Coordinating under constraint: a qualitative study of communication and teamwork along Ethiopia’s cervical cancer care continuum

Abstract

Background

As Ethiopia expands cervical cancer screening services, it urgently needs information to develop appropriate post-screening diagnostic and treatment services for women with abnormal results. Quality cancer care requires extensive coordination among multidisciplinary provider teams. This study explores experiences coordinating care among providers at multiple levels of the cancer-care continuum in Ethiopia.

Methods

From February 2020 to January 2022, we conducted four focus group discussions (FGDs) and ten key-informant interviews with 34 purposively selected healthcare providers: health extension workers (HEWs) in communities; midwives and nurses at health centers; obstetrician-gynecologists at regional hospitals, and oncology nurses, oncologists, and pathologists at tertiary hospitals. FGDs and interviews were conducted in Amharic and audio-recorded. Audio transcripts were then simultaneously transcribed and translated into English for analysis. Investigators performed thematic analysis using inductive and deductive codes.

Results

We found four themes: resource scarcity, care centralization, lack of formal coordination mechanisms, and recommendations. Themes were dynamically connected by eight sub-themes. Providers valued teamwork and coordination. However, severe shortages of cancer specialists and high patient loads left little time for communication and hampered the formation of stable care teams. Facilities lacked formal coordination systems, such as patient navigators and case managers. The relative centralization of cancer care specialists and equipment in the capital exacerbated coordination problems. It impeded pre- and post-treatment care communication between tertiary and secondary facilities and caused secondary facilities to unnecessarily refer patients because they lacked the resources to treat them locally. Referral communication was unidirectional, with lower-level providers communicating regularly to higher-level facilities but rarely receiving feedback. The exception was regular, structured feedback from primary facilities to HEWs. Lower-level providers wanted to learn whether their referrals were appropriate or completed, and many used informal channels to gain this information. Respondents recommend decentralizing cancer care services, significantly increasing staff and equipment investments, and adding liaison staff at secondary hospitals to track and communicate patient progress and counsel patients for referral.

Conclusions

Our findings underscore the need to rapidly increase cancer specialist staff and regional cancer centers in Ethiopia and highlight the importance of developing robust coordination and feedback mechanisms at secondary and tertiary facilities.

Background

Almost 90% of new cervical cancer cases and deaths occur in low- and middle-income countries, with the highest burden in Sub-Saharan Africa (SSA) [1,2,3]. In response to this situation, countries in SSA have initiated cervical cancer prevention programs, hoping to replicate the successes that high-income countries have seen with human papillomavirus (HPV) vaccination and routine screening using HPV DNA testing, cytology, or visual inspection with acetic acid (VIA) [4]. These essential primary and secondary prevention interventions require fewer resources and are more cost-effective than treating invasive cancer [3]. However, most women in SSA who have abnormal screening results already have advanced disease [2, 5, 6]. As primary and secondary prevention strategies are scaling up, tertiary prevention services will, therefore, remain a necessity [7, 8], and countries such as Ethiopia will need strategies to expand and improve post-screening follow-up, retention, and treatment for the increasing number of clients who will require care [9, 10]. The Ministry of Health’s priority in these strategies will be reducing the treatment delays that allow early-stage cancers to progress to fatal cases. Such delays are common in Ethiopia, driven by the scarcity of treatment facilities and specialists and a poorly structured referral network [11,12,13].

Ethiopia’s cancer treatment systems are severely under-resourced. Studies report only a handful of oncologists practicing in the country of approximately 115 million people [14]. Even the country’s flagship cancer treatment facility, Tikur Anbessa Specialized Hospital (TASH)—the only facility in the country that provided radiotherapy at the time of the study—lacks the specialists, equipment, supplies, and centralized record-keeping systems needed to deliver quality care consistently [8, 9]. The country’s challenge is to find strategies to increase the quantity and quality of cervical cancer prevention and treatment under these severe resource constraints. Care coordination has long been seen as an important low-cost quality-improvement strategy because it promotes responsive person-centered care, improves health outcomes, and increases the efficiency of healthcare spending by reducing duplication, medical errors, and delays in care that lead to disease progression [15, 16].

Care coordination involves proactive practices and procedures designed to encourage provider collaboration, streamline service delivery, and facilitate patient navigation to ensure patients receive person-centered, integrated care [15]. It has been conceptualized as having three core aspects: 1) organizing services, which encompass administrative tasks such as making timely referrals and appointments; 2) sharing information between providers, patients, and multidisciplinary team members; and 3) relationships, which refers to the quality and extent of interactions among everyone involved in patient care [17]. Care coordination involves communication and collaboration both within and between organizations as patients move from one facility to another [18]. The outcome of care coordination is continuity of care, or “the delivery of services by different providers in a coherent, logical and timely fashion” [19, 20]. Care coordination is vital to improving health outcomes in general and is essential for cancer care in particular because cancer treatment is complex and often involves multiple providers from different disciplines working in different settings over long periods. Because care coordination interventions can be cost-saving, they are particularly important for resource-constrained settings [18, 21] and are potentially feasible in highly centralized health systems such as Ethiopia’s.

Little research has been conducted on cancer care coordination in SSA [15, 21]. Recent reviews have found only a handful of such studies in Africa, none conducted in Ethiopia, the continent’s second most populous country [21, 22].

