Abstract
Objectives:
Patient and provider barriers to cancer screening in disadvantaged Asian populations are understudied. We conducted a qualitative study of attitudes to screening for colorectal, cervical and breast cancer within low-income communities in Singapore.
Methods:
Interviewers elicited perceptions of barriers/enablers to cancer screening amongst patients eligible for fecal occult blood test (FOBT)/Pap smear/mammograms, and providers serving these low-income communities. Interview transcripts were analyzed thematically using established qualitative methodology.
Results:
Twenty patients and nine providers were interviewed. Patient and provider comments were grouped into seven content areas: primary care characteristics (PCC), procedural issues, knowledge, costs, priorities, attitudes, and information sources. For FOBT, the most frequently mentioned content areas were knowledge (61.2%) and attitudes (16.9%) for patients, and knowledge (19.5%) and PCC (18.6%) for providers. For Pap smears, it was knowledge (23.6%) and costs (20.5%) for patients, and PCC (33.8%) and knowledge (19.1%) for providers. For mammograms, it was knowledge (41.4%) and sources of information (15.9%) for patients, and knowledge (28.8%) and procedural issues (19.2%) for providers.
Conclusion:
While PCC, knowledge and costs were the main contributors as perceived by physicians, knowledge and attitudes were the biggest barriers from patients’ perspectives. Differences in perceptions of cancer screening exist between low-income patients and providers serving them.
Introduction
Inequities in cancer screening access exist in many countries, particularly in disadvantaged communities. 1 However, in Asian societies, although socioeconomic status (SES) and screening participation have been linked in several studies,2,3 few studies focus on screening participation amongst low-SES communities.4–7 Equity in access to screening is important given rising income inequality in urbanizing Asian societies. 8
Singapore is one such multi-ethnic urbanized Asian society. Locally, cancer contributed about one-fifth of all disability-adjusted-life-years (DALYs) lost 9 and colorectal and breast cancers are the most common cancers locally. 10 National population-based screening programmes have been initiated for breast cancer, 11 cervical cancer, 12 and recently, colorectal cancer. 13 In order to reach the disadvantaged, screenings are subsidized and in 2011, the use of Medisave (a national compulsory healthcare savings scheme previously only reserved for defraying hospitalization costs) for screening was approved, further reducing out-of-pocket payments. 14 However, in our previous cross-sectional studies, participation was still much lower in poorer communities compared with higher-income neighbors.15,16 Although interventions providing free, multimodality screening door-to-door in these communities increased screening participation, 16 participation was still lower compared with more well-to-do counterparts. 17 The reasons for these disparities are likely to be multifactorial. Apart from patients’ perceptions and beliefs of health screening, providers’ attitudes are also important. In the local population, patients with a regular primary care doctor were more likely to be aware of cancer screening and to participate in screening. 18 However, there are no local studies on provider perceptions and attitudes toward health screening. There is also a paucity of qualitative studies on attitudes to cancer screening in Asian populations. The bulk of the literature is focused on underserved Asian immigrant populations in Western countries,19–21 where findings may not be easily generalizable due to different sociocultural contexts. Only a limited number of qualitative studies have explored patients’ perceptions of cancer screening in Asian countries,22–24 with only one study exploring providers’ perceptions 25 and none focusing on disadvantaged settings.
As such, we conducted a qualitative study of patient and provider attitudes to screening for three cancers (colorectal, breast and cervical) within low-SES public rental flat communities in Singapore to better understand promoters and barriers to cancer screening in similar Asian resource-poor urban settings.
Methodology
Setting and recruitment
Patients were recruited via purposive sampling techniques from a pre-existing study population staying in two rental flat communities in Singapore. 17 Public rental flats are a good marker of SES, which is closely tied to home ownership in Singapore. The majority of Singaporeans (⩾85%) stay in public housing and due to government subsidies, home ownership is high (90.1%).26,27 Public rental flats provide heavily subsidized rentals (S$26–S$275/month) 27 for the needy. Respondents were chosen to ensure roughly similar proportions of gender and ethnic groups compared with the population at large, with roughly equal numbers of younger (aged 40–59 years) and older (aged ⩾60) participants. These two sites contained all rental blocks in their respective estates and were in the eastern and western zones of Singapore, respectively, which have the largest number of such blocks. 27 Inclusion criteria included: being Singaporean citizens/permanent residents, aged ⩾40 years, having lived in the community for at least three years, and no pre-existing cancer history. Patient participants were recruited via letters of invitation, interviewed at home visits, and were reimbursed $10 for their contributions. This study was approved by the National University of Singapore’s Institutional Review Board, and written informed consent was sought from participants. We had equal numbers of patients who had participated in at least one screening modality, and those who had not participated in any screening.
