Table 1. Summary of included studies.

Author, year, location, settings	Study design Data collection and/or method of analysis	Total participants Ethnic Groups (number and/or %)	Cancer genetic service	Key findings	Study limitations Other comments
Qualitative studies
Atkin et al 22 (2009) UK Primary care	Qualitative in-depth interviews. Thematic analysis	52 women diagnosed with breast cancer and 5 people with other cancers Not all participants had accessed genetic service South Asians (n=35) White British (n=22)	Primary care-based genetic service led by GP (family physician) with Special Interest in Genetics and genetic counsellor Full genetic counselling provided in community setting	• South Asians: lacked awareness of service; cultural taboos and stigma prevented people speaking to family members about cancer and inheritance patterns • Knowledge of cancer appeared lower among South Asians compared with white participants • South Asians accessing the service: experienced poor communication from professionals and wanted a more direct approach to receiving information, which they found vague; and decision-making could be compromised when interpreters are involved	Included 11 interviewed in languages other than English Qualitative study linked to observational study of service29
Ford et al23 (2007) USA Secondary care	Focus groups Content analysis	20 women with above average risk for breast cancer aged ≤50 years African American (n=13) White ‘Caucasian'(n=7)	Integrated primary and specialty service	• Barriers expressed by women who did not receive genetic counselling included: stigma; uncertainty if breast cancer could be prevented by genetic testing; and failure to see the benefits of genetic counselling • Women who received cancer genetic counselling expressed: increased fear and worry about being diagnosed with breast cancer and associated it with death; concern and mistrust about revealing family health information; and faith in the positive role of God in engaging in risk assessment	People without medical insurance or under-insured were not represented in sample
Matthews et al24 2000 USA Secondary care	Focus groups Thematic analysis	Women (n=13) with either personal or family history of breast/ovarian, colon, or prostate cancer Men (n=8) without personal or family history of these cancers All African descent (n=21)	Familial cancer risk assessment service	• Cancer equated with ‘death sentence' and fatalistic views led to screening avoidance • Participation in genetic testing prevented by: distrust and fear of hospitals, modern medicine, being experimented on; anticipated negative emotions/anxiety • Lack of discussion in families of cancer-related issues • Mistrust of use of information • Lengthy process of genetic counselling/testing and complex educational material acted as barriers	Analysis does not differentiate between participants with cancer and those referred to cancer genetics due to familial risk Included cross-sectional patient survey (see below)24
Quantitative studies
Fraser et al26 (2003) UK Secondary care	Cross-sectional self-completed survey (prior to attending clinic)	162 newly referred patients 2.5% reported ethnicity as non-white	Five regional cancer genetics service in England	• Low proportion of referrals involving non-white patients • 55% patients instigated and sought genetics referral from a provider themselves • 52% referred to genetics service by primary care provider (GP) • 38% referred to genetics service by hospital doctors • 2.5% self-referred to genetics service
Wonderling et al1 (2001) UK Secondary care	Cross-sectional self-completed survey (attending clinic) Provider survey of activity during 4 week period	Patients n=869 97% of service users: white Providers n=22	UK Cancer Genetic Services in 1998	• Only 3% reported ethnicity as non-white (almost half-Jewish ancestry) • 49% referrals from GPs • 47% referrals from hospital doctors • Some self-referrals	Little descriptive data for ethnic minorities
Sussner et al27 (2009) USA Secondary care	Cross-sectional Patient telephone survey Univariate and multivariate analysis of predictors of perceived barriers to genetic testing	146 women at increased risk of breast/ovarian cancer 100% African descent (56% US born)	Cancer genetic clinic in New York area	• Perceived barriers to genetic testing (as indicated by avoidance of breast cancer distress symptoms) include confidentiality concerns and family-related guilt Further associations include: • Stigma (with age a strong predictor) • Education (and note foreign-born had less educations) • More anticipation of negative emotions and distress associated with genetic testing for foreign-born women	Small sample size and cohort with higher educational, medical insurance and income status
Gulzar et al25 (2007) UK Primary and secondary care	Cross-sectional self-completed satisfaction survey GP satisfaction survey to 5 primary care clinics Routine data collection of referral activity Descriptive analysis	Patients (n=81) GPs (n=18)	Nurse-led cancer genetic clinics: four primary care-based and two hospital-based in London area (self-referral and proactive identification by GPs)	Overall, of 137 of referrals 17% involved people from ethnic minorities and of these 74% (n=17) self-referred compared to 89% (n=102) white self-referral Patient satisfaction survey: all respondents stated English as language of choice, language not mentioned as a barrier From two focus groups with 9 participants issues were raised regarding their post-clinic letter • Some patients wanted more statistical facts presented • Some were unwilling to share the information in the letter with their family	Minority ethnic composition of catchment population unclear Information for specific minority groups not available from patient or GP survey Project actively sought to recruit patients from minority populations (eg literature and counsellors multilingual) and in London districts with high proportion of ethnic minorities