Studies of cancer care in Ethiopia have focused on patient-provider communication at tertiary facilities and suggest that patients value timely, accurate communication about their care [12, 23], but that workload and privacy constraints prevent such conversations [24]. Other Ethiopian studies have documented care delays due to lengthy referral processes with few mechanisms to follow up with referred patients or to reintegrate them into primary care once their treatment ends [6, 12, 25]. Of the many research studies conducted on barriers to cervical cancer prevention and care in Ethiopia, no studies to our knowledge discuss post-screening care coordination along the care continuum and its implications for patient experience and outcomes. In response to this gap in the literature, this study seeks to understand how cervical cancer diagnosis and treatment are coordinated, the barriers and facilitators to coordination, and the points at which coordination occurs and breaks down. We seek to document the coordination experiences of a variety of providers at all tiers of the healthcare system in order to assist policymakers in pinpointing priority areas for interventions on the care continuum.

Methods

Study context

Ethiopia has a three-tiered national health care system with primary, secondary, and tertiary levels of care. At the primary level, care is delivered by Health Extension Workers (HEWs) in the community and at health posts and by nurses, midwives, and health officers at health centers and district hospitals. At the secondary tier, referral hospitals deliver care to regions of up to 1.5 million people. The tertiary tier consists of specialty hospitals serving 3.5 to 5 million people. Cervical cancer screening and preventative care are provided free of charge at the primary and secondary levels, and tertiary treatment is free for those deemed indigent by local officials [26]. Health insurance is delivered through two government programs: Community-Based Health Insurance and Social Health Insurance to maintain universal coverage [27]. At the time of this study, the only public facility providing comprehensive cancer treatment was Black Lion (Tikur Anbessa) Specialized Hospital (TASH) in Addis Ababa. St. Paul’s Hospital in Addis Ababa also provided cancer care but did not offer a comprehensive suite of services. TASH and St. Paul cancer care units are staffed by an extremely small but dedicated staff of specialists [14]. Regional cancer centers were coming online by the end of the study period.

Research team and reflexivity

The project was developed by an interdisciplinary team of Ethiopian and American public health and clinical researchers (SB, AH, BG, and EM) who had previously conducted cervical cancer research in Ethiopia. The project emerged from a study of cervical cancer patient experiences in Ethiopia conducted by SB, AH, and EM with the assistance of HS [12]. HS, a cis-gender female Ethiopian researcher from the Amhara region, was responsible for recruitment, focus group facilitation, interviewing, and data transcription, translation, and management. She was selected because of her strong qualitative research skills, her fluency in English and Amharic, her multidisciplinary social science background, which includes graduate degrees in gender and human rights studies (in alignment with the study’s patents’ rights motivation), and her deep familiarity with Ethiopia’s health system gained from her previous reproductive health research on maternal mortality and reproductive health care access for LGBTQ individuals. HS had no prior relationship with the vast majority of participants prior to the study.

The data-analysis team was led by SB and AH, American researchers who have lived and worked extensively in sub-Saharan Africa, the Caribbean, and Asia, and who had previously conducted qualitative, reproductive health research with Ethiopian patients and healthcare providers. They brought an understanding of patient and provider experiences to codebook development and analysis. The research assistants who assisted data analysis (KP, NK, SH, RD, GM, SW) were second-year Master of Public Health students who had received qualitative methods and global health training. All but one were also in the final semester of obtaining a clinical degree and conducting rotations. They, therefore, had a solid understanding of the dynamics of working in a clinical care team, which they brought to their analysis. None of the members of the data analysis team had direct contact with participants in the study.

Study design and sampling

To assess provider experiences of care coordination and the referral system in Ethiopia, we conducted a qualitative study involving four focus group discussions (FGDs) and ten individual key informant phone interviews with healthcare providers. Participants were selected via purposive snowball sampling, with the aim of recruiting providers at each level of the health system along the cervical cancer screening, diagnosis, and treatment continuum. This involved recruiting HEWs at the community level; midwives and nurses at primary care health centers; obstetrician-gynecologists (OBGYNs) at secondary referral hospitals, and oncology nurses, oncologists, and pathologists at a TASH.

Eligibility criteria

All healthcare workers who had experience working with cervical cancer patients, who were over the age of 18, and who had worked at their post for more than one year were eligible to participate in the study.

Recruitment strategy

Our research project was based out of the University of Gondar (UoG) in the Amhara region of Ethiopia. We invited participation from individuals throughout Ethiopia. Recruitment for specialist providers was conducted via email, phone, and WhatsApp messaging, with initial “seed” participants drawn from our research team’s internal lists of cancer specialists and reproductive health providers developed in our previous research in Ethiopia. Recruitment for focus groups with the non-specialist HEWs and mid-level providers occurred in person at health centers that were geographically proximate to the secondary and tertiary facilities from which specialists were recruited. Recruitment continued until we had spoken with members from all types of clinicians involved in cancer care and we found we had reached data saturation, with themes being repeated during our ongoing analysis.

Data collection

Data were collected using original semi-structured Amharic-language interview guides, specific to the provider’s health system level (see Supplementary file 1). Each guide contained questions on patterns of provider communication, the extent of feedback on referrals received or given, mechanisms for teamwork and coordination, perceptions of barriers to coordination, and suggestions for improving care and coordination.

