Public interest in and acceptability of the prospect of risk-stratified screening for breast and prostate cancer

Abstract

Background. For risk-stratified screening to be implemented as a screening program for breast and prostate cancer it has to be accepted among the general population. Investigating public interest in stratified screening and its acceptability to the public is therefore essential since as yet little is known.

Method. Cross-sectional web survey sent to a sample of 10 000 individuals (20–74 years of age) representative of the Swedish population as registered in 2009.

Results. Among the responders (28%), a vast majority (94%) expressed an interest in knowing their breast or prostate cancer risk and stated wanting to know to ‘avoid worrying’. Men and women were equally interested in knowing their prostate and breast cancer risk, respectively. However, men showed more certainty. Trusting the healthcare workers with personal information (63%) as well as genetic information (70%), in order to calculate the risk, did not seem to be a major issue. Furthermore, 87% would agree to get screened more often if identified with a high risk, whereas, if identified with a low risk, only 27% would agree to get screened less often. Finally, although a consultation with a physician seemed to be the preferred way to communicate the risk, a majority would agree to receive it via a letter or a phone call.

Conclusion. Risk-stratified screening has the possibility to be accepted by the general public. Knowledge about interest and acceptability of the prospect of risk-stratified screening for breast and prostate cancer will help when implementing new screening strategies.

Research today is devoted to identify risk factors for breast and prostate cancers. Genetic and environmental risk factors can be used as a basis for risk-stratified screening. Cancer screening programs could likely be improved by using screening regimes stratified by risk instead of on the basis of sex and age alone [1–4]. This would include estimating the risk of individuals of developing breast or prostate cancer, grouping them according to their risk and targeting screening according to that risk [5]. Age is only one risk factor for cancer. There is a need to further refine screening by targeting groups with higher risk in order to have to screen fewer individuals with the same cancer detection rates. Thus, relying on age alone to target screening is not sufficient for identification of groups who might benefit from more or less intense screening and the tailoring of interventions to risk levels. However, the growing understanding of the etiology behind breast and prostate cancer may allow improvements in this respect [6]. Moreover, the introduction of personal genome sequencing will also play a role in determining and dividing individuals into higher or lower risk of cancer and thus be offered a corresponding screening strategy [7,8]. Screening programs stratified by risk would ensure that people at ‘higher’ risk are invited for screening, while sparing lower risk people from the inconvenience and harms of screening, thereby also reducing the overall costs as well as over-diagnosis. However, these new individualized screening programs will require a balance between healthcare resources and the patient's right not to know [9].

Risk-stratified screening programs imply that men and women will be asked to provide information on lifestyle along with a DNA sample. Even if screening would be shown to be clinically beneficial, uptake of risk-based screening programs would depend on the attitudes in the population. Interest in genetic testing and screening in the general population has proven to be high [10–14] whereas its actual uptake has been lower [15,16]. Public understanding and interest in participating are essential to the success of risk-based screening, because the individuals will have to understand what is being offered and why in order to consent to it. Moreover, acceptability of stratified screening depends on a recognition that this change in screening routines is in the public's interest. As yet, little is known of the public's reaction to the prospect of risk-based screening.

The aim of the present study was to assess the acceptance of risk-based screening in the general population. First, the interest of the population in knowing their individual breast or prostate cancer risk and the reasons behind their choice were evaluated. Second, the acceptability of risk-based screening and its implications were assessed in the general population.

Methods

Sample and non-respondents

A population-based, cross-sectional survey research design was used in a random sample of 10 000 individuals between 20 and 74 years of age selected from the Swedish population (SPAR registry). They were matched to the age and sex distribution of the Swedish population as registered in 2009 [Statistics Sweden].

Data collection was conducted using a web survey. Individuals in the sample identified were sent a letter informing them about the web survey and asking them to log in to the study web site. The heading of the letter read as follows: “Would you like to influence future mammography/prostate cancer screening?”. The concept of risk-stratified screening was presented in the letter by stating that increased knowledge about risk factors will make it possible in the future to select women and men at risk for breast and prostate cancer, and that it might be possible, based on a risk profile, to offer individualized screening program. The letter also included a log in to the web site. The research data were gathered between November 2011 and January 2012.