Separately, healthcare providers were recruited via purposive sampling to represent various organizations providing medical services in these communities. In Singapore’s healthcare system, the bulk of primary care is delivered via public primary care clinics called polyclinics, as well as private general practitioner (GP) clinics. 28 Tertiary hospitals handle more complex cases which need specialized referrals and treatment. In addition, within these low-income communities, various programs such as free clinics run by voluntary welfare organizations and door-to-door medical consultations provided by grassroots organizations 29 fill gaps in healthcare coverage. We recruited representatives of these organizations to obtain their perspectives. To be included in the study, they must have been in their current roles for at least two years and be directly involved in patient care in these communities. Provider participants were recruited via letters of invitation and were not reimbursed. Additional residents and providers were interviewed until saturation was reached (i.e. when no new themes emerged from the data analysis). 30
Conduct of interview sessions
Individual in-depth interviews (approximately an hour each) for each participant were carried out in the residents’ homes for patients, and at the offices of healthcare providers. Interviewers were medical students who underwent qualitative research training by the senior author prior to study commencement. To ensure that interviewers were more attuned to the sociocultural milieu that their patients came from, patient interviewers must have had at least one year of participation in service-learning programs involving these needy communities.29,31 We also matched interviewers to patient interviewees by race and language, so that the interview could be conducted in the interviewee’s first language. Interviewers used an interview guide developed by the investigators. Information was collected via a series of open-ended questions (Table 1) in English, Chinese and Malay for residents, and in English for healthcare providers, to elicit their feelings about, and experiences with, cancer screening using three screening modalities (i.e. fecal occult blood test (FOBT) for colorectal cancer, Pap smears for cervical cancer and mammograms for breast cancer). All residents were asked questions pertaining to general attitudes toward cancer and screening tests (Table 1, Section A). Residents who were eligible for the various cancer screening modalities were asked questions about the corresponding screening modality (Table 1, Section B). Eligibility was determined based on the local Ministry of Health’s guidelines for health screening. 32 For care providers, similar questions were asked about what they thought their patients’ ideas and thoughts on cancer screening were (Table 1, Section B). Throughout the interview process, interviewers performed member checking with interviewees by paraphrasing and summarizing to clarify points brought up.
Interview guide for residents and healthcare providers.
Qualitative content analysis
Iterative content analysis of the verbatim transcripts of the audiotaped interviews with patients and providers was carried out to identify barriers and enablers to screening participation in the three modalities. For the initial transcripts, pairs of investigators identified and highlighted every codable “unit of text” in the transcripts that represented a singular idea. Each unit of text in the transcripts was then reviewed and a list of themes representing distinct barriers and enablers to screening participation was created from each transcript. All investigators then met to discuss the collated lists of themes and produce a master list comprising all the unique themes identified. The master list was then used to pilot code one patient and one provider manuscript, and any issues encountered were discussed, seeking consensus and refining the master list. All accumulated transcripts were then recoded using the master list. Throughout the coding process, the team met regularly and this process was repeated multiple times, allowing investigators to suggest the addition of new themes to the master list as they arose. The final master list was then used by two senior investigators in the final phase of coding, independently reviewing all transcripts and recoding them according to the finalized list of themes. Finally, two senior investigators met to compare their recoded transcripts, and where there were differences, divergences were resolved through consensus.30,33
Results
Participants’ characteristics
In total there were 29 participants (20 patients, nine providers). Participants’ characteristics are shown in Table 2. Of the patients, the majority were Chinese (85%) which reflects the ethnic distribution of the Singapore population. Slightly under half (45%) were aged ⩾60 years old and a similar proportion were female. These patients were from the lower socioeconomic strata: only a quarter had finished secondary education; two-thirds were currently unemployed, and all had a monthly household income of ⩽$1500/month (compared with the average monthly household income of $7570 34 ). Of the providers, all were healthcare professionals, with the majority being doctors, working in a variety of roles and locations.
Sociodemographic characteristics of study participants, comprising residents staying in two rental flat communities in Singapore (
Receiving financial help was defined as: being a recipient of Public Assistance (a national scheme which provides a monthly allowance to indigent citizens who are unable to work due to old age, illness or disability, have no means of subsistence and have little or no family support) or other forms of financial assistance from the local community center (grassroots organization).
