Factors associated with insufficient awareness of breast cancer among women in Northern and Eastern China: a case–control study

Objectives To investigate the awareness and knowledge level of breast cancer among Chinese participants. Design Case–control study. Settings This study was based on the database of the minister-affiliated hospital key project of the Ministry of Health of the People’s Republic of China that included 21 Chinese hospitals between April 2012 and April 2013. Participants Matched study was designed among 2978 participants with Han ethnicity aged between 25 and 70. Primary and secondary outcome measures Student’s t-test, Pearson’s χ2 test, reliability analysis, exploratory factor analysis, and univariate and multivariate logistic regression analyses were performed to know the level of breast cancer knowledge and find the breast cancer awareness-associated factors. Results 80.0% (2383/2978) of the participants had poor awareness level of breast cancer. In-depth knowledge of breast cancer such as early symptoms and risk factors was poorly found among them. Television broadcast and relatives or friends with breast cancers were the main sources of information about breast cancer. Of all participants, 72.8% (2167/2978) had heard about breast cancer as a frequent cancer affecting women, and 63.3% (1884/2978) knew that family history of breast cancer was a risk factor for breast cancer. Over half of them were aware that a breast lump could be a symptom of breast cancer. Multivariate analysis identified the following variables that predicted awareness of breast cancer: young age (OR=0.843, 95% CI 0.740 to 0.961), occupation (agricultural worker) (OR=12.831, 95% CI 6.998 to 23.523), high household social status (OR=0.644, 95% CI 0.531 to 0.780), breast hyperplasia history (OR=1.684, 95% CI 1.273 to 2.228), high behavioural prevention score (OR=4.407, 95% CI 3.433 to 5.657). Conclusion Most women were aware of breast cancer as a disease, but their in-depth knowledge of it was poor. More publicity and education programmes to increase breast cancer awareness are necessary and urgent, especially for the ageing women and agricultural workers.

Conclusion: Most women were aware of breast cancer as a disease, but their in depth knowledge of breast cancer was still poor. Increased publicity and education programs to increase breast cancer awareness are necessary and urgent, especially targeting aging women and agricultural workers.
Keywords: Breast cancer, Awareness, Knowledge, Chinese women

Strengths and limitations of this study
This was a multicenter, matched case control study designed to investigate the breast cancer awareness of women in 21 hospitals in Northern and Eastern China.
We designed this hospital based study to investigate the level of knowledge of breast cancer in female breast cancer patients. Meanwhile, we compared the results with our previous community based study.
Although we used the same questionnaire in these two studies, we still did not have uniform standards and methods for measuring knowledge evaluation. Thence a standard measurement of breast cancer related knowledge should be developed. Breast cancer (BC) is one of the most common cancers and the leading cause of cancer related death among women worldwide [1] . Although in the past, China had a relatively low incidence of BC, recently BC incidence has been increasing much faster in China than globally [2] . According to latest statistics from the National Central Cancer Registry, breast cancer is the most frequently diagnosed cancer among Chinese women in all age and ethnic groups, accounting for nearly one fifth of all cancer types. Moreover, by 2011, BC incidence had increased to 32.43/100,000, which is higher than the average BC incidence in East Asia (27/100,000) [3] .
Many studies have shown that the early detection of BC plays a vital role in patient survival.
Further, a delay in the diagnosis and subsequent treatment can lead to worsening of morbidity and mortality [4,5] . From the time of onset to that of diagnosis, patients may experience disease progression, which could lead to tumor growth, and consequently, worse outcomes [6] .
Stages at diagnosis differ among countries with different incomes. A report showed that more than 70% of women with BC in developed countries had disease stage I or II, compared with 20% to 60% in low and middle income countries [7] . A commercial report showed that in China, nearly two thirds of BC patients were diagnosed with advanced disease, which was obviously higher than that in the USA [2] . It is well known that BC screening is an effective way to detect early stage disease; however, in contrast to developed countries, China has no such current nationwide BC screening program because of economic and demographic factors [8] . Additionally, people are not considered to consciously have the appropriate attitudes toward such BC screening programs [9] . Many studies have shown that the level of cancer awareness is a significant risk factor for early detection of BC [9 11] . Therefore, it is necessary to implement interventions aimed at increasing the comprehensive knowledge and awareness of BC symptoms and screening methods.
To promote BC awareness among Chinese women and build education programs to prevent delays in diagnosis and treatment, healthcare specialists must know their current level of understanding. Thus, we performed this cross sectional survey in China to assess the level of awareness and knowledge of breast cancer related symptoms and risk factors, and identify awareness related factors.

Study design
We designed a multicenter, case control hospital based study to investigate the awareness of women in 21 hospitals located in 11 provinces in Northern and Eastern China. This study was conducted between April 2012 and April 2013 and was funded by the Ministry of Health of the People's Republic of China.