In order to study reasons for non-participation in the survey, a sample of 300 persons who were eligible but who did not respond to the web-based survey was randomly selected for telephone interviews. Their phone numbers were found through the telephone directory home page. Thus, only individuals listed there were identified. The interviewer tried to reach each person three times if no response was given. Respondents were asked about their reasons for not participating in the web-based survey study and if they would want to know their breast/prostate cancer risk. The study was approved by the regional ethics committee (DNR 2011/630-31/4).

Development of the questionnaire

A panel of clinical and epidemiologic experts developed the questionnaire to be used in the survey. The questionnaire used for women and breast cancer risk is very similar to the one used for men and prostate cancer risk (with the words breast used for women and prostate used for men, respectively). The original questionnaire was written in English and translated into Swedish. Pilot interviews with 20 healthy women and 10 healthy men were conducted according to the EORTC guidelines for questionnaire development [17] to ensure that the cultural and linguistic integrity of questions were maintained. Based on the results of these interviews the questionnaire was revised. The full questionnaire is available from the authors on request. The questionnaire was then adapted to a web-based interface. In order to test the web's functionality as well as the study logistics, a feasibility study was carried out on a random sample of 200 women from the Swedish population. The feasibility study did not reveal any functionality- or user-related problems, thus no further changes were made. The response rate reached during the feasibility study (30%) was used to calculate the size of the sample needed in the main study.

The Questionnaire (Supplementary Table I, available online at: http://informahealthcare.com/doi/abs/10.3109/0284186X.2015.1043024).

The final questionnaire consisted of two parts. Response alternatives were given on a four-point Likert scale, from Agree (or mostly) to Disagree (or mostly) and for some items the option ‘Do not know’ or ‘Either’ was added.

1. Public interest in cancer risk and underlying reasons. This part included three items regarding (a) their intention to know their cancer risk, (b) the underlying reasons for wanting to know their risk as well as (c) the underlying reasons for not wanting to know. Participants who answered ‘Disagree’ as to whether they would like to know their cancer risk or not in item (1a) were not asked to respond to item (1b). Similarly, item (1c) was not to be responded to by participants who answered ‘Agree’ in item (1a).

2. Public acceptability of the prospect of risk-based screening

1. Acceptability of a screening tailored to risk was assessed through four items. Items 2a and 2b assessed attitudes towards increasing and respectively decreasing the frequency of the screenings. Item 3 evaluated the intention to take a daily pill if identified as having a high cancer risk. The fourth item (4) is the only item throughout the entire questionnaire that differed according to gender. Since mammography screening is already in place in Sweden, the women were asked whether they would attend mammograms no matter how often they were asked to. As men in Sweden are not offered population-based prostate cancer screening, they were asked if they would attend a screening program.

2. Acceptability of the communication structures was evaluated by three items. Item 5a and 5b concerned the level of comfort in conveying personal and genetic information respectively to the healthcare system. Item 6 assessed, through a matrix, how and from whom the general population would want their cancer risk to be conveyed.

Demographic variables were assessed through self-reported personal characteristics (age, education, income and marital status) and health-related history (latest prostate-specific antigen or mammogram, previous cancer diagnosis and family history of cancer).

Data analysis

^χ2-tests were used to test for differences according to sex, age-group (20–39 years, 40–59 years or 60–74 years) and education level (low or high). Likert scales were in most cases summarized to a two-point scale: Agree (or mostly) and Disagree (or mostly). A Kolmogorov-Smirnov test was used to test for possible differences between the age distributions for respondents and non-respondents and a χ²-test for the sex distribution among respondents and non-respondents.

Results

Respondents and non-respondents

The overall response rate of the questionnaire was 28% (2822/10,000). Table I shows the characteristics of the respondents. Respondents resembled the non-respondents in age (p = 0.48) and sex (p = 0.49), which were the only demographic variables available for all non-respondents.