Major content areas
For each of the three cancer screening modalities, patient and provider comments fell into seven content areas: primary care characteristics, procedural issues related to screening, knowledge, costs, priorities, attitudes, and information sources. These content areas are listed in Table 3. We also listed the themes within each content area according to how frequently participants mentioned them, and whether they were mentioned as barriers or promoters of screening. For colorectal cancer screening, amongst patients, the most frequently mentioned content areas were knowledge (61.2%), attitudes (16.9%) and procedural issues (15.1%); whereas for providers, the most frequently mentioned content areas were knowledge (19.5%), primary care characteristics (18.6%) and costs (15.9%). For cervical cancer screening, amongst patients, the most frequently mentioned content areas were knowledge (23.6%), costs (20.5%) and attitudes (17.3%); whereas amongst providers, the most frequently mentioned content areas were primary care characteristics (33.8%), knowledge (19.1%) and costs (16.1%). For breast cancer screening, amongst patients, the most frequently mentioned content areas were knowledge (41.4%), sources of information (15.9%) and attitudes (14.6%); amongst providers, the most frequently mentioned content areas were knowledge (28.8%), procedural issues (19.2%) and primary care characteristics (17.8%). Representative quotations of the various content areas are presented in Supplementary Table 1 (patients) and Supplementary Table 2 (providers).
Promoters and barriers to cancer screening identified by patients and providers from two rental flat communities in Singapore from 2009–2011 (
Percentage of times comments fell into this content area, out of the total number of comments; an individual can mention the same content area more than once in the course of an interview.
= Indication that this particular barrier/promoter was mentioned by either patients, providers, or both.
For primary care characteristics, when considering colorectal cancer screening using the FOBT, issues raised by patients centred around elements of the patient–provider relationship—trust in their doctor, reminders from their doctor—whereas for breast and cervical cancer screening, patients considered not just the patient–provider relationship but also whether access to primary care was convenient (Table 3). This was in line with the nature of FOBT as a mail-in test, compared with Pap smears and mammograms which necessitated the patient going down to the screening center. Providers’ perceptions of primary care characteristics largely overlapped with the themes identified by patients, but they also provided additional insight into the barriers/enablers they faced when considering whether to recommend screening to these needy patients (e.g. the need for them to remember eligibility criteria for screening, checklists/prompts to remind them to broach the subject of screening in clinic) (Table 3).
For procedural issues, patients and providers largely concurred on the various aspects of each screening procedure that acted as barriers/enablers to doing the tests (Table 3). Procedural issues identified were modality-specific (e.g. fear of handling faecal material for FOBT; pain and discomfort for mammograms/Pap smears). The same procedural issue could be both a barrier and an enabler (e.g. fear of pain during the test was a barrier, but at the same time patients also brought up that, having experienced the test, they would be open to doing it again as it was less painful than they had imagined) (Supplementary Table 1, Supplementary Table 2).
With regards to knowledge, the main misperceptions were identified by both patients and providers (e.g. “I don’t need screening as I am healthy”; “I’m not aware of screening”) (Table 3). However, providers were more focused on knowledge/ increased awareness as an enabler, whereas for patients lack of knowledge was a barrier. For costs, again patients and providers concurred that the cost of screening and further treatment, if the screening test was positive, were barriers to cancer screening; while providing free/subsidized screening was viewed as an enabler (Table 3). Patients and providers also identified that, given many other pressing issues that these low-income residents faced, such as family priorities, job security, and fulfilling the breadwinner role (Supplementary Table 1, Supplementary Table 2), health screening might not necessarily be patients’ top priority. In terms of attitudes, both patients and providers concurred on the main negative attitudes to screening (Table 3)—that screening was not needed because disease was something that was fated and hence something that was not preventable, or that patients stayed away from screening because of a fear of cancer, given its poor prognosis and potentially distressing side effects of treatment (Supplementary Table 1, Supplementary Table 2). Both patients and providers identified the same sources of information about cancer and cancer screening: word-of-mouth via friends, relatives and family, media outlets, such as television and newspapers, and community outreach efforts, such as talks, posters and flyers (Table 3). However, patients also indicated that sources of information could be conduits for both encouraging and discouraging screening participation. In particular, because they trusted their loved ones and friends, they tended to adopt their friends’/loved ones’ attitudes to screening, regardless of whether those attitudes were positive or negative (Supplementary Table 1, Supplementary Table 2).
