A systematic review of barriers and enablers to South Asian women’s attendance for asymptomatic screening of breast and cervical cancers in emigrant countries

Objectives The aim of this review was to identify the cultural, social, structural and behavioural factors that influence asymptomatic breast and cervical cancer screening attendance in South Asian populations, in order to improve uptake and propose priorities for further research. Design A systematic review of the literature for inductive, comparative, prospective and intervention studies. We searched the following databases: MEDLINE/In-Process, Web of Science, EMBASE, SCOPUS, CENTRAL, CDSR, CINAHL, PsycINFO and PsycARTICLES from database inception to 23 January 2018. The review included studies on the cultural, social, structural and behavioural factors that influence asymptomatic breast and cervical cancer screening attendance and cervical smear testing (Papanicolaou test) in South Asian populations and those published in the English language. The framework analysis method was used and themes were drawn out following the thematic analysis method. Settings Asymptomatic breast or cervical screening. Participants South Asian women, including Bangladeshi, Indian, Pakistani, Sri Lankan, Bhutanese, Maldivian and Nepali populations. Results 51 included studies were published between 1991 and 2018. Sample sizes ranged from 25 to 38 733 and participants had a mean age of 18 to 83 years. Our review showed that South Asian women generally had lower screening rates than host country women. South Asian women had poorer knowledge of cancer and cancer prevention and experienced more barriers to screening. Cultural practices and assumptions influenced understandings of cancer and prevention, emphasising the importance of host country cultures and healthcare systems. Conclusions High-quality research on screening attendance is required using prospective designs, where objectively validated attendance is predicted from cultural understandings, beliefs, norms and practices, thus informing policy on targeting relevant public health messages to the South Asian communities about screening for cancer. PROSPERO registration number CSD 42015025284.


Strengths and limitations of this study
• Separate outcomes were compared of integrative reviews of inductive, predictive, comparative and intervention studies to assess consistencies between methods.
• Deductive studies were either atheoretical or used generic health psychology theories, such as the health belief model, that were validated on western samples and not adapted for South Asian populations.
• Studies were primarily conducted in Canada and the USA where a payment-oriented healthcare is available which may not be comparable to other health services.
• Due to the small number of published studies, it is difficult to identify factors unique to groups of South Asian women based on nationality, geographical region or religion.
• High quality research on screening attendance is required using prospective designs. South Asian women in the UK have higher breast and cervical cancer mortality and are more likely to present with advanced disease 5 . This is partly attributable to their lower likelihood of attending routine mammographic and Papanicolaou (Pap) screening. Screening is widely available in most high income countries [6][7][8] and probably reduces mortality and morbidity through early detection and treatment 9 . South Asian women in England show lower uptake of breast screening services than the host population [10][11][12][13] , particularly those from lower socioeconomic groups 10 14 15 and a higher proportion than the host population have never received cervical screening 16 .
It is unclear why screening rates in South Asian women are lower than host population rates. Possible explanations have included poorer individual knowledge and awareness of breast and cervical cancer [17][18][19] , lower community awareness, poor communication between health professionals and patients and less access to appropriate cancer health services 20 21 . Some South Asian women cannot speak or read in the host language 22 23 . Another body of research focuses on South Asian women's attitudes, beliefs and behaviours relating to cancer screening 24 25 . Crawford et al. 25   behaviour in immigrant groups are unique to them or are shared with host or other immigrant groups. Shared factors include relative economic deprivation 22 28 or social and cultural adjustment challenges 29 . It is also important to review reports of intervention studies to examine how successful previous interventions (or their individual components) have been in improving screening rates in South Asian populations.

Aims of the Review
We examined cultural, social, structural and behavioural factors that influence asymptomatic breast and cervical cancer screening in South Asian populations, to explain why attendance rates are lower than host country women. We performed separate integrative reviews of inductive, predictive, comparative and intervention studies, and compared outcomes of these reviews to assess consistencies between methods. Our aim was to identify the cultural, social, structural and behavioural factors that influence asymptomatic breast and cervical cancer screening rates in South Asian populations to improve screening rates and to propose priorities for further research. Our objectives were to: • critically review and integrate findings of inductive, predictive, comparative and intervention studies on asymptomatic screening; • document consistent and inconsistent findings across methods; • make theoretical and methodological recommendations for the conduct of future research. We conducted literature searches using multiple databases to overcome problems associated with inadequate indexing 27 30 and to ensure a more exhaustive scope 27 31

Selection criteria
The review included studies of attendance at asymptomatic screening of female breast and/or cervical cancers by routine mammography and cervical smear testing (Papanicolaou test). It was confined to host countries where mass screening programmes are available to the general public. The populations of interest were Bangladeshi, Indian, Pakistani, Sri Lankan, The review did not include breast self-examination, diagnostic screening or visual or tactile examinations by healthcare professionals. We excluded studies of women in known high-risk groups who were engaged in monitoring programmes for genetic risk factors, hereditary breast and ovarian cancer syndrome, premenopausal or familial breast cancer. We excluded homogenous samples restricted to particular demographic groups (e.g. a study of dental students).

