Factors Affecting Delayed Presentation and Diagnosis of Breast Cancer in Asian Developing Countries Women: A Systematic Review

Background: Advance in screening strategies and management had steadily decreased the mortality rates of breast cancer. In developing countries, conducting screening and early diagnosis of breast cancers may face several problems. This systematic review aims to determine factors affecting the delayed diagnosis of breast cancer in developing countries in Asia. Methods: Literature research was conducted through Pubmed, ScienceDirect, Scopus, EbscoHost, Cochrane Library, and Google Scholar. The main keywords were “breast cancer”, “delayed diagnosis” and “developing countries”. Both quantitative and qualitative studies were included. Results: A total of 26 studies were included. The definition of delayed presentation or diagnosis varied from 1 month to 6 months. Among all the factors from patients and providers, breast symptoms and examinations consistently showed a significant contribution in reducing delayed diagnosis. Strengthened by qualitative studies, patients’ knowledge and perception also had a major role in delayed diagnosis. Conclusion: Among Asian developing countries, breast symptoms and examination, as well as individual knowledge and perception, are the main factors related to delayed diagnosis of breast cancer.


Introduction
Breast cancer is the second most common type of cancer globally, after lung cancer (Bray et al., 2018). From nearly 10 million deaths due to cancer worldwide, 6.6% are attributed to breast cancer. Significantly higher mortality-incidence ratios and lower survival rates of breast cancer in low-and middle-income countries (LMIC) or developing countries have been highlighted (Panieri, 2012;da Costa Vieira et al., 2017;Rivera-Franco and Leon-Rodriguez, 2018;Francies et al., 2020).
In the last few decades, advances in screening strategies and management have steadily decreased the mortality rates of breast cancer (Kohler et al., 2015). Screening and early diagnosis hold vital roles as the best survival rates are seen in early-stage breast cancers (Richards et al., 1999). Late diagnosis leads to a longer time interval to treatment (surgery or adjuvant therapies), ultimately worsening the overall survival and recurrence rates (Bleicher et al., 2016;Eaglehouse et al., 2019).
In developing countries, conducting screening and REVIEW

Factors Affecting Delayed Presentation and Diagnosis of Breast Cancer in Asian Developing Countries Women: A Systematic Review
early diagnosis may face hindrance from two main groups of factors. Patient factor groups such as socioeconomic background, family support, social stigma, and culture may influence the patients' decision to seek professional health care Iskandarsyah et al.;Poum et al.;Huo et al.;Roy et al., 2015;Kumar et al., 2019;Zhang et al.). On the other hand, affordability of service, quality of workforces and resources, compliance to available clinical guidelines in practice, and referral system are health care factors that contribute equally to patient factors (Sacerdote et al., 2013;Kumar et al., 2019;Zhang et al., 2019;Ho et al., 2020;Songiso et al., 2020). Even though multiple factors affecting delayed diagnosis in Asian developing countries have been studied, the results are heterogeneous and indecisive. Therefore, we conducted a systematic review to determine factors affecting breast cancer delayed diagnosis in Asian developing countries. We specifically chose Asia due to a common shared historical and cultural root in the region, hence avoiding overly diverse results.

Quality assessment and data extraction
Two independent reviewers (FBS and HG) assessed the risk of bias using the Newcastle-Ottawa Scale (NOS) (Wells et al., 2014). Seven items were assessed for risk of bias in each included study. Every item gave a certain quantity of stars. Studies with a total of ≥7 stars were considered of having a low risk of bias and vice versa. In this systematic review, stars in the instrument were represented by numbers.
For data extraction, three independent reviewers (FBS, RCRAP, and HG) were involved. Disagreements between reviewers were reconciled by discussion. Another independent reviewer was sought if an agreement was not achieved. Confirmation to the corresponding author was done if there were incomplete or ambiguous data. The extracted data included age at the time of diagnosis, education, economic status, residential status, marital status, age at first birth, parity, post-menopause, breast symptoms, the regularity of self and clinical-breast examination, self and family history of breast disease, knowledge and perception of breast cancer, type of payment, first consultation provider, chances of consultation, false-negative diagnostic test, and alternative therapy or traditional medicine. Other relevant factors were also extracted. Endpoint statistical measures included were odds ratio (OR), relative risk (RR), mortality rate, survival rate, proportion, and p-value for quantitative studies and description of causal factors in qualitative studies. If OR was not stated in the studies, manual calculations were done. If the risk comparison between numerator and denominator was different, the OR was recalculated for adjustment.

Data synthesis
If there was adequate data after compilation and analysis, all the data was synthesized into a meta-analysis.
A fixed-effect model was used for relatively homogeneous data, while a random-effect model was used for heterogeneous data. Test of heterogeneity is using I 2 statistics and Cochran's Q test. A funnel plot analysis is used to investigate the possibility of any publication bias in meta-analysis. No predefined subgroup analysis were planned.

