Health System Factors That Influence Treatment Delay in Women With Breast Cancer in Sub-saharan Africa: A Systematic Review


 BackgroundBreast cancer patients in sub-Saharan Africa experience long delays between their first presentation to a health care facility and the start of cancer treatment. The role of the health system in the increasing delays in treatment has not been widely investigated. This review aimed to identify existing information on health system factors that influence treatment delays in women with breast cancer in sub-Saharan Africa to contribute to the reorientation of health policies in the region.MethodsPubMed, ScienceDirect, African Journals Online, Mendeley, ResearchGate and Google Scholar were searched to identify relevant studies published between 2010 and July 2020. We performed a qualitative synthesis in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. Related health system factors were extracted and classified according to the World Health Organization’s six health system building blocks. The quality of qualitative and quantitative studies was assessed by using the Critical Appraisal Skills Program quality-assessment tool and the National Institute of Health Quality Assessment Tool, respectively. In addition, we used the Confidence in the Evidence from Reviews of Qualitative Research tool to assess the evidence for each qualitative finding.ResultsFrom 14,184 identified studies, this systematic review included 28 articles. We identified a total of 36 barriers and 8 facilitators that may influence treatment delay in women with breast cancer. The principal health system factors identified were mainly related to human resources and service delivery, particularly difficulty accessing health care, diagnostic errors, poor management, and treatment cost.ConclusionThe present review shows that treatment delay among women with breast cancer in sub-Saharan Africa is influenced by many related health system factors. Policy makers in sub-Saharan Africa need to tackle the financial accessibility to breast cancer treatment by adequate universal health coverage policies and reinforce the clinical competencies for health workers to ensure timely diagnosis and appropriate care for women with breast cancer in this region.

In sub-Saharan Africa, delays in diagnosis and treatment and barriers to care experienced by breast cancer patients have been previously explored [6,10], and many ndings show that patients experienced long delays between initial symptoms and presentation to a health care facility (patient delay) and between rst presentation and de nitive diagnosis or treatment (provider or health system delay) [6,10].
Most studies conducted among sub-Saharan African women were focused mainly on patient delay factors (sociodemographic, cultural and economic related factors), early detection [11,12], and knowledge, attitudes, and practices of women regarding breast cancer and self-examination [13,14].
In addition, most of the review's studies conducted in the sub-Saharan Africa region were focused on time to presentation, diagnosis and related factors and stages of diagnosis [11,12]. However, the role of health system-related factors in treatment delays (from presentation to rst treatment) has not been well investigated. To the best of our knowledge, this is the rst systematic review to focus on health system factors that may explain treatment delays in women with breast cancer in sub-Saharan Africa. The results from this study will contribute to the improvement of breast cancer health policies in this region.

Methods
This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement by Moher et al [15].

Study selection and data collection process
Zotero reference manager software [16] was used to organize and detect duplicate references. We identi ed eligible articles by using the PRISMA ow diagram. The rst and second authors independently screened all titles and abstracts identi ed by the search, and those clearly irrelevant to the topic were excluded. The full texts of all potentially eligible papers were retrieved and reviewed for inclusion in this review according to the inclusion criteria. All included studies were independently reviewed by two authors to con rm eligibility (GG and MB).

Data extraction and items
For the included studies, two authors (GG and MB) independently extracted information such as the characteristics of the study (title, authors, year of publication, country, study design, research method, age group, participants, and sample size), health system factors (barriers and facilitators), and the time to treatment (if available).
All health system factors were classi ed according to the World Health Organization (WHO) Health Systems Framework's six building blocks [17] as follows: 1) health service delivery, 2) health workforce, 3) heath information systems, 4) access to essential medicines and technologies, 5) health system nancing, and 6) leadership and governance. Any discrepancies in the process of selection and extraction were resolved through discussion, if necessary, with two other authors (MK and IH).
The treatment delay is often de ned as the interval of time between the onset of breast cancer symptoms and the initiation of therapy [18]. In this review, we focused on the treatment delay from the rst presentation to the start of treatment. This period is often divided into two time frames: the rst is the period between the rst presentation to the health care provider and the con rmation of the diagnosis of breast cancer, and the second is the period between the con rmation of the diagnosis and the start of treatment [9,11,19].