Screening
Team members screened titles and abstracts to identify potentially eligible studies and two reviewers independently considered the eligibility of each of the titles and abstracts. Outputs were compared to detect discrepancies and the agreement rate was 90%. Disagreements over selection of abstracts were resolved by consensus between the team. Calibration of the selection criteria was performed after the first 50 and 100 papers and taking a small sample (15%) of reports from grey and unpublished literature. Two reviewers independently assessed the full text of relevant studies using a standardized, pilot-tested screening form agreed with the steering group. Disagreements were resolved by consensus or by referral to a third-party arbiter. EndNote (X5) reference manager was used to manage citations and view abstracts and full-text articles.

Quality evaluation
Each study was evaluated for quality specific to the method used, with validated checklists developed from the Critical Appraisal Skills Programme 33 . Inductive studies were generally found to be good. Predictive, descriptive and intervention studies had theoretical, sampling, design and measurement limitations. We did not exclude studies that used poor methodologies, but extensively describe these problems and consequent interpretive limitations in the results.

Data extraction, synthesis and analysis
All studies included in the review are included in summary tables (Tables 1 to 4). Four reviewers completed data extraction for each study type and reviewed the variable headings upon completion 34 . Subsequently, tables were adapted and the following variables were recorded for all studies: region, study design (sample size and sampling); demographic and clinical characteristics of women selected; setting; data collection instruments; analytic method; nature of asymptomatic screening (mammogram or Pap smear test); definition of timely screening attendance; theoretical focus; key findings; study limitations and quality rating. For predictive studies we recorded outcome variables, rate of screening attendance and all predictors for and against screening. Intervention studies included a description of the intervention concerned.
Syntheses were made using thematic analysis within each methodology type 27 35 .
Syntheses were initially structured from the summary tables, beginning with a period of data familiarisation, during which researchers listed ideas about emerging themes which formed the basis of a thematic framework. At this point, the analysis returned to the full papers, where the developing thematic framework was tested and refined against the initial data.
Themes were developed, reviewed and refined by analysing the data synthesised within each code and testing for 'internal homogeneity' and 'external heterogeneity' 36 . The research group met continuously to check and discuss the meaning and interpretation of the data. Participants were recruited from community and healthcare settings and had a mean age of 18 to 83 years. Eight were inductive, 21 predictive (containing analysis of predictors of and risk factors for attendance), 9 comparative, and 6 intervention studies. No further studies were found from the grey literature search.

Overview
Inductive studies provided rich insights into cultural practices and assumptions, and the problems of adjusting to a new social and healthcare system that might inhibit screening in South Asian women. Largely, though, deductive studies failed to exploit these insights in hypothesis testing. Deductive studies were either atheoretical or used generic health psychology theories, such as the health belief model, that were validated on western samples and not adapted for South Asian populations.
Nonetheless, common findings emerged across methodologies. The extent to which women understood the causes of cancer and the benefits of screening were important.
Inductive studies revealed cultural constraints on understanding, whilst comparative studies showed South Asian women faring worse on measures of knowledge than host country women. Predictive studies showed that those with more complete understandings of cancer and screening were more likely to attend screening. Similarly, both inductive and deductive studies showed that perceived barriers inhibited screening, and that South Asian women typically perceived more and different barriers to host country women. Inductive studies showed the cultural origins of barriers, describing how traditional beliefs about risk, illness, female roles and family structures mitigated screening interest and attendance. Predictive studies showed that the number of perceived barriers inhibited screening, and that South Asian women who were more acculturated to western host countries, operationalized as time spent in those countries, were more likely to attend screening.

