Health Related Quality of Life and Its Determinants Among Breast Cancer Patients in Africa: A Systematic Review and Meta-Analysis

based on 14 modied of life checklist. The pooled estimate quality of life by open Meta with The results of included studies not for meta-analysis The heterogeneity of included studies was evaluated with I 2 statistics.

di cult time of their disease [12]. The aim of this systematic review and meta-analysis was to assess HRQOL among breast cancer patients and to identify factors affecting their QOL in Africa. The ndings of this systematic review and meta-analysis may help policy makers in planning and implementing strategies to reduce factors affecting QOL and improve HRQOL of breast cancer patients.

Study protocol
The review protocol was performed in accordance with preferred reporting items for systematic review and meta-analysis (PRISMA) [13].
Screening and eligibility of studies TB selected the title of the study. Three authors (TB, WA and TB) screened the title and abstracts of the studies based on the inclusion and exclusion criteria. TB, WA and TB also collected the full texts, evaluated the eligibility of the studies for nal inclusion and assessed the quality of the study. TB and WA analyzed the data. EE solved the disagreement between the authors.

Inclusion and exclusion criteria
Observational studies (cross-sectional and prospective observational studies) that assessed QOL in breast cancer patients and those published only in English language without time limit were included. The exclusion criteria applied in this review and meta-analysis was: the study which did not investigate HRQOL in people with breast cancer; the study which did not provide any data about HRQOL among study population; reviewed articles; the study which was not a journal article (for example conference abstracts) and the study which was not done in Africa.

Data source and search strategy
A systematic search was conducted from August 10 to September 5, 2020 to review the studies that evaluated HRQOL among breast cancer patients in Africa without limitation of publication year. International data bases like Medline/PubMed, HINARI, Science direct and Google scholar were used for literature searching. These searching strategies were performed through alone and a combination of medical subject heading terms and free keywords: [("breast cancer" or "breast neoplasm" or "breast tumor" or "breast carcinoma") and ("quality of life" or "health related quality of life") and ("determinants" or "predictors" or "factors of quality of life") and (Africa)]. In addition to this, the reference lists of the reviewed article were manually scrutinized in order to identify and included potentially relevant studies. All published and unpublished articles were searched.

Data extraction
Titles and abstracts of the studies were screened in order to identify all potential eligible studies using prede ned data extraction form. Depending on the data extraction form, the following data were extracted: rst author's last name, year of publication, study location (country), study design, type of data collection method, types of questioner, study setting, sample size, age of participants, and type of score with its measurement domains and determinants of quality of life of breast cancer patients.

Methodological quality assessment of studies
The methodological qualities of included studies were assessed based on a modi ed checklist developed to assess the methodological quality aspect of QOL reporting. It was assessed according to a prede ned 14 item checklist. The 14 items include: two items (sampling), one item (QOL measurement tool selection), two items (data collection method), two items (response rate), one item (group comparison), ve items (reporting clarity) and one item (prognostic factor determination).
A score of 1 or zero was given for each item. A score of one was given for an item if meeting the methodological criteria. A score of zero was given for an item if an item neither met the criteria nor described the related parameter su ciently. A score of ≤ 6(lower than 50% of the maximum attainable score) indicated as low quality. A score 7 to 9 (between 50% and 75% of the maximum attainable score) and ≥ 10 (≥ 75% of the maximum attainable score) indicated as moderate and high quality respectively [14].
List of criteria to assess the methodological quality of studies on QOL of breast cancer patients 1. Socio-demographic and medical data is described (e.g., age, race, employment status, educational status, tumor stage at diagnosis etc.) 2. Inclusion and/or exclusion criteria are formulated 3. The process of data collection is described (e.g., interview or self-report etc.) 4. The type of cancer treatment is described 5. The results are compared between two groups or more (e.g., healthy population, groups with different cancer treatment or age, comparison with time at diagnosis etc.) . Mean or median and range or standard deviation of time since diagnosis or treatment is given 7. Participation and response rates for patient groups have to be described and have to be more than 75% . Information is presented about patient/disease characteristics of responders and non-responders or if there is no selective response 9. A standardized or valid quality of life questionnaire is used 10. Results are not only described for quality of life but also for the physical, psychological and social domain 11. Mean, median, standard deviations or percentages are reported for the most important outcome measures 12. An attempt is made to nd a set of determinants with the highest prognostic value 13. Patient signed an informed consent form before study participation 14. The degree of selection of the patient sample is described Outcome measurements QOL is the primary outcome of this review and meta-analysis. Mean is the summary measure.