Data were collected from February 2020 to January 2022. Data collection commenced with in-person focus group discussions (FGDs) with HEWs and midlevel providers. FGDs were held at primary care health centers in Addis Ababa, the capital, and Bishoftu, a town 30 miles south of the capital. Each focus group was approximately 90 min long, and each interview was approximately 40 min. Due to the COVID- 19 pandemic, we replaced FGDs with semi-structured phone interviews with key informants in mid- 2020 for data collection from the remaining providers. We recruited oncologists and oncology nurses from TASH and OBGYNs and pathologists from regional and secondary hospitals in Addis Ababa, Debre Markos, Debre Tabor, and Mekelle.

Both FGDs and interviews were conducted by the study’s Project Coordinator, HS. She was assisted in FGDs by a cis-gender Ethiopian female note-taker. Interviews and FGDs were audio-recorded with prior written (for FGDs) or verbal (for interviews) consent.

Data analysis

HS transcribed and translated audio recordings from Amharic to English simultaneously with the assistance of a professional translator. Weekly debriefing phone meetings were held between HS and the lead project investigator (SB), after which FGD and interview guides were revised as needed and recruitment plans updated.

Approach and frameworks

We analyzed transcripts continuously during data collection using mixed inductive and deductive thematic coding, following Braun and Clarke’s recommendations [28, 29]. We chose a thematic analysis approach because it is recognized as an effective strategy for exploring the experiences and practices of participants and appropriate for studies like ours that are descriptive rather than hypothesis-generating [29,30,31]. We approached the data with an experiential orientation that prioritized participants’ own accounts of their attitudes toward and experiences of coordinating care. Within the thematic analysis method, we used a mixture of reflexive and codebook approaches to coding. We started coding using theoretically and empirically driven deductive codes but then allowed team members to follow an inductive, reflexive approach in which codes were understood as representing individual researchers’ understanding of data patterns, and a codebook was developed over multiple iterative rounds of data analysis. The evolution of codes into themes occurred at the end of the coding process through a series of discussions.

Our analysis was guided by several empirical and theoretical sources. Empirically, we drew on research with patients and providers in Ethiopia and sub-Saharan Africa that documented gaps in cervical cancer care [12, 23, 32]. Theoretically, our interview guide and the subsequent codebook drew on definitions of care coordination and its constituent elements laid out in the WHO’s Continuity and Coordination of Care Practice Brief [15], as well as the cancer-coordination-specific definitions described by Walsh et al. [17]. We also drew extensively on McDonald et al.’s syntheses of coordination frameworks, which highlights the multiplicity of individuals involved in coordination, their interdependence in providing care, the importance of shared knowledge regarding roles and resources in this network, and the essential nature of information sharing to facilitate coordination [33]. Elements of McDonald et al.’s subsequent Care Coordination Measurement Framework were also used to guide instrument and codebook development, particularly its discussion of points of transition in care as important sites of coordination failure [34].

Coding process

The project leadership (SB, AH, EM) and HS developed an initial list of codes based on our theoretical and empirical sources and interview guide questions regarding communication, referral, and teamwork. SB then led two-person teams of Master of Public Health students from Touro University California in developing a codebook containing these initial codes and new codes that emerged from reading the transcripts. This occurred in three rounds over the data collection period with three teams of coders (KP and NK; SH and RD; GM and SW). Coding team members first read a sample transcript individually and developed codes for it independently. These codes were compiled into a draft codebook along with the initial deductive codes. In twice-weekly meetings, SB and coding team members discussed codes, refined and culled them, added definitions, created sub-codes. Once consensus was reached on the codebook, team members blind-coded a single transcript to test for inter-coder reliability. Once this was established, they independently coded the remaining transcripts and recoded any previously coded transcripts with the updated coding scheme. Once transcripts were coded, project investigators met over several weeks to develop and refine themes and identify patterns and relationships between them. All coding was conducted using Dedoose software.

Results

Characteristics of study participants

We conducted four FGDs and 10 interviews. Each FGD had six participants. In total, we sampled 34 providers. We had a 100% response rate. All providers we approached who met our eligibility criteria agreed to participate. Most respondents were women (76%) who worked at the primary-care level (71%). Approximately a fifth of the sample worked at TASH (Table 1).

Table 1 Characteristics of study participants (n= 34)

Overview of themes

The central and overriding theme emerging from FGDs and interviews was that there were too many patients and insufficient staff and equipment to coordinate their care adequately. There was a circular relationship between this overriding resource constraint, the resulting centralization of cancer care at the few secondary and tertiary facilities with treatment resources, and the resulting unnecessary referrals to these facilities, which, in turn, then exacerbated their problems with patient load (Fig. 1).

Fig. 1

Relationship between themes and sub-themes

The lack of staff and equipment, coupled with the enormous patient loads, made it all but impossible for providers at the tertiary level to coordinate with each other or with external providers as they lacked the time. This lack of formal coordination at the tertiary level contrasted with the well-established coordination and feedback systems at the primary level in which health center staff worked closely with community-level HEWs to follow up with patients. In lieu of formal coordination and feedback mechanisms, providers often took personal initiative to follow up on patients. The few feedback systems that existed were characterized by unidirectional communication from the lower levels of the health care system upward, with limited feedback to providers on the appropriateness of referrals and no formal system for following up with patients who had received treatment and were being returned to the community for ongoing monitoring. This lack of feedback and communication inhibited provider learning, further exacerbating the problem of unnecessary referrals. Suggestions for improving coordination centered on increasing cancer diagnosis and treatment resources and training and decentralizing treatment resources to regional levels to ease workloads, better prepare patients for treatment, and reduce referrals. In addition, a latent theme emerged on adding navigators, liaisons, or treatment health educators at secondary facilities to better prepare patients for treatment and reduce the counseling load at tertiary facilities.