Inductive studies
The seven inductive studies (  Knowledge, attitudes, understanding of cancer and cancer prevention: Neither cancer nor intimate body parts are commonly discussed in some South Asian cultures 38 42 . All studies showed that women lacked basic understandings of cancer, cancer prevention or early detection. Breast cancer was viewed by some women as a 'white woman's disease' 37 , that did not occur in their community 41 . Others considered cancer to be incurable and early detection and intervention futile 41 . Cervical cancer was often not known or understood. For example, some Bhutanese refugees in the USA had not heard of cervical cancer 41 . Indian Sikhs in Canada, living in a culture where sexual and reproductive health is rarely discussed, referred to the cervix as an 'unknown' and unspoken part of their body 42 . Those aware of cervical cancer perceived the principal risk factors to be inseparable from those for general health, rarely mentioning the discrete risk factors of having multiple sexual partners, not using barrier contraception or screening 41 .
Studies attributed a lack of understanding of cancer to two main factors. First, religious fatalism meant that cancer was seen as predestined, as divine retribution for sins, or as a dearth of moral character 38 . Second, all studies pointed to the curative focus of healthcare in countries of origin as a reason for some women's failure to understand the concept of prevention 38 44 and consequent belief that healthcare seeking is unnecessary in the absence of symptoms 37 40 42 .
Culture: Family responsibilities were salient to women. This had three implications, one positive and two negative. First, women felt strong responsibilities to remain in good health and to protect family members from cancer 40  cervical cancer were seen as stigmatising 42 and to some women this extended to screening 38 41 . Indeed, some Sri Lankan Tamils worried that attending a mammogram would lead people to think they already had breast cancer 38 . Third, women's behaviour was often subject to influence from male members of the family. Women frequently followed family advice for healthcare provided by males and elders, generally against screening, and felt the need to avoid conflict within the family associated with assertions of independence 37 42 .
The process of Cultural Adaptation: South Asian nations have largely curative health systems, in which health costs are required to be paid by patients and there is no free access to healthcare. This contrasts with preventive healthcare models in host countries, and as with other health issues South Asian women showed little understanding or orientation toward cancer prevention 42 , although this evolved with time as awareness of the culture of the host country increased 39 43 . Women appreciated healthcare professionals who understood and respected values of personal modesty/shyness 41 . South Asian women in Canada emphasised the value of being chaperoned to screening appointments that may have been located away from their local community, for assistance with language barriers, to alleviate feelings of personal vulnerability and to avoid being alone with doctors 37 .

Deductive Studies
Study Quality: Predictive and comparative studies contained similar limitations to quality.
The first limitation was the poverty of theory. With the exception of Pons-Vigues and colleagues, whose deductive study 45 46 , with limited applicability to South Asian populations. Similarly, the concept of acculturation was invoked in predictive studies, but was operationalised in a limited way, focussing on time spent in the emigrant country and language preferences. Other deductive work was not theoreticallybased.
Studies were also affected by methodological limitations. A database linkage study 47 and two cluster sampling studies 48 49 provided samples with a potentially high degree of population representativeness, with random digit dialling techniques providing some confidence that samples may be representative 50 51 . Other studies used poor sampling techniques, including selection of South Asian names from phone directories or sampling at cultural events or other locations with high proportions of South Asian women, providing less confidence. Definition of a South Asian population differed between studies, some examined women born in South Asia, others second generation immigrants, and some examined selfidentified ethnicity.
It is important that attendance is recorded objectively 52 . All studies but one 47 used non-verified self-reported attendance and one used a hypothetical scenario of an offer to attend screening 50 . These outcomes included timely screening attendance (e.g., previous screening was within a specified time period or reported regular timely testing) or whether women had ever been screened in the past.
Predictive Studies: It is strongly recommended that predictive studies be conducted prospectively to eliminate the problem of reverse causality 52 . All of the 21 predictive studies were cross-sectional and causal interpretation is difficult. Lower screening rates were noted among women with no health insurance, younger women and women with lower levels of education. Studies did not provide consistent evidence that low knowledge predicted reduced likelihood of attendance. Lower knowledge was associated with a reduced likelihood of mammography screening in two studies 53 54 , but did not predict the likelihood of hypothetical acceptance of a cervical screen 50 . Lower attendance was associated with a greater number of self-reported barriers to screening [53][54][55][56][57] .
However, the instruments used to assess barriers were largely based upon existing instruments developed among western samples that do not reflect South Asian concerns such as adapting to a new culture, language or health system.
Two methodologically rigorous comparisons between South Asian and other minority groups 47 48 used population sampling and statistically adjusted for demographic differences between samples. Vahabi, et al 47 also used an objectively verified indicator of mammography attendance. Both showed South Asian women to have lower attendance rates than other immigrant women. In two studies, Indian women had lower knowledge of cancer and screening than Chinese or Malays 48 61 . Pons-Vigues et al. 45 and Teo et al 58 showed Indian and Pakistani women perceived fewer barriers arising from lack of knowledge about preventative screening than other immigrant groups, and highlighted that many of the women thought that routine blood tests and urine tests would detect broader health issues such as cancer 45 . In another study 59 , Indian women perceived themselves to be less vulnerable to getting breast cancer, did not view breast cancer as a serious illness and were more likely to claim that they did not know 'where to find a mammogram'.
Intervention Studies: Community educational programmes promoted breast and cervical cancer screening across the six intervention studies. Three of the studies were pre-and postcommunity-based interventions [62][63][64] , two were randomised control trials 44 65 ; and one a time series study 66 . Sampling was predominantly among South Asian women as a group, which eliminates comparisons between the different South Asian populations. Studies employed various methods of socioculturally-tailored, language-specific health education materials and participants were recruited from primary care or South Asian community venues and residences. Recruitment was opportunistic via local newspapers, surveys conducted in community settings, South Asian nurses and link health workers. No study examined age trends 64 ; and participants had met the researchers before which may constitute a bias 44 . Controlled studies were conducted in close-knit communities which may have led to intervention contamination into the control groups. Increased screening rates were reported for three studies but many were self-reported 62 63 ; rather than from objective indicators 44 .
No long-term change in screening uptake was reported for three studies [64][65][66] , but they showed an increase in knowledge of breast cancer among South Asian immigrant women and reduced the misperception of short survival after diagnosis.