Quality assessment
The quality of the qualitative studies was assessed by using the Critical Appraisal Skills Program (CASP) quality-assessment tool (http://www.casp-uk.net) [20]. The quality of the quantitative studies was assessed using the National Institute of Health (NIH) Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies [21]. Study quality was assessed according to the following criteria: research question, study population, eligibility criteria of the population, sample size justi cation, outcome measures, response and follow-up rates, statistical analyses, and ethical issues. In addition, we used the Con dence in the Evidence from Reviews of Qualitative Research (CERQual) tool to assess the evidence for each qualitative nding. Based on the assessment of four components (methodological limitations, relevance, adequacy, and coherence), all qualitative ndings were classi ed into three categories: high con dence, moderate con dence, or low con dence [22].

Study characteristics
The main characteristics of the included studies are summarized in Table 1. Among the 28 studies included in the review, 11 were quantitative, 12 were qualitative, and ve were studies with mixed methods (quantitative and qualitative approaches). They were conducted across 13 countries in sub-Saharan Africa. Thirteen (46%) studies were conducted in East Africa, 10 (35%) in West Africa, 3 (10%) were conducted in Southern Africa and 2 (7%) were multicountry studies mixing countries from three parts of sub-Saharan Africa (East, West and southern Africa). The study publication dates ranged from 2013 to 2020. The sample sizes of the studies ranged from 64 to 1429 for the quantitative and mixed studies. Eleven (69%) quantitative and mixed studies were cross-sectional surveys, while three were cohort studies. The qualitative articles used focus group discussions and in-depth interviews to explore different factors in uencing treatment time in breast cancer patients. In 21 (75%) studies, their study populations exclusively included women with breast cancer, whether newly diagnosed or not, whereas the seven other studies included physicians, health care workers, family members and women without breast cancer. The average (mean/median) age of the women with breast cancer was 40 and over in the large majority of studies.

Factors in uencing time to treatment in breast cancer patients
There were a total of 36 barriers and eight facilitators identi ed across all studies. Factors identi ed in each study and classi ed according to the WHO Health Systems Framework's six building blocks are summarized in Table 2 and Table 3, respectively. Among the barriers, misdiagnosis was the most common (16 studies), followed by provider attitude (11 studies) and the high cost of investigation and treatment (11 studies). The appearances of other barriers and facilitators ranged from one to seven times.

Health service delivery
Health service delivery was addressed by 23 of the studies included in the review, and the factors identi ed can be grouped into two major themes: logistic and infrastructure.
Logistical di culties included different waiting times for an appointment (medical or specialist consultation), investigations (imaging, biopsy), test results, referral or treatment (surgery, radiotherapy) [23][24][25][26][27][28][29][30][31]. A long waiting time for test results, more speci cally for the biopsy results, was a factor in uencing the time to treatment in 11 included studies [24,[26][27][28][29][30][32][33][34][35][36]. Seven included studies reported that referral was delayed for women who had visited other health facilities 1 to 4 times or more before visiting the diagnostic centre [23,24,30,[36][37][38][39]. In one of these studies, for instance, more than 80% of 499 patients had at least 3 visits before diagnosis, and repeat visits in the referral system appear to have been caused by misdiagnosis and delays in appointments and test results [24]. Pace et al. found that patients who visited other healthcare facilities ≥5 times before diagnosis were more likely to experience system delays of ≥6 months (OR, 2.69; 95% CI, 1.24-5.84; p = 0.01) [37]. Studies have also shown that the type of healthcare facility and the type of health worker visited at the rst consultation had an impact on diagnosis and treatment delays [39,40], and women who rst visited a community health centre or general practitioner experienced longer delays than those who rst saw a specialist (surgeon or oncologist) [39,40]. Being referred by an oncologist or surgeon or having received the rst consult evaluation in a public hospital have been identi ed as factors facilitating diagnosis and access to treatment [39,40]. Logistic problems also included poor organization and unavailability or shortage of breast cancer services [30,31,41].
The geographical inaccessibility of healthcare facilities and the insu ciency of diagnostic centres are part of the infrastructure problems [28][29][30][31]41]. In Uganda, for instance, one patient interviewed said, 'The health centre nearby the community does not offer screening services, and someone may nd it hard to leave this place [Ssisa sub-county] to go to Kampala; but if they bring the services closer to the community, some will nd it easier to visit them.' [Semistructured #2] [41].