Theme 1: scarcity of resources

The most frequent and forcefully expressed perception of cervical cancer care in Ethiopia was one of chronic and severe scarcity of resources. Repeatedly, our respondents told us they lacked the human, financial, and material resources required to coordinate care.

Sub-theme 1.1: lack of time for coordination

The severe shortages of providers limited coordination and teamwork all along the care continuum but was particularly acute at the tertiary level. At tertiary facilities, the dearth of cancer specialists and high patient loads left little time for communication and hindered the formation of stable care teams. The pervasive lack of resources meant that providers did not have the time or sufficient staff to develop and maintain formal coordination systems such as patient navigators and case managers.

There are two main things that I think hamper teamwork. The first one is patient load because when there are a lot of patients, it is difficult to have time to coordinate things in a focused manner, as expected. …. And the second factor that inhibits this teamwork is a shortage of human resources adequate to provide service to the caseload. There needs to be an adequate number of specialists, physicians, oncology nurses, radiotherapy technicians, operating physicians, etc., but that is not the case. As a result [of the lack of staff], there is a long queue, and the patients have to wait a longer time to get their treatment, and it makes it difficult to have a system to coordinate care and teamwork (Interview, Oncologist- 7).

For example, at the time of our interviews, only five oncologists and two radiotherapy technicians at TASH were responsible for providing chemotherapy and radiation services for all cancer cases referred to the facility. The radiotherapy technician shortage was repeatedly highlighted as crucial as they were responsible for ensuring the function of radiation equipment. If one of the technicians was absent, for example, in a coordination meeting, and the machine malfunctioned, no patient care could occur.

Yeah, the limitation in coordinated teamwork [is] the number of oncologists and nurses. For example, as an oncologist, I can say our numbers are few but, nonetheless, not decreasing, but the radiotherapy technologists, however, used to be three, and now they are only two. So, what I am trying to say is that professional distribution sometimes does not go the way we want it and hampers coordination (Interview, Oncologist- 5).

The severe constraints at the tertiary level also made communication with referring providers at the secondary level difficult. OBGYNs at referring facilities said they rarely tried to follow up with patients at tertiary facilities because these providers were too busy. As one OBGYN stated, “Even if I called, things would not change because these hospitals are overburdened” (Interview, OBGYN- 8).

Theme 2: lack of formal coordination mechanisms

Due to resource scarcity, our respondents report that there were few formal care coordination mechanisms within or between tertiary and secondary facilities. Providers recognized the importance of referral networks, coordination, and teamwork for quality of care and were aware of their lack of coordination mechanisms. They felt this lack of coordination was stressful for patients and harmed patient-provider relationships.

There is no official mechanism for us to follow up on the women we refer to get screened for cervical cancer (FGD- 2, Midlevel Providers, R5).

It [care]is not coordinated… for example, she comes to me, and I do the surgery, and I am done when she goes to Oncology, then I have no role anymore. And that is very hard for the woman and has a negative impact on patient psychology because it took her time to get used to me to trust me, and she has given me her body for surgery and the same with the anesthesiologist. And after trusting and getting used to us, we give her to another doctor, and all contact with her stops, and we have no contact with the oncologists afterward (Interview, OBGYN- 6).

Ideally, once a patient is confirmed to have cervical cancer, there is a multidisciplinary team, or what we call a tumor team, that looks at the patient and the test results, comes together to look at and discuss the case and make a consensus on the treatment and the sequence of steps that should be undertaken. This is the procedure of how things are done in more developed countries; it is done as a team once a patient is diagnosed. This is not a common procedure practiced here in the Black Lion Hospital (Interview, Oncologist- 7).

As Fig. 1 indicates, the relationship between poor coordination and resource constraints is bi-directional, with resource constraints hampering coordination and poor coordination exacerbating resource constraints because patients become lost in the system and receive duplicative care.

The development of care coordination mechanisms gained pace during the data collection period, and by late 2020, respondents were reporting that TASH was holding interdisciplinary team meetings to review some complex cancer cases and help prioritize patients needing treatment. Electronic patient records were also being used for some cancer treatments.

Sub-theme 2.1: unidirectional referral communication

The resource constraints, centralization of cancer treatment services, and the lack of coordination and feedback mechanisms meant that communication about patients was unidirectional in the health system, with lower-level providers communicating regularly with higher-level providers but rarely receiving feedback on patient care.

So, you can imagine they cannot give so much emphasis to writing feedback on all of those cases. Having one center for 110 million people, you can’t expect feedback and expect the providers to give all this coordinated service (Interview, OBGYN- 8).

The system at primary care health centers was the exception to one-way communication and a lack of feedback. Here, there was a written feedback card that was either left in an envelope in screening rooms for HEWs to collect on their weekly rounds or brought by midlevel providers to health posts during community outreach sessions. HEWs viewed this system positively and valued it.

So, when we refer a woman to come here to get screened, the feedback is kept in the pocket that is hung on the wall of the screening room. And we do a round to collect our feedback from each pocket in the various rooms once a week (FGD- 3, HEWs, R2).

In other cases, the nurses themselves bring it [written feedback] to us and tell us that they have received a certain number of referrals from us as general feedback (FGD- 3, HEWs, R5).