Discussion
Prominent across study types were the findings that South Asian women had poorer understandings of cancer and cancer prevention and that they perceived greater cultural and structural barriers to screening than host country women.
Lack of understanding by South Asian women about the need for asymptomatic screening has important ramifications. Predictive studies showed greater knowledge to be associated with screening attendance. The inductive research yielded some plausible reasons for this. Many women held fatalistic views or beliefs that cancer is incurable, whilst others believed that cancers could be identified in routine health testing. Others were unaware of the existence of cervical cancer in particular, and did not perceive threat to themselves or their communities. Whilst there is a clear need to change such beliefs, the inductive studies showed this to be a challenging task for two reasons. First, understandings were embedded within religious and cultural traditions, and cannot be addressed in isolation to those traditions. Thus, a simple educational intervention is likely to have limited effect.
Accommodations will need to be reached with communities that allow a creative integration of cancer awareness within existing belief structures. Second, some women were largely unaware of the concept of disease prevention. Thus, the promotion of specific cancer awareness and understandings are unlikely to be helpful until a wider understanding of prevention is reached.
Predictive studies showed the importance of perceived barriers, but these barriers pertained only to generic barriers faced by all women, irrespective of culture. Acculturation, in terms of time spent in the host country and mastery of the language was associated with increased screening likelihood, but these issues are likely to exist for all immigrant women and fail to reveal specifically South Asian issues. Inductive studies provided more subtle and culture-specific indications of the barriers perceived by women. Many were cultural. In particular, women spoke of the importance of female modesty and stigma associated with cancer that also affected willingness to be screened. Whilst the importance of female testing staff from South Asian backgrounds and use of South Asian chaperones is emphasised, this cannot address the wider cultural issues of modesty and stigma. One finding that offers encouragement is that personal health is important to South Asian women because it helps them to care for their families.
Interventions will need to be conducted more widely than merely targeting women and their beliefs. Males occupy decision-making roles in some South Asian families and women may not wish to challenge this (see also 67 68 . Thus, addressing the views of male family members and other community opinion leaders is also important.

Limitations
The following limitations were identified within the review. First, studies were primarily conducted in Canada and the USA. North American conditions, such as payment-oriented healthcare and whether screening is promoted by a central agency or not, are not necessarily replicable in other health systems. Second, due to the small number of published studies, it is difficult to identify factors unique to groups of South Asian women based on nationality, geographical region or religion. By necessity, we discuss findings in terms of a generic 'South Asian' population, but are aware of variance between South Asian populations according to nationality, region, culture and religion.

Future Research
Stratifying the analysis by study methodology brings two benefits; greater confidence can be placed on findings that transcend methodologies than those that are contained within one method, and studies with similar methodologies can be critiqued in ways appropriate to those  Failure to incorporate inductive findings into the design of deductive studies means that many inductive findings are untested in a population context. Further, deductive studies themselves used flawed designs as they were generally atheoretical or based upon health behaviour models developed in western populations and thus potentially lacking insight into South Asian issues. Translation of inductive findings to a deductive context will require the development of valid and reliable instruments to assess cultural understandings, beliefs, norms and practices.
These studies will also need to use better empirical methods. Few studies used sampling techniques that can be confidently claimed to be population-representative. It is important to employ best practice in study design for screening attendance research; the use of prospective predictive studies and objectively verified reporting of attendance from clinical records 52 . Adequate sampling frames need to be established. Firstly, this involves a distinction between South Asian women as a minority group or as an immigrant group. The former can comprise women with high degrees of familiarity with the host country, but who nonetheless may be faced with cultural barriers deriving from their countries of origin. The latter group will reflect the problems of adjustment faced by recent immigrants. Studies will also need to use population-representative sampling techniques.