Misdiagnosis, Misinterpretation and Mismanagement
Among health workforce factors, misdiagnosis was the most common and appeared in 16 studies [24, 25, 29, 30, 32, 34-36, 38-40, 42-46]. Patients reported being inappropriately reassured by health workers that their breast lump was benign without a biopsy or with an incorrect biopsy interpretation.
"… I noticed something, a small lump on my breast … I woke up in the morning and went to my doctor.He told me it could be a tumour.I asked him if it could be a cancer because I heard about it on TV.He told me it is not a cancer." (P07) [29] Symptom misinterpretation and misdiagnosis were the most frequent reasons for prolongation of the primary care interval. In six studies, caregivers mismanaged breast cancer patients by giving incorrect medical prescriptions or incorrect advice [27,28,30,36,43,44]. For instance, Pruitt et al. reported that the majority of women described receiving oral medication or injections, usually antibiotics, sometimes for months or years before being referred or making an independent decision to seek care elsewhere [27].
'I went to the Referral Hospital in XXX. They gave me drugs which I took for 3 months. During that period, I did not see any amelioration…3 months later I went back to my doctor and he gave me other prescriptions for blood analyses…' (Patient 7) [44].

Poor knowledge and skills
Misdiagnosis is often attributed to poor knowledge and lack of health workers training about breast cancer; most health workers provided incorrect information to patients [30,32,35,41]. Some of them were unable to examine the patient appropriately [25]. As an illustration, in one focus group, a health worker said: 'For me, I have never gotten training on breast cancer detection but I just hear that breast cancer is very dangerous and it is good for someone to go for check-ups but I have never received training on breast cancer examination.' [FGD CHWs #7] [41].

Attitude of health workers
The attitude of providers was also an important factor in uencing women's access to treatment. Some studies reported that poor attitudes and corruption among health workers were factors that accounted for delays in the start of de nitive treatment. Patients also said some health workers disrespect them, refused to answer their questions or did not treat them well, which forced them to delay the start of treatment [28], while others had good communication with health care providers, which made it easier for them to manage their breast cancer [46].

Strikes and shortage of caregivers
The shortage of health workers and strikes were also identi ed as factors delaying women's access to breast cancer treatment [25,27,28,32,36,42,44]. Studies have reported that various hospital departments turned women away because a doctor was not available [32]. Women faced system delays in receiving their test results and had di culty accessing care due to strikes by various members of the healthcare team, including consultants and residents [27].

Health nancing
Six key elements were identi ed among nancing-related factors: high costs of treatment and investigations, lack of insurance or limited insurance coverage, expensive private insurance and discrimination by private insurance. The high cost of treatment was reported as an important factor in uencing women's access to breast cancer treatment in 11 studies [27-31, 38, 46-50]. One woman in the study reported by Sanuade et al. said, 'The chemo is expensive. The trauma and money you spend is a problem too. If you do not have at least 200 Ghana cedis, you cannot buy the drugs. When someone hears all this, the individual would opt for herbal medicine or prayer. So, as for me, I think that lack of money is a factor. I paid 1000 Ghana cedis to use the chemo machine. If you do not have money you would go home. So, money is a serious factor.'' (FGD 4-R1) [28].
Lack of or limited insurance coverage was also mentioned as a barrier to care. In the study conducted by Subramanian et al. in Kenya, 78% out of 400 women with breast cancer reported borrowing money from family or friends to cover out-of-pocket medical and related expenses [47]. In another study conducted in the Ivory Coast, 36% out of 126 patients declared having had a delayed diagnosis due to lack of nancial resources [34]. Many women reported having no insurance coverage [47][48][49]. For instance, Okoronkwo et al. found that 71.8% out of 194 patients studied in Nigeria did not have health insurance coverage [48].