Table I. Characteristics of the respondents.

	Men n = 1384 %	Women n = 1438 %
Age
20–39 years	25	29
40–59 years	41	39
60–74 years	33	31
Education^a
Low (≤ 12 years)	52	43
High (> 12 years)	42	50
Income^a
Low (< 300 000 SEK)	30	46
High (≥ 300 000 SEK)	67	50
Marital status
Single/divorced/widow(er)	23	24
Married/cohabiting	77	76
Habitat
Urban (> 100 00 pers)	34	33
Rural (< 100 000 pers)	66	67
Performed a PSA test or a mammography during the last 2 years
Yes	24	65
No	76	35
First degree family history of BC/PC^a,b
Yes	11	13
No	80	84
Previous cancer
Yes	5	7
No	95	93

^atotal do not add up because category “Other/Do not know” not represented. ^b BC, breast cancer; PC, prostate cancer.

Of the 300 persons selected for the non- respondents telephone interview, 32% (n = 96) responded. A majority (75%) stated wanting to know their cancer risk. The most common reason for non-participation was ‘lack of time’ which was stated by 56% (n = 54) of the telephone interview respondents. However, 18 women (32%) stated that on top of lack of time, the subject induced discomfort and thus contributed to them not answering to the survey.

Public interest to cancer risk and underlying reasons

A vast majority (n = 2661, 94%) of the respondents showed interest (i.e. responded ‘Absolutely’ or ‘Yes maybe’) in wanting to know their cancer risk (Table II). However, female respondents were less certain than male respondents, with a higher proportion of women that responded ‘Yes, maybe’ (n = 399, 29% vs. n = 319, 22%, p < 0.0001) to this question. Neither age nor education predicted willingness to know cancer risk. However, an older age predicted being more certain of the standpoint whereas a higher education was associated with less certainty. The major reasons for wanting to know one's risk were to avoid worrying (n = 2493, 90%), to get a realistic view of the future (n = 2244, 83%) and to get rid of uncertainty (n = 2221, 82%) (Supplementary Table II, available online at: http://informahealthcare.com/doi/abs/10.3109/0284186X.2015.1043024).

Table II. Public interest to know their cancer risk and acceptability of a screening tailored to risk (n = 2822).

	Sex			Age-group^a				Education^b
Questions	Male n = 1438%	Female n = 1384%	p-Value	Young n = 765%	Intermediate n = 1139%	Old n = 915%	p-Value	Low n = 1334%	High n = 1298%	p-Value
Would like to know your BC/PC risk?^c			> 0.0001				0.006			> 0.0001
Absolutely	73	65		64	70	72		73	65
Maybe	22	29		31	25	21		22	29
Probably not	5	5		4	4	6		4	5
Definitely not	1	2		1	1	1		1	1
If identified as having a low risk, would you agree to undergo screening less often than average?			> 0.0001				> 0.0001			> 0.0001
Absolutely	34	20		31	28	23		25	30
Maybe	49	43		46	45	46		46	46
Probably not	11	22		16	15	17		16	15
Definitely not	6	16		7	12	14		13	8
If identified as having a high risk, would you agree to undergo screening more often than average?			> 0.0001				> 0.0001			0.14
Absolutely	83	90		83	88	88		86	88
Maybe	14	8		15	10	9		12	10
Probably not	2	1		1	2	2		2	2
Definitely not	1	1		> 1	> 1	1		1	> 1
Would you agree to take a daily pill with some side effect to lower your risk?			0.25				0.0003			0.02
Agree (or mostly)	90	89		86	90	92		91	88
Disagree (or mostly)	10	11		14	10	8		9	12

Percentages do not always add up to 100% due to rounding.

^aYoung = 20–39 years; Intermediate?40–59 years; Old = 60–74 years.

^bLow = primary and secondary school; High = University level. The category “Other” is not represented. Omitting them did not affect the p-values.