Recommendations for Practice
Findings from all study types demonstrate that interventions should be sensitive to cultural norms. In particular, studies emphasised the importance of language, female practitioners and the importance of community approval and involvement. Interventions at the community level will be necessary to surmount the cultural barriers identified in the inductive studies. It is worrying that the findings indicated that younger women and women with lower levels of education were less likely to attend for screening. Interventions need to be targeted at educating South Asian women who are younger, not married and less educated.
Encouraging female family members to become more involved as chaperones and translators could also be helpful, and may form a mechanism for educating young women simultaneously.
Information aimed at South Asian women who are invited for breast and cervical screening should highlight the presence of female practitioners and exclusively female environments at breast and cervical screening sites in the UK 69 . There is limited use of written communication in South Asian languages, although 70% of screening units across the UK want to provide information in patient's language 70 . This may help improve South Asian women's knowledge, make informed choice/ consent, have better patient experience and eventually help in improving their screening uptake rates.
Interventions to increase uptake rates need to be long-term, multifaceted and tailored to the specific needs of the local community by, for example, developing close links with the community through Health Education workers. South Asian community members, including males and opinion leaders, should be encouraged to be involved and co-produce engagement strategies within community settings. Reducing ethnic inequalities in uptake rates of breast cancer screening needs to remain a policy priority of the Government and the Breast Screening Programme.

Competing Interests
The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.

Conflicts of Interest
The Authors declare no conflict of interest.

Data sharing
No additional data available.      Attendance patterns were not maintained although women had positive experiences. Challenges for ongoing success: 1) maintaining the continued involvement of stakeholders in developing long-term strategies to enhance community awareness about cervical cancer; 2) creating mechanisms to strengthen support from physicians in the community; 3) meeting the needs of the underserved within a specialised health service for SA immigrant women.

Systematic Review Registration: PROSPERO CSD 42015025284
Strengths and limitations of this study • Separate outcomes were compared of integrative reviews of inductive, predictive, comparative and intervention studies to assess consistencies between methods.
• Inductive studies provided nuanced and detailed insights into cultural, social, structural and behavioural factors influencing screening attendance.
• Deductive studies did not use insights gained from inductive research, were either atheoretical or used generic health psychology theories that were validated on western samples and were generally poorly designed.
• Due to the small number of published studies, it is difficult to identify factors unique to groups of South Asian women based on nationality, geographical region or religion.
• We provide specific advice for high quality deductive research on screening attendance that will allow estimation of the prevalence of factors that facilitate or inhibit screening attendance and the magnitude of their influence on attendance.  In the UK, South Asian women have higher breast and cervical cancer mortality than the host population, worse cancer-related health outcomes, with the exception of some Indian groups, and are more likely to present with advanced disease 3 5 . Whilst South Asian and host populations may differ over a range of factors that influence mortality, such as tumour subtype and HPV status (Gomez 2010), one potential cause of greater mortality is that South Asian women show a lower likelihood of attending routine mammographic and Papanicolaou (Pap) screening. Screening is widely available in most high income countries [6][7][8] . Some research shows shows population mortality benefits of screening programmes 10 11 , although other studies find no effect 9 . Importantly, greater mortality benefits are found at the individual level, where studies confine analyses to women who accept screening rather than those who are merely invited (because some women decline screening) 12 . Compared with the host population, South Asian women in England show lower uptake of breast screening services [13][14][15][16] , particularly those from lower socioeconomic groups 13 17 18 and a higher proportion have never received cervical screening 19 . This is also the case in the USA 20 .  26 32 or social and cultural adjustment challenges 33 . It is also important to review reports of intervention studies to examine how successful previous interventions (or their individual components) have been in improving screening rates in South Asian populations.

Aims of the Review
We examined cultural, social, structural and behavioural factors that influence asymptomatic breast and cervical cancer screening attendance in South Asian populations, to explain why attendance rates are lower than host country women. We performed separate integrative reviews of inductive, predictive, comparative and intervention studies, and compared outcomes of these reviews to assess consistencies between methods. Our aim was to identify the cultural, social, structural and behavioural factors that influence asymptomatic breast and cervical cancer screening rates in South Asian populations to improve screening rates and to propose priorities for further research. Our objectives were to: • critically review and integrate findings of inductive, predictive, comparative and intervention studies on asymptomatic screening; • document consistent and inconsistent findings across methods; make theoretical and methodological recommendations for the conduct of future research.