Medication access and technologies
Unavailability of drugs and equipment failure were the most common factors among those related to medication access and technologies [28,30,32,44,46,47]. Diagnostic assessments were not available in small health centres, and multiple visits were required for X-rays or blood tests or to obtain a biopsy sample, according to interviewees. Various hospital departments turned women away because the computer was broken or the X-ray machine was not working [28,30,32]. Study results also identi ed drug shortages as an important factor in treatment delay [32,44,46].
For instance, key informants in one study conducted in Uganda reported that the existing health system is not equipped to manage breast cancer [41].

Leadership and Governance
Lack of cancer policy, low prioritization of non-communicable diseases, and lack of provider training on breast cancer were the common factors identi ed in the included studies [41]. Key informants in one study conducted in Uganda highlighted the lack of cancer policy-providing guidelines for cancer management across each spectrum of the cancer care continuum [41]. 'There is no such policy on cancer screening or cancer prevention; there's nothing like that.' [key Informant #6] [41].
Information System The information system was addressed by only one study among those included in the review [37]. The key factor identi ed was the delay in administrative procedures. In this study, conducted in Rwanda, 27% out of 113 women interviewed said that, to receive public insurance coverage for care provided at a district hospital, a referral form from a health centre was required. Many patients described the need for a transfer form as a reason for the delay.

Quality appraisal of the included studies
The majority of the quantitative studies were rated as good quality based on the NIH study quality assessment tools for Observational Cohort and Cross-Sectional Studies (Supplementary data, Table S1). Most qualitative studies were of high quality on the CASP checklists (Supplementary data, Table S2).
There were no low-quality among studies. Based on the four components (methodological limitations, relevance, adequacy, and coherence) of the CERQual approach, the con dence in the evidence for most of the qualitative ndings (16 of 21) was rated as low. (Supplementary data, Table S3).

Discussion
This review aimed to identify the health system factors that in uence treatment delay in women with breast cancer in sub-Saharan Africa. The qualitative synthesis of studies identi ed 44 factors, including 36 barriers and 8 facilitators. These factors are mainly related to health service delivery, health workforce and nancing, followed by factors related to medication access and technologies, governance and leadership, and the information system.
Poor organization of health service delivery was responsible for increasing the waiting times for different investigations (often the biopsy, consultation or surgery appointment, and referrals). This was mainly attributed to misdiagnosis, mismanagement or misinterpretation and a long primary care interval. The geographical inaccessibility of different services was also found to be an important barrier to care. These results may be a reasonable representation of the phenomenon of interest according to the level of con dence in the qualitative evidence, as assessed by the CERQual tool. However, they are supported by quantitative and mixed studies that were mostly assessed as good quality studies. These results are consistent with those reported in a recently published systematic review by Nathan R. Brand et al. in LMICs. In their review, 92 studies looked at breast cancer, 10 identi ed reduced access to primary care, 6 identi ed limited access to diagnostic services, and 14 identi ed geographic inaccessibility as a factor related to treatment delays [51]. Similar results were reported in studies conducted in Morocco [52][53][54], Palestine [55] and Brazil [56]. These ndings suggest the need for efforts to be deployed by decisionmakers to ensure the availability and quality of screening services and specialized and comprehensive care for breast cancer in sub-Saharan Africa.
The factors related to the health workforce have been identi ed as a major handicap to women's access to treatment in our review. These are mainly diagnostic errors and inadequate care, especially at rst contact with the health system, followed by the attitude of providers and the lack of human resources. According to the CERQual assessment tool, these data provide a moderate level of con dence in the synthesis of the qualitative ndings, and they are also found in quantitative and mixed studies of good methodological quality according to the NIH assessment tool. In addition, it is important to note the variations in some attitudes of health professionals that were identi ed by these studies, particularly in terms of communication, trust, patient information and corruption. These variations would be related to the differences in contexts and experiences lived and reported by women. However, inappropriate diagnosis remains the most common factor in our studies and is also found in other settings. In their critical review, Unger-Saldaña K.
identi ed medical errors in initial diagnosis, screening interpretation and pathology review as factors related to access or quality of care de ciencies that have been associated with breast cancer provider delay in different countries, such as the United States, England Thailand, Scotland, Netherlands, Canada and Mexico [9]. Our results are also in line with those of the review conducted in Africa by Espina et al. [11]These very alarming ndings should arouse enormous interest among professionals in health care and health policy to give importance to initial and continuing medical training in breast cancer.
As a factor related to nancing, this review identi ed the probable in uence of treatment costs. Lack of health insurance or limited health insurance coverage forces women to pay for services out of pocket, which is not always easy for the most vulnerable individuals. These results are consistent with those of a review conducted in the Middle East and North Africa (MENA) region [57] and with a similar study conducted in India [58]. The nancing-related factors identi ed in our review have been reported by mostly high-quality qualitative studies according to the CASP assessment tool and good-quality quantitative studies according to the NIH assessment tool. It is therefore necessary to accelerate the race for medical coverage to ensure adequate and timely care for patients, as stipulated in the sustainable development goals (SDGs) established by the WHO [59].
The remaining three health system building blocks (medication access and technologies, governance and leadership, and information systems) were weakly represented. This is likely the in uence of the lack of health policy and the possible in uence of other factors, such as the lack or shortage of drugs, care equipment failure and the low prioritization of non-communicable diseases. The only factor related to the information system is the di culty of administrative procedures. This factor was reported in one mixed study of good methodological quality. The relevance of factors related to medication access and technologies, governance and the information system remains limited. Therefore, more primary studies on these factors and, above all, adapting health policies to local speci cities should be recommended.
The few facilitators identi ed in this review re ect a variation and inequity in women's access to care between some health structures but also between countries in Sub-Saharan Africa. These facilitators were related to free health care, su cient health coverage and the availability of resources. In a review conducted in the MENA region, having health insurance facilitated access to mammography [57].
Studies that reported these same factors in high-income countries are mostly more than 10 years old [60,61]. This would indicate two major conclusions: rst, our phenomenon of interest is no longer a real problem for them, and second, there has been a clear advancement in the management of women with breast cancer in high-income countries compared to low-and middle-income countries. However, recent studies conducted in these countries have shown inequalities in access to breast cancer diagnosis and care for vulnerable populations such as immigrant women, residents of rural areas and black individuals [62][63][64].