Sub-theme 2.2: lack of learning

Midlevel providers at health centers and secondary hospitals voiced frustration at the lack of feedback they received on referred patients because it hampered their learning, preventing them from learning if their referrals were appropriate or completed. While feedback was valued for all patient referrals, they were seen as particularly important for cervical cancer screening referrals because most providers were newly trained in screening, relatively inexperienced, and unsure about what conditions required referral.

There is no feedback from the doctor, and we will never know what kind of care she was provided. We don’t know whether she got surgery, a polypectomy, a hysterectomy, or the like, and it is important to get feedback to see if there are any gaps. I try to improve the care provision that I provide by going to the Ghandi Hospital and trying to get feedback. But if you don’t have feedback and accountability, how can you improve the gaps that may exist? (FGD- 4, Midlevel Providers, R3).

You see, feedback is important for many reasons, you want to know how accurate your diagnosis was, and there is no one who checks on the necessity of referrals made. And referral of cervical cancer, in particular, does not have review meetings like other cases. There is nobody who asks, “What are the challenges of cervical cancer screening providers who are trained to provide this service?” (FGD- 4, Midlevel Providers, R3).

Participants reported that receiving feedback motivated and incentivized them to continue their efforts to educate, screen, and treat patients.

…[the] most important incentive is getting feedback on the patient that you refer. Especially to hear that the patient you screened and referred has gotten the proper care and treatment and has recovered from their illness. That is success (FGD- 2, Midlevel Providers, R6).

Sub-theme 2.3: personal initiatives to give and receive feedback

As seen in the sub-theme on lack of learning, lower-level providers wanted to know whether their referrals were appropriate or completed, and many used informal channels to gather this information. An indication of how much providers valued feedback was provided in the statements that, particularly at the community level, HEW and midlevel providers often used their personal phones or made home visits to contact providers and patients. They also used their personal networks to get advice on complicated cases and to arrange referrals. This ad hoc, informal system could only be used by providers who had personal connections with providers at referral facilities and with women who were willing to share their personal contact information.

I do this [follow up] personally using my own cellphone, and with regards to the feedback, I check with her again, and she tells me whether the patient got the required service and tells me things in detail. So, I do this personally (FGD- 4, Midlevel Providers, R2).

So, for example, as [participant] mentioned, [she] knows the nurse there and has her telephone number, so it is easy to communicate, but that is not the case for most providers, even if you may know the nurses at the referral hospital you may not have their number, and you will not call them asking for progress, there is no such system, so it is very rare that happens (FGD- 4, Midlevel Providers, R1).

Theme 3: centralization of diagnosis and treatment services

One of the immediate impacts of the health system’s lack of human and financial resources was that specialized cervical cancer care was centralized in the capital at TASH and St. Paul’s Hospital. Respondents told us that centralizing cancer care specialists and equipment in the capital city exacerbated the challenges of coordinating care. For example, several specialists at TASH reported that centralization impeded pre- and post-treatment care communication with providers at secondary facilities because it meant they had few peers with whom they could collaborate meaningfully. Moreover, patients needing chemotherapy and radiation were referred from facilities nationwide with widely differing capacities, making it difficult for oncologists to communicate with the vast number of referring providers. Once patients completed treatment, providers at tertiary facilities had no way to follow up on their post-treatment care.

You see, in other countries, for example, the oncologist and the gynecologist make decisions together, but that is not the case here in our country. The reason for that is that the gynecologist in Gondar cannot work with me [in Addis Ababa], and this is because we are very few, and we cannot do that (Interview, Oncologist- 5).

First of all of them [patients] come from very different hospitals, and secondly, our referral system is limited to writing a feedback form, but there is no system where you can check where they really go, so maybe that person instead goes to traditional healers, and we do not check whether they went or not (Interview, Oncologist- 7).

The lack of specialists at secondary facilities also meant that patients could not receive counseling and education before they were referred to TASH. They often arrived without fully understanding their diagnosis, prognosis, or care plan. Several providers noted this lack of preparation and the additional strain it placed on the small number of specialists at TASH to educate and counsel patients.

An example of when the providers undermine the situation is when they tell them [patients] that there is a growth, and they will just write a referral and tell them that they will get treated when they arrive at the Black Lion Hospital. So, when they come to us, they do not know if they have cancer and, at times, what cancer is at all. So, we have to explain to them from scratch, and that is very difficult (Interview, OBGYN/Oncologist- 12).

Time is very important for patients, and they need to be given time to discuss their case, [for us] to inform them on the status and provide them with the necessary understanding of the treatment as well as provide them with counseling, which also includes supporting their family is necessary. And this is not easy to do considering the patient load, where you give enough time to talk about things in detail and answer all of their questions (Interview, Oncologist- 7).

Sub-theme 3.1: numerous unnecessary referrals

Centralization also caused providers to unnecessarily refer patients because their facilities lacked resources to care for patients locally. For example, although health centers were tasked with cervical cancer screening under the Ministry of Health’s strategy, in practice, providers told us that few had the capacity to perform screening, and that they often referred patients to secondary facilities for screening and diagnosis services. In addition to unnecessary screening referrals, respondents also noted unnecessary treatment referrals. For example, gynecologists at secondary hospitals noted that patients requiring relatively simple treatments, such as loop electrosurgical excision procedures (LEEP) or straightforward surgeries, are often referred because the hospital does not have the resources to treat them locally. Even basic screening services were often referred:

There are times when the women are referred from a health center. There are trained nurses at the health center, but usually, they may not have the cryotherapy machine, and also they might not provide VIA, so they refer them to us (Interview, OBGYN- 9).