Data analysis
The results of included studies which were not suitable for meta-analysis were synthesized narratively. The pooled estimate GQOL and QOL scales were analyzed by open Meta analyst software and presented with forest plot. Dersimonian and laird's random effect model was used [15].
The heterogeneity of included studies was evaluated with I 2 statistics. Based on I 2 statistics, a value less than 25% were considered low heterogeneity, between 50 and 75% medium heterogeneity and greater than 75% were considered as high heterogeneity [16]. Subgroup analysis was performed based on country to determine heterogeneity.

Study selection
Based on the search strategy in the data bases, a total of 1081 studies were retrieved initially. About 754 studies were remained after 327 duplicate studies were removed. 737 studies were excluded due to unrelated title and abstracts. Then, the remaining 17 studies were evaluated based on the eligibility criteria for inclusion and exclusion. Finally, ten studies met the eligibility criteria and included for nal review and meta-analysis (Fig 1).

Study characteristics
A total of 10 articles which reported QOL of 2,190 breast cancer patients were analyzed. The mean age of patients was 46.97 years. From ten included articles, eight articles were cross-sectional studies and the remaining two articles were prospective observational studies. All included studies were done in Hospitals.
The studies were conducted on ve countries (four from Ethiopia; two from Egypt; two from Nigeria; one from Kenya and one from Morocco). The included studies collected the data through interview. To measure health related quality of life of breast cancer patients, studies were used EORTC QLQ-C30, EORTC QLQ-BR23, FACT B, FACT G and WHOQOL-BR23 questionnaires (Table 1).

Methodological quality
Based on the 14 quality assessment criteria, eight studies had high quality (attained a score of 10 or higher) whereas the remaining two studies had moderate quality (attained a score of 9). Most of (about 90%) the studies didn't compared two groups. All of the included studies lacked information on the characteristics of non-respondents ( Table 2).

Quality of life based on EORTC QLQ-C30 standard tool
Seven studies reported QOL in breast cancer patients based on EORTC QLQ-C30 questionnaires. Based on this standard tool, the pooled estimate of the mean score of GQOL was 52.77(95% CI: 42.199 to 63.345; I 2 =99.21%, P < 0.001) (Fig 2).
A leave-one-out meta-analysis was performed to show how each individual study affects the pooled estimate of the remaining studies. From the analysis, there was no change in the pooled estimate mean GQOL of breast cancer patients. There was no single study that substantially in uenced the overall effect size. The pooled estimates mean GQOL was between the con dence interval of the pooled mean GQOL when one study was leaving out (Fig 4).  (Table 4).
In one study, QOL was dichotomized into poor and good QOL. This study showed that about 79.2% of breast cancer patients faced nancial di culties whereas nausea/vomiting was least affected symptom scales of 26.6 (66%).

Breast cancer-speci c functional and symptom scores based on EORTC QLQ-BR23
Five studies reported cancer speci c functional and symptoms based on EORTC QLQ-BR23. The pooled estimate score of body image and future perspective were highest whereas breast and arm symptoms were lowest (

Determinants of HRQOL of breast cancer patients in Africa
There are different factors that are associated with HRQOL in breast cancer patients (Table 6)