Search strategy
We conducted literature searches using multiple databases to overcome problems associated with inadequate indexing 31 34 and to ensure a more exhaustive scope 31

Selection criteria
The review included studies on the cultural, social, structural and behavioural factors that influence asymptomatic breast and cervical cancer screening attendance and cervical smear testing (Papanicolaou test) in South Asian populations. It was confined to host countries where mass screening programmes are available to the general public, including South Asian To ensure that the studies pertained to screening attendance, we excluded those that did not specifically refer to screening. Thus, studies solely covering general attitudes to breast or cervical cancer were excluded. The review did not include breast self-examination, diagnostic screening or visual or tactile examinations by healthcare professionals. We excluded studies of women in known high-risk groups who were engaged in monitoring programmes for genetic risk factors, hereditary breast and ovarian cancer syndrome, premenopausal or familial breast cancer. We excluded homogenous samples restricted to particular demographic groups because these are not population representative (e.g. a study of dental students).

Screening
Team members screened titles and abstracts to identify potentially eligible studies and two reviewers independently considered the eligibility of each of the titles and abstracts. Outputs were compared to detect discrepancies and the agreement rate was 90%. Disagreements over selection of abstracts were resolved by consensus between the team. Calibration of the selection criteria was performed after the first 50 and 100 papers and taking a small sample (15%) of reports from grey and unpublished literature. Two reviewers independently assessed

Quality evaluation
Each study was evaluated for quality specific to the method used, with validated checklists developed from the Critical Appraisal Skills Programme 37 . Inductive studies were generally found to be good. Predictive, descriptive and intervention studies had theoretical, sampling, design and measurement limitations. We did not exclude studies that used poor methodologies, but extensively describe these problems and consequent interpretive limitations in the results.

Data extraction, synthesis and analysis
All studies included in the review are included in summary tables (Tables 1 to 4 Syntheses were made using thematic analysis within each methodology type 31 39 .
Syntheses were initially structured from the summary tables, beginning with a period of data familiarisation, during which researchers listed ideas about emerging themes which formed the basis of a thematic framework. At this point, the analysis returned to the full papers, where the developing thematic framework was tested and refined against the initial data.
Themes were developed, reviewed and refined by analysing the data synthesised within each code and testing for 'internal homogeneity' and 'external heterogeneity' 40 . The research group met continuously to check and discuss the meaning and interpretation of the data.

Patient and Public Involvement
The   Eight were inductive (see Table 1), 25 predictive (containing analysis of predictors of and risk factors for attendance) (see Table 2), 10 comparative (see Table 3), and 8 intervention studies (see Table 4). No further studies were found from the grey literature search.

Overview
Inductive studies provided rich insights into cultural practices and assumptions, and the problems of adjusting to a new social and healthcare system that might inhibit screening in Asian women who were more acculturated to western host countries, operationalized as time spent in those countries, were more likely to attend screening.

Inductive studies
The eight inductive studies (Table 1)  Knowledge, attitudes, understanding of cancer and cancer prevention: Neither cancer nor intimate body parts are commonly discussed in some South Asian cultures 42 46 . All studies showed that women lacked basic understandings of cancer, cancer prevention or early detection. Breast cancer was viewed by some women as a 'white woman's disease' 41 , that did not occur in their community 45 . Others considered cancer to be incurable and early detection and intervention futile 45 . Cervical cancer was often not known or understood. For example, some Bhutanese refugees in the USA had not heard of cervical cancer 45 . Indian Sikhs in Canada, living in a culture where sexual and reproductive health is rarely discussed, referred to the cervix as an 'unknown' and unspoken part of their body 46 . Those aware of cervical cancer perceived the principal risk factors to be inseparable from those for general health, rarely mentioning the discrete risk factors of having multiple sexual partners, not using barrier contraception or screening 45 .
Studies attributed a lack of understanding of cancer to two main factors. First, religious fatalism meant that cancer was seen as predestined, as divine retribution for sins, or as a dearth of moral character 42 . Second, all studies pointed to the curative focus of healthcare in countries of origin as a reason for some women's failure to understand the concept of prevention 42 48 and consequent belief that healthcare seeking is unnecessary in the absence of symptoms 41 44 46 49 .
Culture: Family responsibilities were salient to women. This had three implications, one positive and two negative. First, women felt strong responsibilities to remain in good health and to protect family members from cancer. 40 In some cases this facilitated screening attendance, however, some women found no time to attend screening due to family responsibilities. 44  women may be more likely to attend a cervical smear if the family doctor was female 49 . Both breast and cervical cancer were seen as stigmatising 46 and to some women this extended to screening 42 45 . Indeed, some Sri Lankan Tamils worried that attending a mammogram would lead people to think they already had breast cancer 42 . Third, women's behaviour was often subject to influence from male members of the family. Women frequently followed family advice for healthcare provided by males and elders, generally against screening, and felt the need to avoid conflict within the family associated with assertions of independence 41 46 .
Another study showed that women felt family members to be supportive 49 .
The process of Cultural Adaptation: South Asian nations have largely curative health systems, in which health costs are required to be paid by patients and there is no free access to healthcare. This contrasts with preventive healthcare models in host countries, and as with other health issues, South Asian women showed little understanding or orientation toward cancer prevention 46 , although this evolved with time as awareness of the culture of the host country increased 43 47 . Women appreciated healthcare professionals who understood and respected values of personal modesty/shyness 45 . South Asian women in Canada emphasised the value of being chaperoned to screening appointments that may have been located away from their local community, for assistance with language barriers, to alleviate feelings of personal vulnerability and to avoid being alone with doctors 41 .