Srinivasa et al29 (2007) UK Primary care	Before and after observation study Descriptive analysis	Referred patients No data on ethnicity of service users. 7% of the region ethnic minorities, predominately South Asian, with greater proportion in study area	Primary care-based genetic service led by GP with Special Interest in Genetics and genetic counsellor in North Kirklees district of Yorkshire, England 2004–2006	Ethnic minorities accessing regional genetic service increased from 0% at baseline to 6% at 2 years after introduction of the service. The 6% comprised six families	Percentages difficult to interpret as no denominator supplied at baseline or follow-up Linked to qualitative study above (Atkin et al22) Modest uptake for amount of resources deployed in demonstration project
Jacobs et al30 (2007) UK Primary care (and community settings)	Cross-sectional Patient self-completed satisfaction survey Routine data collection of referral activity	Patients (n=71)	Nurse-led familial cancer risk assessment service in the local community in two districts of London Patient access service through self-referral or through community-based health professionals 2005–2006	Patient survey: • 99% found service helpful • 96% found easy to access location • 94% found appointment time convenient • Ethnicity not reported Routine data: 415 patients seen in community clinics 46% referral non-White, 30% of these patients defined their origin as one of Black ethnic groups 194 patients of 415 seen were assessed as moderate or high risk of breast cancer/complex family history and 70/415 patients at moderate or high risk of colorectal cancer/complex family history. Ethnicity not reported	Although proportion of ethnic minorities recruited not identified, study was in a district with a high proportion of ethnic minorities (47% residents African Caribbean and other ethnic groups) People were invited to phone in to make an appointment which may have been a barrier for non-English-speaking members of the community
Hughes et al34 (2003) USA Secondary care	Structured telephone interview preceding genetic risk assessment for breast cancer Data from study records Descriptive analysis	28 self-referred women with >10 probability of BRACA1/2 gene 100% African descent	Genetic counselling and testing service at a cancer centre in Washington	61% (17) participated in risk assessment Comparing those that participated in genetic risk assessment to non responders • More fatalistic • More positive about preparing for the future (positive temporal orientation) • When considering using faith to cope with difficult situations: 70% of those that used faith participated in genetic risk assessment whereas this fell to 20% in those who did not utilise this approach Further: 41% who endorsed familial interdependence received BRCA1/2 test results (acceptors) compared to 91% who did not endorse this belief	Small sample size and recruited older, higher income and well educated women
Culver et al28 (2001) USA Secondary care	Interviewer-completed patient survey Descriptive analysis and logistic regression analysis of the predictors for accepting genetic counselling	97 self-referred women, recruited through paper and internet publicity, both with or without; European Americans (n=37), African Americans (n=15), Native Americans (n=15) and Ashkenazi Jewish (n=30)	Offered free genetic counselling as participants identified at increased familial risk for breast cancer Seattle area 1996–1998	Sociodemographic profile • Correlation between educational level and acceptance of genetic counselling • Only 13% of African Americans had completed college compared with 73% Ashkenazi Jewish women Uptake of service • In recruited women no difference in acceptance rate for genetic counselling (overall 52%) between ethnic groups, when adjusted for education status. Interest in genetic counselling was demonstrated for all ethnicities	Self-referred sample not representative of population Accepted 5 women with education beyond undergraduate degree and 1 with less than a high school degree following attempt to recruit women within restricted educational criteria
Matthews et al24 (2000) USA Secondary care	Cross-sectional postal survey Descriptive analysis	Women (n=13) with either personal or family history of breast/ovarian, colon, or prostate cancer Men (n=8) without personal or family history of these cancers. Participants recruited through paper and radio publicity 100% African descent	Familial cancer risk assessment service	Strong family history of cancer in recruited cohort (25% had 2+ affected relatives) but only: 10 (48%) of 21 discussed cancer within the family: 14 (67%) had annual medical check up; 12 (57%) had annual cancer screen Awareness of increased morbidity for breast and prostate cancer in African Americans: 13 (62%) knew risks in men and 14 (67%) in women Important factors affecting decision to participate in genetic testing: risk of cancer for family members; effects on family/themselves of test results; knowledge of whether other family members need testing; future planning; and test accuracy	Self-referred participants and small sample size Qualitative results reported above (see Matthews et al24)