Age differences among patients
Elderly participants aged ⩾60 tended to bring out fatalism as one of the barriers more frequently than their younger counterparts. Younger residents tended to discuss lack of priority and time as barriers instead:
I don’t think I need this test because I am old, it doesn’t matter anymore. I’m not scared of getting illness, If get these nothing can be done anyway. (Mdm A, 78 years old) I work half days on Saturdays and they don’t open on Sundays. (Mr B, 51 years old) Because of time and we housewives are very busy with housework and the kids, so I don’t go. (Mdm C, 40 years old)
Discordance between patients and providers
The most frequently mentioned content areas were very different between patients and providers. For colorectal cancer, aspects of primary care—such as initiation of discussion about screening by the healthcare provider, relationships and communication between the provider and patient—were frequently mentioned by providers as important factors influencing screening. However, only a very small minority of comments by patients were related to primary care characteristics (18.6% vs. 2.9%, respectively). Similarly, for cervical cancer and breast cancer, a significant proportion of providers’ comments were related to primary care characteristics, but this content area formed only a small minority of patients’ comments (cervical cancer: 33.8% vs. 11.0% respectively; breast cancer: 17.8% vs. 10.2% respectively). With regards to procedural issues relating to screening, while aspects of the screening process—for example, painful test, bringing convenient access via mobile screening—were mentioned in 19.2% of providers’ comments, procedural issues formed only 3.8% of patients’ comments (largely concerns about pain). While knowledge was mentioned by both providers and patients alike as a significant factor influencing screening, it formed a much larger proportion of patients’ comments compared with providers’ comments. While providers tended to focus on knowledge as an enabler, in which increased awareness of screening would result in increased screening participation, knowledge was a barrier for most patients. Misperceptions, such as breast self-examination being an adequate substitute for mammograms, or that mammograms cause cancer, also formed a significant proportion of themes for non-participation in mammograms. For attitudes, while a large proportion of patients’ comments reflected negative attitudes like fear and fatalism (17.3% for cervical cancer, 14.6% for breast cancer), only a small proportion of providers’ comments (8.8% for cervical cancer, 5.5% for breast cancer) reflected that attitudes towards screening might be an issue. For sources of information, in cervical cancer screening, while patients highlighted that word-of-mouth and experiences about screening narrated by their friends, relatives or family were the more significant sources, providers thought that patients were influenced more by media and public outreach efforts.
Discussion
Previous work has established that disparities in access to screening do exist in Singapore,15,16 despite access-enhancing interventions, structured national multi-disease screening programmes, and the availability of subsidized screening for the disadvantaged. Nationally 9.1% had had FOBT, 59.2% had Pap smears and 40.9% had mammograms according to national guidelines. 35 In our population of low-income Singaporeans staying in public rental flats, only 7.7% had FOBT, 20.4% had gone for Pap smears and 14.3% had mammograms according to national guidelines. 17 The causes of this disparity are likely to be multifactorial in nature. Amongst patients, knowledge and attitudes featured prominently as reasons for not participating in screening, across all three modalities. Similar to findings in other studies,24,36 many felt they were not “at risk” because they lacked signs and symptoms of cancer and therefore did not need screening. They did not understand that screening detected early, pre-symptomatic disease and did not grasp the concept of secondary prevention. Attitudes such as fear and fatalism were also prominent barriers to screening in this underserved community, as demonstrated in these results and elsewhere. 37 In better-educated populations, fear was actually perceived as an enabler, in that patients’ fear of getting cancer pushed them to seek screening and early detection. 38 Given the gaps in knowledge in these communities, however, without first communicating that screening allows for early detection and prevention, increased fear of cancer is unlikely to encourage members of this community to go for screening.