Deductive Studies
Study Quality: Predictive and comparative studies contained similar limitations to quality.
Lower screening rates were noted among women with no health insurance, younger women and women with lower levels of education. Studies did not provide consistent evidence that low knowledge predicted reduced likelihood of attendance. Lower knowledge was associated with a reduced likelihood of mammography screening in two studies 61 62 , but did not predict the likelihood of hypothetical acceptance of a cervical screen 57 . Lower attendance was associated with a greater number of self-reported barriers to screening [61][62][63][64][65] although one study found the opposite 66 . However, the instruments used to assess barriers were largely based upon existing instruments developed among western samples that do not reflect South Asian concerns such as adapting to a new culture, language or health system.
Where acculturation was examined, less time spent in the host country was the strongest predictor of non-attendance, although one study cited lower preference for the host language (usually English) compared to women's native language 65 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  Indian and Pakistani women perceived fewer barriers arising from lack of knowledge about preventative screening than other immigrant groups, and highlighted that many of the women thought that routine blood tests and urine tests would detect broader health issues such as cancer 50 . In another study 68 , Indian women perceived themselves to be less vulnerable to getting breast cancer, did not view breast cancer as a serious illness and were more likely to claim that they did not know 'where to find a mammogram'.

Discussion
Prominent across study types were the findings that South Asian women had poorer understandings of cancer and cancer prevention and that they perceived greater cultural and structural barriers to screening than host country women.
Lack of understanding by South Asian women about the need for asymptomatic screening has important ramifications. Predictive studies showed greater knowledge to be associated with screening attendance. The inductive research yielded some plausible reasons for this. Many women held fatalistic views or beliefs that cancer is incurable, whilst others believed that cancers could be identified in routine health testing. Others were unaware of the existence of cervical cancer in particular, and did not perceive threat to themselves or their communities. The role of males was also important, with male family members sometimes negative about screening and women unwilling to provoke conflict within the family by attending. Whilst there is a clear need to change such beliefs, the inductive studies showed this to be a challenging task for two reasons. First, understandings were embedded within religious and cultural traditions, and cannot be addressed in isolation to those traditions.
Thus, a simple educational intervention is likely to have limited effect. Accommodations will need to be reached with communities that allow a creative integration of cancer awareness within existing belief structures. Second, some women were largely unaware of the concept of disease prevention. Thus, the promotion of specific cancer awareness and understandings are unlikely to be helpful until a wider understanding of prevention is reached.
Predictive studies showed the importance of perceived barriers (e.g. lack of education, no health insurance, no family history, lower mammogram importance, less years living in host country, unmarried, language barriers, low self and outcome efficacy for screening), but these barriers pertained only to generic barriers faced by either all women or all immigrant women, irrespective of culture. Acculturation, in terms of time spent in the host country and  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   22 mastery of the language was associated with increased screening likelihood, but these issues are likely to exist for all immigrant women and fail to reveal specifically South Asian issues.
Inductive studies provided more subtle and culture-specific indications of the barriers perceived by women. Many were cultural. In particular, women spoke of the importance of female modesty and stigma associated with cancer that also affected willingness to be screened. Whilst the importance of female testing staff from South Asian backgrounds and use of South Asian chaperones is emphasised, this cannot address the wider cultural issues of modesty and stigma. One finding that offers encouragement is that personal health is important to South Asian women because it helps them to care for their families.
Interventions will need to be conducted more widely than merely targeting women and their beliefs. Males occupy decision-making roles in some South Asian families and women may not wish to challenge this (see also 78 79 ). Thus, addressing the views of male family members and other community opinion leaders is also important.