^cBC, breast cancer; PC, prostate cancer.

Public acceptability to the prospect of risk-based screening

To date, there is no organized national screening program for prostate cancer for Swedish men. Albeit, if such a program was introduced, 69% (n = 989) stated that they would like to participate and 28% (n = 396) expressed that they probably would. As national screening programs for breast cancer already exist in Sweden, women were instead asked if they would participate in screening programs no matter how often the mammograms where scheduled. A vast majority (80%, n = 1109 answered that they would and 17% (n = 230) that they might.

If predicted to have a high risk for breast or prostate cancer, regardless of sex, age or education, 87% (n = 2448) of the respondents would agree to be screened more often than average (Table II). However if predicted to have a low risk, only 34% (n = 492) of the men would agree to get screened less often than average and only 20% (n = 273) of the women. Moreover, younger and higher educated respondents were less likely to favor a reduction of the screenings frequency. Finally, when asked if they would take a daily pill to lower their risk, provided no major side effect, 89% (n = 2524) answered that they would.

A majority of the respondents answered being comfortable in conveying their personal information (63%, n = 1788) as well as their genetic information (70%, n = 1981) to the healthcare workers to assess their risk (Table III). However, young respondents appeared to trust the health care more than older respondents (66% vs. 60%, p < 0.001) when it comes to providing personal information whereas older respondents were more often answering that they neither agree nor disagree (32% vs. 23%, p < 0.001). Moreover, higher educated respondents held this opinion more often than lower educated respondents (68% vs. 59%, p < 0.0001).

Table III. Level of comfort in conveying personal and genetic information (n = 2822).

Questions	Sex		p-Value	Age group^a			p-Value	Education^b		p-Value
	Male n = 1438%	Female n = 1384%		Young n = 765%	Intermediate n = 1139%	Old n = 915%		Low n = 1334%	High n = 1298%
Would you be comfortable in conveying:			0.15				0.001			> 0.0001
– Personal information?
Agree (or mostly)	62	65		66	64	60		58	68
Disagree (or mostly)	9	9		11	9	8		9	9
Neither agree nor disagree	29	26		23	27	32		33	22
– Genetic information?			0.18				0.02			> 0.0001
Agree (or mostly)	71	69		72	70	69		69	72
Disagree (or mostly)	8	10		10	10	7		7	11
Neither agree nor disagree	21	21		18	20	24		24	18

Percentages do not always add up to 100% due to rounding.

^aYoung = 20–39 years; Intermediate?40–59 years; Old = 60–74 years

^bLow = primary and secondary school; High = University level. The category “Other” is not represented. Omitting them did not affect the p-values.

The preferred mode of being informed of one's cancer risk was via a consultation with a cancer specialist (93%, n = 2511) (data not shown). Nevertheless, mails and phone calls were options accepted by the respondents. No matter by whom the risk information would be conveyed (i.e. a physician, a nurse, a genetic counselor or a cancer specialist), over half of the respondents were positive (i.e. responded ‘Yes, absolutely’) as mail or phone call as a way to be communicated one's cancer risk (Figure 1).

Figure 1. Proportion who responded ‘Yes, definitely’ as to whether the mode of communication mail, phone call and consultation could be used to convey their risk.

Discussion

The results suggest that those who have responded were interested in and accepted the prospect of risk-based screenings, i.e. knowing one's risk, having to share personal and genetic information as well as having a screening frequency adapted to that risk. Although the response rate itself might have reflected a low interest in these issues in the population, the interview of the non-respondents pointed towards a shared interest in the matter, but a lack of time to respond to the web survey. However, the possibility to drawing conclusions on a population level is limited, as is also the generalizability of the findings due to the low response rate.

Although these data provide useful information on the current climate regarding risk-stratified screening in Sweden, there may be a discrepancy between the respondents’ standpoints based on hypothetical circumstances and their actual behavior when risk-based screening programs become reality. In addition, the degree of acceptability might have been overestimated due to the social desirability effect [18].