Results
A total of 4,322 studies were identified at the beginning and after a thorough process (shown in Figure 1), 26 studies were included (shown in Table 1). None of the included studies had incomplete information so further confirmation to the author was not required. Further details on the steps of the study can be reviewed in Supplementary Table 1. Details of excluded studies were in Supplementary Table 2. Most of the studies were conducted in Malaysia and India. The definitions of delayed diagnosis varied from 1 month to 6 months (shown in Table 2). From 14 quantitative studies, 12 were cross-sectional studies and the other two were observational ones. Out of 12 cross-sectional studies, 5 of them had a high risk of bias, while the rest had a low risk of bias (shown in Table 3). Factors related to delayed diagnosis were summarized into socioeconomic status, maternity status, breast symptoms and examinations, healthcare-related, patients' knowledge and perceptions, and other factors as seen in Table 4. None of those factors were found to be strongly related to delayed diagnosis of breast cancer, but breast symptoms and examinations, and patient knowledge and perception, were major contributors. Results from qualitative studies, as shown in Table 5, validate the information on how different those factors were between the various study settings.
Meta-analysis was not possible due to the varying definitions of the delayed diagnosis (some studies used three months as a cutoff, while others used six months or more), uneven distribution of the location of the studies (mainly in Southeast Asia and South Asia), and high risk of bias in the majority of the studies.    On the other hand, the weakness of this study was the unfeasibility of conducting a meta-analysis due to the aforementioned reasons. The risk of publication bias was relatively low because some factors in the included studies were not statistically significant, yet the studies were still published. However, most of the included studies had a high risk of bias, especially in the aspect of patient selection, and thus might have led to selection bias. We hypothesized that the high risk of selection bias was probably caused by the difficulty in collecting accurate pre-diagnosis information from patients. Moreover, the information was collected retrospectively, causing a risk of recall bias. Some studies