Limitations
The following limitations were identified within the review. First, many of the included studies were conducted in the USA, where screening services can require payment, which may not be comparable to other health services. Second, due to the small number of published studies, it is difficult to identify factors unique to groups of South Asian women based on nationality, geographical region or religion. By necessity, we discuss findings in terms of a generic 'South Asian' population, but are aware of variance between South Asian populations according to nationality, region, culture and religion. Finally, few studies used sampling techniques that are population representative, employing samples based around community activities. This may introduce unknown biases in findings associated with nonsampling of women who are less likely to attend such activities. group or as an immigrant group. The former can comprise women with high degrees of familiarity with the host country, but who nonetheless may be faced with cultural barriers deriving from their countries of origin. The latter group will reflect the problems of adjustment faced by recent immigrants. Studies will also need to use populationrepresentative sampling techniques.

Recommendations for Practice
Findings from all study types demonstrate that interventions should be sensitive to cultural norms. In particular, studies emphasised the importance of language, female practitioners and the importance of community approval and involvement. Interventions at the community level will be necessary to surmount the cultural barriers identified in the inductive studies.
It is worrying that the findings indicated that younger women and women with lower levels of education were less likely to attend for screening. There is some evidence that South Asian women might experience breast cancer at an earlier age 80 , thus interventions may need to be targeted at educating South Asian women who are younger. Encouraging female family members to become more involved as chaperones and translators could also be helpful, and may form a mechanism for educating young women simultaneously. Encouraging female family members to become more involved as chaperones and translators could also be helpful, and may form a mechanism for educating young women simultaneously. Interventions to increase uptake rates need to be long-term, multifaceted and tailored to the specific needs of the local community by, for example, developing close links with the community through Health Education workers. South Asian community members, including males and opinion leaders, should be encouraged to be involved and co-produce engagement strategies within community settings. Reducing ethnic inequalities in uptake rates of breast cancer screening needs to remain a policy priority of breast screening programmes.    1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47 o n l y 37  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47 o n l y 44  Pre and post intervention The Asian Grocery Store based cancer intervention program -incorporating an educational program into women's routine shopping activities Barriers to mammography screening Shift toward screening uptake for Chinese and Vietnamese American women who were nonadherent at baseline but no change for Asian Indian and Japanese American women at follow up.
Time series of service use. Community initiative led by SA community health nurses in collaboration with influential women in the SA community, local physicians, and health board authorities. Qualitative interviews with 20 women who attended the Pap test clinic

Knowledge of early intervention
Only 6 (5%) of those not contacted and 14 (11%) of those sent leaflets had a smear test during the study. Health education interventions increased the uptake of cervical cytology among women in Leicester who had never been tested. Visits and videos were most effective.

Ornelas et al, 2017
Greater Seattle US 40 SA women, 20 Karen-Burmese and 20 Nepali-Bhutanese; 21-58 years (mean age = 35years); living in US for 5 years on average. Most did not speak English well or at all (75%); 8 years average of education; 65% married.73% had Pap test since arriving to US, 70% in last 3 years.
Pre and post survey The two health educators recruited participants through personal contacts they had in their community, as well as referrals from community advisors and participants with whom they had completed data collection. A pilot study to evaluate the acceptability and efficacy of the 17 minute videos provided in their native language.
Behavioural Model Changes in cervical cancer awareness, intention to be screened for cervical cancer, cervical cancerrelated knowledge Nepali-Bhutanese were significantly more likely to have been screened than Karen-Burmese (90% v 55%). Women showed significant increases in knowledge for all the individual items, as well as the mean composite knowledge scores (5.6 to 9.3, p < .001) after viewing the video. There were also increased in knowledge for individual items across ethnic groups; however, not all were significant. Mean changes in the knowledge score were significant for women in each ethnic group (5.4 to 9.2, p < .001 for Karen-Burmese and 5.8 to 9.5, p < .001 for Nepali-Bhutanese). Women indicated high satisfaction with the video length and very few women reported about anything they did not like. Most SA women spoken to by a SA HA indicated a willingness to get screened for breast or cervical cancer and some went on to action their screening intention. Making phone calls to patients to invite them for screening had the most reach and most appeal. The initiatives were reported to be resource intensive for physicians even with voluntary SA HAs involved. However, using SA HAs showed promise to increase awareness and willingness    1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60