groups is especially timely [9], [10], [39], [40], [41].

Conclusion
Overall, our study reveals a wide expansion in the scope-of-practice of nurses during the course of ART scale-up in Uganda. Nurses were thrust in roles of HIV disease management that were traditionally the preserve of medical doctors. Our study highlights the importance of reforming regulatory frameworks governing nursing workforce scope of practice in Uganda such as the need for evolving a policy on task shifting which is currently lacking in Uganda.
No: 191994), Sida (Grant No: 54100029) and MacArthur Foundation Grant No: 10-95915-000-INP. The funders did not in any way contribute to the design of the study and collection, analysis, and interpretation of data.

Authors' contributions:
HZ conceptualized this paper, analyzed the data and produced the manuscript draft.   Table 5, mentioned on page 9, was omitted by the authors in this version of the paper.