Discussion
This systematic review was the first one to summarize factors associated with delayed diagnosis of breast cancers in Asian developing countries. Clinicians and stakeholders can use the information gained from this study, to design preemptive models against elements that cause the delay.
The inclusion of qualitative research strengthened this study by enhancing information obtained through quantitative research. It gave a more unquantified perspective on how patients' perception, fear of cancer diagnosis, knowledge, and healthcare accessibility played an important role in preventing delayed diagnosis of breast  Although strength of evidence for each factor related to breast cancer delayed diagnosis are various, we had seen some tendency as will be discussed further in this section. Three months as the cut-off to define delayed diagnosis was mostly used in included studies. The previous metaanalysis clarified that globally, a delay (defined as the delay from the onset of symptoms to the start of treatment) of more than three months was strongly related to a poor survival rate (Richards et al., 1999).
Principally, reasons for a delay in diagnosis were grouped into patients' and providers' delays. However, we found some factors fell into a "grey area". For example, a patient's visit to a non-physician at the time of the onset of symptoms can be influenced by the patient's perception. However, it also can be caused by the low quality of available health providers or the long distance to the nearest provider. Thus, both factors should be carefully investigated when optimizing the screening and early diagnosis of breast cancers.
A past systematic review reported that single marital status and advanced age were the only sociodemographic factors that seemed to be strongly associated with patient delay (Richards et al., 1999). However, further studies have been shown to obtain contradictory results (Rivera-Franco and Leon-Rodriguez, 2018). Meanwhile, higher formal education and socioeconomic status could reduce delayed diagnosis, even though several studies, contrastingly, did not report similar results (Bodapati and Babu, 2013;Lim et al., 2015;Roy et al., 2015;Dyanti and Suariyani, 2016;Solikhah et al., 2020). Residential status and distances to healthcare facilities were also among factors related to delayed diagnosis in breast cancer (Bodapati and Babu, 2013;Thakur et al., 2015). Among maternity status, age at first birth significantly reduced delayed diagnosis, but its categorization (>20 years old and 21-25 years old) was not well related to breast cancer's biomolecular background (Poum et al., 2014;Kumar et al., 2019). Menopausal status that strongly affected delayed diagnosis of breast cancer might be explained by less breast mass density in postmenopausal women which allows for early detection during breast self-examination (BSE) or clinical breast examination (CBE).
In our opinion, the delay from the provider is rooted in two stages. The first stage is primary health care. Ideally, every country should have a breast cancer screening program that is bound by the law to reinforce the screening uptake. Since approximately one-third of Asian countries have been categorized as low-and lower-middle-income countries, screening modalities should be tailored according to the economic capability of the country (World Bank, 2020). One excellent example is the recommendation from World Health Organization (WHO) that such countries may attempt screening by CBE for women aged 50-69 years old if mammography is too expensive and not feasible (World Health Organization, 2014). Moreover, the primary healthcare providers' CBE skills should be assessed and evaluated periodically. To ensure non-substandard CBE skills, the associated educational institution can aid in the form of hands-on training or other kinds of knowledge-sharing sessions. A clear referral system with geographically accessible secondary or tertiary healthcare facilities is also needed.
The second stage of provider delay is the referral stage. The problem at this stage stems from the inability of secondary healthcare to provide an accurate and quick diagnosis of breast cancer due to a lack of skills or facilities/resources. Therefore, both institutional and healthcare centers must be improved to rectify this problem. An initiative from the government for equal distribution of secondary and tertiary healthcare providers (e.g., general surgeons or surgical oncologists) and healthcare facilities are also required. Overall, while funding systems may initially become a problem in developing countries, a well-executed established system can serve as a driving force for physicians to act according to the clinical guidelines. Eventually, proper and on-time   The Government's commitment holds incredible importance in the timely management of cancer in general. As an example, the law in Brazil states that the time from diagnosis to treatment of breast cancer should be less than 60 days (Ferreira et al., 2020). In the United Kingdom, suspected cases of colorectal cancer have to be referred within two weeks (Thomas and Burnet, 2001). While they have flaws, those programs showed the willingness of regulators to be involved in the management of cancer. On the other hand, a system that is too strict and non-adaptive may cause further delays. An example of this is the referral system in Indonesia, which is based on the authors' experience. The waiting time for diagnosis of individuals suspected to have breast cancer is unnecessarily prolonged due to the inability of patients to be referred directly to a cancer center or similar health facility with definitive diagnostic capabilities, because of administrative reasons from National Health Insurance. This also hampers the physician's ability to quickly perform the required diagnostic procedures. For instance, a biopsy cannot be performed in the same visit as initial laboratory and radiologic workups due to a "cost package limit".
An important finding in this systematic review was that the presence of breast symptoms (any symptoms including breast mass, ulcer, or pain and nipple discharge) was consistently related to less delayed diagnosis, both in quantitative and qualitative studies (Norsa'adah et al., 2011;Poum et al., 2014;Huo et al., 2015;Lim et al., 2015). However, because symptomatic breast cancer may be indicative of a more advanced stage of cancer, other factors that can encourage individuals to get diagnosed early should be explored.
Breast cancer detection through BSE, although not routinely done, was among the most effective methods for early detection of breast cancer (Ghazali et al., 2013;Dyanti and Suariyani, 2016). CBE is also important for the early detection of breast cancers as it significantly reduced delay in diagnosis and presentation (Norsa'adah et al., 2011;Huo et al., 2015). Both of them should be the main strategies in breast cancer detection in developing   Patients' knowledge and perception, or a good literacy on breast cancer, especially about its symptoms, are related to less delayed diagnosis. From qualitative evidence, those factors were known to influence the act of seeking a professional health provider and avoiding alternative therapy, in addition to a less negative attitude toward medical services. Psychosocial factors, such as the perception of fear, denial, embarrassment, ignorance, and emotional burdens were reinforcing factors that need support and motivation from close relatives and the environment (Norsa'adah et al., 2012;Bodapati and Babu, 2013;Iskandarsyah et al., 2014;Huo et al., 2015). In most Asian countries, family bonds along with husbands' or fathers' decisions have a strong impact on women's actions (including health issues); therefore, education on the importance of breast cancer screening and early diagnosis should reach all family members and community members.
Health service quality is another key factor in the prevention of delayed diagnosis. Still, the interpretation of quality was varied among studies, such as competency of health workers, time and distance to reach healthcare services, and quality of the diagnostic test. In addition, there was a tendency in Asian cultures to visit alternative or traditional medicine (Norsa'adah et al., 2011;Ghazali et al., 2013;Thakur et al., 2015). The intention to visit alternative therapy may have arisen from either personal preferences or due to unsatisfied services received from a previous healthcare provider (Norsa'adah et al., 2012;Thakur et al., 2015;Akhtar et al., 2018).
Further meta-analyses should be conducted whenever more data are available. On the other hand, high-quality observational studies may be undertaken to obtain specific information on how to optimize breast cancer detection. It ought to have an adequate sample size and consider both patient and provider factors so that the data is more comprehensive. More evidence is needed because different approaches to prevent the delayed diagnosis of breast cancers should be considered in developing countries as compared to high-income countries. Furthermore, to improve patients' knowledge and perception, every country should conduct individualized anthropological research to discover an appropriate communication, information, and education model which takes into account the local culture.
In conclusion, among Asian developing countries, breast cancer symptoms and examinations, as well as individual knowledge and perception, are the main factors related to delayed diagnosis of breast cancer. Socioeconomic, maternity, and health service quality have various impacts on the delayed diagnosis of breast cancers, and thus need specific approaches, tailored to the local public health situation.

Author Contribution Statement
FBS, AB, SSP, DA, HG, PWY, AAN, RCRAP, and ADW contributed equally to the process of drafting the protocol. Both RCRAP and ADW contributed equally for developing the search strategy and running the search. FBS, RCRAP, and HG selected studies for inclusion and extracting data. The risk of bias was assessed by FBS and HG. All authors carried out and interpreted the analysis. FBS, PWY, and HG wrote the manuscript. All authors have read and approved the manuscript.