Discussion
The basic goal of healthcare is to improve the patients' quality of life and it is speci cally important in the case of breast cancer because the disease is more likely diagnosed at late stage than other forms [23]. The aim of this systematic review and meta-analysis is to investigate health related quality of life of breast cancer patients in the African countries. This review and meta-analysis analyzed 10 studies that reported QOL in breast cancer patients in Africa. About 2,190 breast cancer patients were involved for the analysis of the pooled estimation of QOL.
Standard tools introduced to quantify health related QOL in breast cancer patients had frequently developed for the last decade. All studies were used standardized data collection instrument and the most frequently used standard tool to measure the QOL in this review were EORTC QLQ-C30 and EORTC QLQ-BR23 which is similar to a review done in Spain [24]. But this nding was slightly different from the review of the review conducted [25] which stated that frequently used speci c QOL instrument in breast cancer patients were FACT-B and EORTC QOL-BR23.
In systematic review and meta-analysis, data collection instrument showed that the scores of QOL of African breast cancer patients differ from country to country. Therefore, the mean score of QOL varies from 28.38 to 65.48. Good scores were recorded at Kenya, Ethiopia and Nigeria and the lowest mean score was recorded among Egyptian breast cancer patients (28.38). Differences observed in HRQOL scores among these African breast cancer patients may be related differences in the time since diagnosis, disease stage, treatments they received and variation in socio-demographic characteristics of participants.
Based on EORTC QLQ-C30, the pooled mean score of GQOL in this review was 52.772 (95% CI: 42.199 to 63.345). This nding was lower than a systematic review and meta-analysis done in Eastern Mediterranean region in which the mean overall QOL was 60. 5 [26]. This difference may be due to the total sample size included in the study which was 6,034, better quality of care provided during the course of the disease and differences in socio-demographic characteristics between African and Eastern Mediterranean region participants. . These may be because of majority of African populations including Ethiopia belonged to lower-middle class families which may pose additional nancial burden for the cost of the disease management. Diarrhea (mean 14.75 with 95% CI: 6.55, 22.96) and nausea/vomiting (mean 18.81with 95% CI: (9.83, 27.79) were the least breast cancer speci c symptoms consistent with nding of meta-analysis conducted in Eastern Mediterranean region in which diarrhea was the least frequent cancer speci c symptom (mean score of 16.7). The reason for nausea/ vomiting and diarrhea were the least breast cancer speci c symptom may be the symptoms are experienced within a week period of taking chemotherapy.
According to the study conducted by han et al., 2010 stated that a woman who has good body image and better conceptualization of it can cope up the cancer better [28]. But woman with poorer body image in the breast cancer disease had greater psychological distress and greatly associated with depression and poorer QOL [29]. In our systematic review and meta-analysis, based on EORTC QLQ-BR23 questioner, body image (mean = 62.47; 95% CI = 46.33, 78.62) and future perspective (mean = 53.12; 95% CI = 31.75, 74.49) were the highest functional and breast cancer symptom whereas, breast symptom and arm symptom were the lowest with pooled mean estimate of 26.56; CI = 16.07, 37.05 and 26.7; CI = 20.41, 32.98 respectively. This was consistent with individual study conducted by sun et al., 2014 that body image was the highest breast cancer symptoms with mean score of 80.6 and breast symptom and arm symptom were the lowest with mean score of 9.4 and 14.9 respectively. This study (sun et al., 2014) states that superiority body image is the greatest strength of breast cancer patients [30]. In addition to this, body image is closely linked to identify, self-esteem, attractiveness, sexual functioning and social relationships [31].
In this systematic review and meta-analysis, one study which was done by koboto et al., 2020 was used WHOQOL-BR23 data collection tool to assess health related QOL of breast cancer patients. In this study, environmental domain (mean ± SD = 93.31 ± 19.76) followed by physical health domain (mean ± SD = 88.26 ± 21.61) were the highest mean score [11]. Whereas, psychological and social domain (mean ± SD = 68.20 ± 19.07 and 36.69 ± 7.62) respectively were the lowest [32]. These differences may due to cancer stigma and cultural view of the community. Findings of this study is slightly different from the study done in Srilankan in which environmental and social domains were higher in their mean score than physical and psychological domains [33]. This may be one indication of differences in culture, religion and social value among different countries.
Based on the ndings of this systematic review and meta-analysis, several factors were associated with HRQOL in breast cancer patients. Regarding to sociodemographic characteristics, there was no consistencies between studies in case of age. Three studies [17, 19, and 21] suggest that, cancer patients' HRQOL was negatively affected in older patients; this may be due to inability to tolerate adverse effects of chemotherapy and inability to perform their daily activity.
While two other studies [8,22] found that younger patients' QOL were more affected than older patients and the reason behind to this suggest that inability to ful ll themselves as wives and they need better physical appearance than older patients. The nding of these studies was supported by other systematic review conducted in Middle East of breast cancer patients which states that there was inconsistency between studies regarding the effect of age on HRQOL [34]. The review also identi ed that patients with higher level of education has better QOL than illiterate; this is due to those who are educated may have better access to salaried and employment and get better economic resource that brings good sense of control. This result is also supported by the study conducted in Korean breast cancer patients [35] and other studies done about education and QOL [36]. Studies conducted in Kenya stated that married participants had better QOL than unmarried and divorced participants. The possible explanation for this was, married ones may get nancial support from families and married by itself is a form of social support that led to positive in uence on QOL.
This review tried to address all relevant information regarding HRQOL among breast cancer patients in African countries. However, it has limitations. Considerable heterogeneity was existed in the included studies. The observed heterogeneity can be described by differences in quality of the study, the study design used and sensitivity. Subgroup analysis was performed based on country only because of variation of variables from study to study. This was also another limitation of our study. analyzed the data. EE solved the disagreement between the authors and gives general advice and corrections to this review and meta-analysis.   Pooled estimate of the mean score of GQOL based on EORTC QLQ-C30 standard tool Subgroup analysis based on country in breast cancer patients in Africa A leave-one-out meta-analysis of breast cancer patients in Africa

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. PRISMAchecklist.docx