Explaining why he made referrals for surgeries that could be performed at his hospital, one OBGYN noted resource shortages that forced the triaging of operating table space.

The cervical cancer operation takes a minimum of four to five hours, while a C-section takes about 25 min. So, in a setup like ours where there are one or two operating tables, if you use it for that long very often, you will not be able to provide emergency surgeries for mothers needing immediate C-sections or women having bleeding and the like. And there are also only two gynecologists, and there will be no one to help others for that long. We can do so many other operations in the time it takes to do this one operation on cervical cancer (Interview, OBGYN- 8).

Providers told us that many of these referrals could have been avoided if lower-level facilities had sufficient staff and equipment. Centralization, therefore, seemed self-reinforcing. It maintained shortages of expertise and resources at lower healthcare system levels, leading providers there to refer people up the care chain, further centralizing care, and straining resources at tertiary facilities. These referrals were not only inefficient, they also burdened patients and harmed their care.

And what this [unnecessary referrals] means is that they may die while awaiting their treatment. That is what is happening (Interview, OBGYN/Oncologist- 12).

Theme 4: recommendations for change

Sub-theme 4.1: decentralize services

The most strongly and consistently voiced recommendation for increasing coordination and improving the quality of care was to decentralize cancer screening, diagnostic, and treatment staff, equipment, and training to primary and secondary facilities. For example, the HEWs in our study said that they wanted screening and cryotherapy services to be offered at primary health centers as the MoH planned. They also recommended that health officials better incorporate cervical cancer screening into other routine reproductive health care such as post-abortion care and family planning visits.

As has been mentioned, the screening and treatment for cervical cancer should also be decentralized to the lower level health facilities like that of health centers….You see, most of the time, the women we refer to this hospital do not come to get screened because they find it to be too much of a hassle when they even think about it because it’s far for them. And when they think of a hospital, they are reminded of the long queues. But you see, the people in the community can go to the health centers much more easily, and they go there for other reasons. It would be easy to convince them to get screened because they can just do it when they go to get other services (FGD- 1, HEWs, R5).

OBGYNs and cancer specialists recommended the rapid implementation of Ethiopia’s planned regional cancer treatment facilities. They also recommended developing systems to enhance coordination and communication between cancer centers. Respondents argued that decentralization of cancer treatment would empower local providers and improve patient experience and continuity of care while reducing fragmentation of care across different facilities or multiple units within crowded secondary and tertiary hospitals; although they also noted that having such facilities would increase, rather than decrease the need for formal coordination mechanisms. Expanding cancer care facilities was understood to require political will and policy prioritization that many saw as lacking.

And if there were [cancer] centers that provide similar care around the country, the providers there could call and consult the senior oncologists when they have a complicated case. I think that would be a good system (Interview, OBGYN- 9).

If such a cancer center existed, it would provide care only for this illness [cancer], so the patients will not be mixed with other patients, they won’t have to go to different departments, and they would get their treatment in one place from A to Z. That is what is recommended…(Interview, OBGYN- 6).

This is something that requires national coordination, particularly in relation to the referral of patients that we were talking about earlier. Currently, there are 5 or 6 cancer centers in different parts of the country that are being made ready to provide the services we provide. And to avoid or reduce the existing mismanagement of patients and their time, there needs to exist a system of coordination among these centers…. So that a patient in Gondar, without ever needing to come to Addis Ababa, can be treated there (Interview, Oncologist- 7).

Respondents had several suggestions for using decentralization to strengthen care coordination specifically. They suggested building on the strong connection between HEWs and community members to improve collaboration at the community level. Respondents suggested that HEWs and Women’s Development Army volunteers could be incorporated into referral and feedback mechanisms to provide patient navigation and follow-up services. In community-based navigation programs, HEWs could provide palliative care to terminally ill patients, follow up with patients after cryotherapy and treatment, and provide psychosocial support to the newly diagnosed.

Sub-theme 4:2: provide more human resources, equipment, and training

In addition to more decentralized staff, providers also recommended expanding the supply of treatment equipment and the provision of training to a broad spectrum of providers at all levels of the health system. Respondents consistently expressed the need for more cryotherapy and radiotherapy machines. This sub-theme echoed the earlier sub-theme of unnecessary referrals in which providers, particularly those at primary and secondary facilities, noted that they made referrals not only because they lacked trained staff but also because they did not have the equipment and space to provide care.

WHO indicates the need for at least one radiotherapy machine for one million people, so in Ethiopia, we need 100 radiotherapy machines to provide timely care and make an impact (Interview, Oncologist- 7).

In contrast, providers at TASH prioritized hiring and training staff (particularly radiotherapy technicians) over providing equipment because they believed trained personnel would create demand for the infrastructure and equipment.

[…]we need to have trained human resources. We need trained surgeons, trained radiotherapists, and the like in the long term because I believe that, more than anything, we need people who can provide the service. Because once we have the human resources, the infrastructure and the machineries will be available at the request of these professionals (Interview, OBGYN- 6).

In addition to increasing training, there was an expressed need to better match training with the provision of equipment so that trained workers would have the tools needed to provide care and equipment would not sit unused due to a lack of trained staff. Providers also recommended that training, particularly screening and cryotherapy training, be provided broadly to all relevant staff at health centers and primary hospitals and that “screen and treat” service provision responsibilities be rotated through different facility staff. Doing this, they argued, would ensure all staff could practice their skills and prevent missing staff from creating care delays. The desire for more training was related to the desire for more supportive supervision and follow-up after training to identify and address service delivery challenges services. In supporting and evaluating staff, respondents noted that post-training evaluations should be based on the quality of care provided, not only the number of patients seen.