Our results showed that patients and providers saw enablers and barriers to cancer screening somewhat differently. In particular, for all three modalities, while providers felt that primary care characteristics were important factors influencing screening participation, only a very small proportion of patients commented on primary care characteristics. These findings differ from those seen in Western populations, where both patients and providers identified aspects of primary care as important factors in encouraging members of underserved populations to go for colorectal cancer screening,37,39,40 cervical cancer screening, 41 and breast cancer screening.41 –43 We hypothesize that this could be due to several reasons. Perhaps convenience was less of an issue in highly urbanized Singapore, where primary care clinics are situated in close proximity to patients’ homes and linked by a dense public transport network. In the two sites of recruitment, one site had a private GP clinic 0.3 km away and the polyclinic was 2 km away, a ten minute journey by bus; in the other site, there were two private GP clinics, one immediately adjacent and one 0.5 km away; the polyclinic was reachable on foot, also 0.5 km away. Another factor could be lower levels of trust between primary care physicians and their patients, with a lack of trust cited by low-income patients as a reason for not going for cancer screening across all modalities, regardless of whether national subsidized cancer screening programs mainly utilized private GPs or polyclinics. Our previous studies in low-income populations 44 found that some low-income patients felt primary care practitioners were not fully in tune with their social circumstances, leading to care plans that might not be fully tailored to their best interests. This was due to physical distance and psychological distance. In addition, a survey of local attitudes towards doctors revealed that around 40% had low-to-neutral trust in their primary care doctor, and while patients rated their doctors highly on expertise, they scored less well on communication and cooperation, 45 which are important aspects of encouraging behavioral change in patients.39,46,47 While there are few cross-cultural studies of trust in primary care, another study found low levels of trust in physicians amongst urbanized Chinese populations, 48 and one found that Japanese trusted their physicians in Japan less compared with Japanese-Americans in the US, postulating that decreased cultural familiarity with their Western-trained doctors might play a role. 49 Although most local screening programs are run at the national level, they still involve primary care physicians to some extent (e.g. in the usage of resources, such as the utilization of polyclinic facilities; or in the explanation of results and their significance). Furthermore, upon receipt of the letter encouraging screening participation from national screening organizations, low-income residents usually may need additional convincing/explanation; without a primary care provider that they trust and can turn to for consultation regarding these matters, they may not participate. Finally, deficits in knowledge about cancer screening and its utility meant that the majority of the population might have been in the pre-contemplative phase of behaviour change. Elements of primary care, such as convenience and having a trusted healthcare provider, were thus less important because a significant proportion had yet to accept the necessity of screening or appreciate that they were at risk of cancer.18,50,51
In this underserved population, both providers and patients acknowledged the importance of family and friends in influencing screening. While providers mainly focused on the positive role, patients acknowledged that the influence of social networks could either promote or discourage screening. Other studies show that social networks can promote screening outside of the medical consultation and physician encounter, 52 especially in underserved populations53,54 with little physician contact and high illiteracy. To effectively utilize social networks in encouraging cancer screening in low-income Asian communities, education efforts have to include not just patients, but also their friends and family, in order to provide additional impetus for participating in screening.
Our study has its limitations. Using a qualitative approach allowed us to yield rich and detailed data on the perceptions of both patients and providers regarding screening, which can inform further quantitative studies. We recognize that these methods alone may have limited direct generalizability. However, qualitative research is useful to obtain an open and undirected response and is suitable as a pilot study to identify possible barriers and promoters to screening. Subsequently, we aim to use data gleaned from this study in identifying possible interventions to improve screening uptake in this low-income community. In addition, there is the possibility of researcher bias in interpreting our findings, which we sought to minimize through a step-wise approach to data analysis, with multiple iterations of checking and cross-checking amongst researchers. Finally, the private GPs contacted for this study did not respond, limiting our perspectives to healthcare professionals working in public institutions/non-profit groups. We note, however, that the majority of this low-income population would generally seek care from public primary care clinics where costs are cheaper, rather than from private ones.
In conclusion, in our study of patient–provider attitudes to cancer screening for colorectal, cervical and breast cancer in a medically underserved Asian community, we demonstrate that there are significant differences between patients and their physicians regarding perceived barriers and enablers to cancer screening. While primary care characteristics, knowledge and costs were the main factors as perceived by physicians, lack of knowledge and attitudes such as fatalism and fear of diagnosis were the biggest barriers from the perspective of patients. Given the gap between patients and providers, provider-based interventions to increase knowledge and boost screening in these disadvantaged communities may not achieve the same success as they have elsewhere.55–57 Given the role that social networks can play in encouraging screening in this community, a tripartite partnership between patients, their loved ones and their physicians might help raise screening rates amongst the medically underserved.
Footnotes
Acknowledgements
We thank the Neighborhood Health Screening Organizing Committee 2012; Yong Loo Lin School of Medicine, Saw Swee Hock School of Public Health; Taman Jurong Community Centre, Macpherson Community Centre, Singhealth Polyclinic (Geylang), National Healthcare Group Polyclinic (Jurong), for supporting this study.
Declaration of Conflicting Interest
The authors declare that there is no conflict of interest.
Funding
The Yong Loo Lin School of Medicine provided funding support for the reimbursement of patient participants.