Strengths And Limitations Of This Review
To our knowledge, this is the rst systematic review focused on health system factors that in uence time to treatment in women with breast cancer in Sub-Saharan Africa. The majority of the included studies were of good methodological quality. Our review also has several limitations. First, the non-inclusion of grey literature does not exclude the risk of publication bias. Second, the review included studies from only 13 of the 48 sub-Saharan African countries, and the phenomenon studied could be worse or better for countries in which there are no data available. This reveals the paucity of published data on this topic, limits the relevance of our results and suggests the need for more primary research on the topic in the region. Third, most included studies explored more than 5,000 women with breast cancer at the hospital level, which could constitute a selection bias. In fact, only women who succeeded in accessing healthcare facilities and obtaining a diagnosis were enrolled in the included studies, further limiting the representativeness or generalizability of the data. This limitation has been underlined by several of the included studies. The heterogeneity noted in the de nition and quanti cation of the different delays did not affect the importance of the factors identi ed or hinder our initial objective, which was to identify any factor that could in uence women's access to treatment.
Implications for health policy and health system research Despite the methodological limitations identi ed in some studies, our review ndings suggest the need for developing appropriate breast cancer policies that take into account the reduction of nancial and geographical accessibility barriers, the strengthening of the management of health service delivery to ensure the availability, the quality of timely screening services, specialized and comprehensive care for breast cancer in sub-Saharan Africa. Attention should also be paid to the continuing education, formative supervision for frontline health workers. Our review has also indicated some research gaps such as the importance of exploring the role of medication access and technologies, governance and leadership, and information systems in hindering or facilitating women's access to appropriate breast cancer care in Sub-Saharan Africa.

Conclusion
Our review indicates that treatment delay among women with breast cancer in sub-Saharan Africa is in uenced by many related health system factors. When women manage to overcome their fear, lack of knowledge, socioeconomic and cultural conditions, they also end up being challenged by overwhelming health system factors that they cannot cope with. Our review shed a light on the underlying factors that explain the delays and health system challenges women face in terms of nancial and geographical access to care, diagnosis errors, inappropriate facility management and lack of appropriate cancer health policies.

Consent for publication
Not applicable Availability of data and materials All data generated or analysed during this review are included in this document and its supplementary data.

Competing interests
The authors declare no con icts of interest.