The other thing I want to emphasize on what she has said is that, yes, there are trained providers in the different health centers, but that does not mean they have the treatment equipment available. So there is no use if they are trained, but the government did not avail the cryotherapy machine, and I do not think the problem only has to do with the unavailability of space, and I don’t think the problem is only limited to the weakness of the providers (FGD- 4, Midlevel Providers, R2).

I think the care will be improved in general if all the providers are trained to give the service. You see, if the person that provides the service does not come [to work] for some reason, let’s say if she was sick, there will be more people that can provide the service if everyone is trained (FGD- 3, HEWs, R5).

Sub-theme 4.3: liaisons/navigators at secondary hospitals

A latent theme that emerged from provider discussions of poor patient preparation for treatment and long wait times was the desirability of having liaison staff at secondary hospitals to track and communicate patient progress and educate and counsel patients for referral. Patients could enter the treatment queue at TASH while at these secondary facilities, reducing their wait time in the capital, and be triaged in this queue so that women without advanced cancer could get care immediately before their cancer progressed. Moving counseling and health education to secondary facilities would free staff time at TASH, reducing waiting times, which providers saw as an urgent problem, noting that women “often died waiting their turn to get treatment” (Interview, OBGYN 8).

Time is very important for patients, and they need to be given time to discuss their case, [for us] to inform them on the status and provide them with the necessary understanding of the treatment as well as provide them with counseling, which also includes supporting their family is necessary. And this is not easy to do considering the patient load, where you give enough time to talk about things in detail and answer all of their questions…. It would be good if it could be possible to give them the time to talk about everything and to calmly ask questions and understand their status and the treatment (Interview, Oncologist- 7).

Discussion

Studies have shown that after screening, patients with abnormal results in Ethiopia often experience long referral times, severe treatment delays, and poor communication with providers regarding their diagnosis and treatment plans [12, 24, 25]. This study examined the role that care coordination could play in reducing these barriers by describing provider experiences of care coordination.

Our primary finding is that there is a lack of formal cervical cancer care coordination mechanisms and practices in Ethiopia, which results in extended referral times, unnecessary referrals, and unidirectional communication between facilities. The most robust coordination occurred between HEWs and midlevel providers at the primary-care level. In contrast, only nascent coordination mechanisms existed within or between secondary and tertiary facilities. Poor coordination was driven primarily by Ethiopia’s severe resource constraints, specifically the lack of trained staff, space, and diagnostic and treatment equipment. A lack of resources prevented coordination directly by limiting the time available for coordination and indirectly by promoting the centralization of treatment in a single non-specialized facility where communication with referring facilities was particularly challenging.

Weaknesses in cancer care coordination and referral systems and their negative impact on health outcomes have been previously documented in SSA and remain an ongoing challenge [2, 14, 21]. In a review of 57 African studies on late presentation to cancer care, inadequate care coordination was a consistent reason for delays in patients reaching cancer care services and receiving quality care once at a healthcare facility [5]. Late care initiation, in turn, is the primary cause of low survivorship among cervical cancer patients in Ethiopia [6]. Infrequent, non-standardized, and unidirectional communication between referring clinicians is a documented and longstanding global challenge in cancer care, including in high-income countries [21, 32, 35].

The emphasis that respondents in our study placed on resource constraints as the main barrier to coordination is also to be expected, given the well-documented, acute shortage of healthcare professionals and poor health infrastructure in Africa [36, 37]. The continent’s 54 countries are serviced by only 3% of the global healthcare workforce and have approximately 25 physicians per 100,000 people [38]. Oncologists and other cancer care specialists are in short supply, and in the short term, increased screening is likely to exacerbate this workforce shortage by increasing demand for services, intensifying the documented burnout and de-motivation among Ethiopia’s cancer providers [14]. Workload and lack of time have been found to be barriers to care coordination in health systems globally [39,40,41].

Policy implications and recommendations

Resources

Despite their challenging conditions, we found that respondents valued collaboration and attempted to coordinate with each other informally when possible. They had thoughtful suggestions for improving care coordination. Given the intensity of health system resource scarcity, it is unsurprising that these suggestions centered on significant staffing, training, and equipment increases as the primary means of improving coordination. Although this general suggestion was consistent across providers, we note that those at lower levels of the health system placed greater emphasis on obtaining equipment and space so that personnel had the resources to implement services. In contrast, those at higher levels of the system emphasized the need for more staff, particularly specialist staff, arguing that once in place, staff would advocate for the infrastructure they needed.

While our findings suggest a clear need for additional resources to stimulate Ethiopia’s coordination efforts, resources are not sufficient for the development of effective cancer care coordination systems. Care coordination is a persistent challenge even in well-resourced health systems, and there is a large variation between and within high-income countries on coordination performance [42, 43]. High-income countries often perform no better than their low-income counterparts on key coordination metrics despite having adopted initiatives such as patient navigators and cancer care coordinators [44, 45]. Poor communication and lack of understanding of roles and responsibilities are common challenges to cancer care coordination in high-income settings, particularly between primary care clinicians and oncologists [46,47,48]. The mixed experience of high-income countries in coordinating care suggests that rather than simply adding resources to Ethiopia’s existing care structures, making ongoing systems- and practice-level policies will be necessary, particularly with regard to data access roles and responsibilities. Ethiopia may be in a unique position to adopt such changes. Because its health system is young and rapidly growing, coordination initiatives can be embedded in new systems, facilities, and professional training plans rather than being layered upon or inserted into long-established, complex systems of training, record-keeping, communication, and care. In addition, the lack of fragmentation in Ethiopia’s health system increases the feasibility of broadly sharing information. For example, Ethiopia’s unitary nature of health system decision-making and record-keeping would allow the adoption of a single national electronic health records system and national policies for allowing primary care providers broad access to oncology records.