When considering the relationship between different factors and the acceptability of participating to risk-based screening programs, it is important to keep in mind that the temporal relationship cannot be established in a cross-sectional survey.

Worrying was the major reason for wanting to know their cancer risk followed by a willingness to get a realistic view of the future and a will to get rid of uncertainty. Cancer worry has previously been shown to predict a greater interest in genetic testing than risk perceptions [19]. In our study, concerns about cancer and the future were more often reasons for wanting to know one's risk than aspects concerning lifestyle and family. These findings might be explained by common misbelieves concerning genetic determinism and risk factors in the population. As discussed elsewhere, the factors driving the decision making process are not fully understood and further research is needed [12,20].

Risk-based screening, if adopted, would lead to changes in the frequency of the screening. A person with a high risk would be screened more often whereas a person with a lower risk would be screened less often. As suggested by Dent et al. [21], where there is already an established screening program, such as that for breast cancer, there may public opposition to a reduction of the screening offered to low risk groups because women could regard this reduction as service rationing. Similar problems are less likely with prostate cancer screening, where there is no existing screening program. The introduction of risk-based screening for prostate cancer might therefore be easier than the one for breast cancer screening programs. Dent's suggestion is confirmed by our results; a reduction of the screenings appeared to raise more opposition and hesitance among women than men. Moreover, the population seemed to agree to undergo screening more often, whereas only a third of the population would agree to undergo screening less often. The risks and benefits of screening were, however, not stated neither in the invitation letter, nor in the questionnaire. As for the question about taking a pill, this item only touched a much broader aspect of cancer prevention where, e.g. tamoxifen has been discussed as a breast cancer preventive agent for women at risk [22]. The intention was to get a pointer in the matter and to add some understanding as to what men and women were ready to do to prevent prostate and breast cancer, respectively.

A Swedish study showed that swedes trust researchers’ and healthcare personnel's capability in regulating or conducting genetic research [23]. Our results confirmed that the Swedish population trust the healthcare workers since a majority of the respondents expressed feeling comfortable in sharing personal and genetic information. Interestingly, a slightly larger number of people reported feeling comfortable in sharing genetic information (70%) than personal information (63%). Moreover, Swedes with a higher level of education tended to be more comfortable in sharing information than individuals with a lower education.

The study population was matched to the age and sex distribution of the Swedish population and randomly chosen in order to be representative this population. Due to the low response rate, however, the results might be affected by selection bias. It is possible that the respondents were more interested in and positive to cancer screening and therefore more prone to answer. Moreover, as this was a web survey, the level of computer experience might have biased the demographic of the respondents. Nevertheless the age repartition among the respondents and the non-respondents were similar, thus pointing towards a study sample representative of the Swedish population in terms of age and gender.

Stratified screening will most likely be implemented in the future. Our results indicated a high interest among both women and men in participation in future risk-stratified screening and that the majority accepted to convey genetic material as well as personal information in order to calculate their individual risk. A consultation with a physician appeared to be the preferred way of risk communication, but a majority would agree to receive it via a letter or a phone call. The study was, however, hampered by the low response rate, thus further studies are needed within this new field. Knowledge about interest and acceptability of the prospect of risk-stratified screening for breast and prostate cancer will help when implementing new screening strategies.

Supplementary material available online

Supplementary Table I and II available online at: http://informahealthcare.com/doi/abs/10.3109/0284186X.2015.1043024

Acknowledgments

We would like to express our gratitude to the individuals that participated in the study. We would also like to thank Mathilde Koitsalu for her precious help with the telephone interviews. This work was carried out under the auspices of the Cancer Risk Prediction and Prevention (CRisP) centre at Karolinska Institutet supported by the Swedish Research Council as a Linnaeus Centre. This work was funded by The Swedish Cancer Society (110742), King Gustaf V Jubilee Fund (124032) and The Swedish Council for Working Life and Social Research (2012-0073). The funders had no role in study design, data collection and analysis, preparation of the manuscript or decision to publish the manuscript.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.