Decentralization

Other than the general call for more resources, the most frequently and forcefully expressed recommendation for improving care coordination and quality of care was decentralizing diagnostic and treatment resources within Ethiopia to provide care at regional, specialized cancer hospitals and secondary and primary facilities. Inequitable internal distribution of diagnostic and treatment resources has been noted as a problem in other African settings [21], and several studies have documented the challenges Ethiopia has faced in providing screening and preventative treatment at the health center level [12, 49, 50]. Decentralization of cancer treatment services is already underway in Ethiopia with the creation of regional cancer centers, and our findings point to the urgency of this task.

Secondary and regional-level facilities emerged as critical points in the care continuum where there are the fewest reported coordination mechanisms and the most opportunities to improve collaboration. Most discussions of gaps in feedback and follow-up communication in our study concerned referral to and from secondary-level facilities. Participants at primary facilities rarely received information on the appropriateness and completion of their referrals to secondary facilities, providers at secondary facilities received no feedback from TASH on referred patients, and providers at TASH noted that patients often arrived unprepared for treatment requiring intensive counseling and education on site. Previous studies of cancer care in Ethiopia have reported this latter finding [14, 24] and have noted the significant challenge of preparing rural women and their families for treatment. In addition to reducing the strain on cancer specialists and the logistical burden on patients, improving the care capacity at lower-level facilities could also reduce the high prevalence of misdiagnosis among cervical cancer patients, a significant source of care delay in Ethiopia [5, 12].

Navigators

While Ethiopia’s creation of regional cancer centers may mitigate coordination problems at the secondary/regional level of care in the long term, several shorter-term solutions emerged from our respondents’ discussions of their care duties. A key solution was providing patient navigators and counselors at secondary facilities who could educate and counsel patients on what to expect during treatment. Because secondary facilities still provide most diagnostic and cryotherapy services, navigators at secondary facilities might also be able to coordinate with primary care facilities on follow-up post-screening diagnostics and secondary preventative treatment plans. Having patient navigators within TASH and new cancer treatment facilities was also suggested by our respondents.

Employing patient navigators is a well-known and highly effective strategy that increases the timeliness of cancer care and improves treatment completion rates, retention in care, and access to supportive and palliative care [21, 51]. Navigators have been found to be acceptable and effective in other African settings for cancer care [21]. Ethiopia is particularly well placed to build community-based navigation and patient support systems for cancer care because it has a solid community-level outreach infrastructure staffed by HEWs. More robust counseling and patient education are vital for cervical cancer care because low cervical cancer knowledge and the high levels of stigma and fear associated with it are associated with delays in care [5, 12, 52].

Our participants’ recommendations to shift counseling and patient education tasks to the primary and secondary levels of care and increase diagnostic and treatment capacity at lower-level facilities align with international calls to action to employ task shifting to scale up cancer care in LMICs [37]. Such task-shifting strategies have been used extensively and successfully in Ethiopia for communicable diseases, reproductive health, and maternal and child health care to mitigate the shortage of specialists and reduce delays in care due to distance [53,54,55,56]. Task-shifting initiatives that integrate health education, the promotion of screening and vaccination, and patient navigation for follow-up services could improve treatment outcomes and bolster prevention initiatives by showing the life-saving potential of early tumor detection [10]. In developing task-shifting and navigation programs, Ethiopia will have to be mindful of the challenges faced by high- and low-income countries in implementing such initiatives, in particular, confusion about roles and responsibilities and unplanned role expansion [57, 58].

Limitations

Recruitment into this study was disrupted due to the COVID- 19 pandemic and political violence. The first impact of this was that our data collection period was extended, and changes to care delivery systems occurred during this period. This complicated our analysis and made it difficult to assess the extent and types of care coordination occurring at any particular time during the study period. The second impact was that we could not speak to healthcare workers in rural areas. Our findings reflect the experiences of providers at facilities in towns and cities. The challenges of communicating with busy providers during the study period also meant that we were not able to give participants the opportunity to provide feedback on our analysis. The study’s strength is its attempt to describe coordination experiences across the care continuum with various providers.

Conclusions

Due to expanding populations and demographic shifts, LMICs will have a growing proportion of cervical cancer cases in the following decades. Without concerted action, millions of women already living with cervical cancer will endure profound physical and emotional suffering. Preventing this suffering and combatting low survival rates will require targeted infusions of resources, adopting proven strategies such as care coordination, and local innovation to develop new interventions. The history of the global response to HIV/AIDS shows that with creativity, even severe health system weaknesses can be mitigated, allowing providers to deliver complex, long-term services. Decentralizing diagnostic and treatment resources, advocating for better access to drugs and equipment, exploring task-shifting for screening, counseling, and patient education services, increasing human and technical resources, and creating formal coordination support mechanisms at all health system levels are crucial strategies for improving cervical cancer outcomes while retaining the essential workforce. We find an abundance of support for such initiatives from Ethiopian providers.