Study population
All participants were of the Han ethnic group. Cases and controls were matched 1:1 for age (±3 years) and timing of examination (within 2 months). The inclusion criteria for BC group were as follows: (1) newly diagnosed and histologically confirmed BC and (2)  criteria for the control group were as follows: a neoplastic disease at any other site, history of cancer or other major chronic diseases. We collected data strictly according to the inclusion and exclusion criteria, and then excluded subjects whose data were incomplete or lacking.
Finally, a total of 1489 case control sets were involved.

Data collection
A self designed structured questionnaire was previously developed to record information through person to person interviews. The theoretical bases of this interview questionnaire were numerous published articles and the opinions of a variety of experts in breast surgery, epidemiology, statistics, nutrition, and molecular biology. Several similar questions were asked in different sections of the questionnaire to minimize recall bias. Previously, we conducted an investigation to assess the practicality and effectiveness of the survey and by which we validated the questionnaire [8] . The final interviewer administered questionnaire was composed of the following parts: (1) demographic characteristics, physiological and reproductive factors, such as current age, age at menarche, age at menopause, menopausal status; (2)   immunohistochemical diagnosis were also evaluated.

Scoring scheme
Awareness and knowledge of BC were assessed through 15 items on risk factors and early symptoms of BC included in the questionnaire (Table 2). For each item, if respondents gave a correct response ("yes"), they scored one (1) point; if a wrong response ("no" or "do not know") was given, the score for the item was zero (0). Total scores thus ranged from 0 to 15.
Then, we set a score to identify the status of respondents' awareness and knowledge of BC.
Respondents with scores ranging from 0 to 8 were considered to have poor awareness and knowledge, whereas those with scores ranging from 9 to 15 points were considered to have high awareness and knowledge. Behavioral prevention was scored cumulatively by five items: participation in BC screening, breast self examination (BSE), clinical breast examination (CBE), radiographic breast examination, and breast ultrasound examination. Its scoring rules were the same as the 15 item questionnaire, and the total scores ranged from 0 to 5. The overall life satisfaction score was cumulatively based on 12 items; high scores meant low life satisfaction, and low scores indicated high life satisfaction.

Quality control
Interviewers were selected by medical professionals and medical post graduate students. All interviewers had completed standardized training and were certified to conduct surveys independently. The questionnaires and forms were coded twice and were entered twice by different clerks. If there were inconsistent records, professionals would manually check and correct these. We also used computer software to check the logic and reasonable range of responses throughout the questionnaire to identify contradictory responses.

Statistical analysis
EpiData3.1 was used to create the database. Statistical methods, including Student's t test, Pearson's χ2 test, reliability analyses, exploratory factor analysis, and univariate and multivariate logistic regression analyses, were used to identify factors related to the knowledge of BC. Odds ratios (OR) with 95% confidence intervals were also calculated. All data analyses were performed using SPSS21.0.

RESULTS
In this survey, 2978 women were all included in our final analysis. The demographic characteristics are shown in   Regarding the awareness of early symptoms of BC, more than half (52.7%) of the subjects were aware that a breast lump could be a symptom of BC. Awareness of other symptoms was higher when compared with our previous community based survey results ( Figure 1). In summary, 595 women (20.0%) showed high BC awareness and 2383 (80.0%) poor BC awareness.  (Table 3). History of breast hyperplasia, reference: absence of breast hyperplasia history.
Behavioral prevention score was a cumulative score of 5 items; reference: low scores.
RMB, Chinese Yuan Renminbi; BMI, body mass index; WHR, waist hip ratio Reliability and construct validity and internal consistency reliability estimates of the 15 item scale of awareness and knowledge of BC were calculated using Cronbach's alpha. The α coefficient for the total scale was 0.902, which was considered acceptable for internal consistency reliability [12] . Exploratory factor analysis was conducted to explore construct validity. The Kaiser-Meyer-Olkin measure produced a coefficient of 0.883, indicative of excellent sampling adequacy. Bartlett's test of sphericity produced a value of 23825.328 ( < 0.001), indicating that the correlation matrix was unlikely to be an identity matrix and was therefore suitable for factor analysis [13] .

DISCUSSION
In this hospital based study, we evaluated the level of BC awareness among Chinese women.
Results showed that most participants had poor awareness regarding BC (80.0%), which is similar to our previous community based study (81.4% of subjects had poor awareness). This seems to be a common phenomenon both in developing and developed countries, although the proportion of women with high awareness in developed countries was reportedly higher than that in developing countries or regions [14,15] .
It has also been reported that different approaches for obtaining cancer related knowledge influenced the level of awareness of BC [16] . We investigated the resources by which women obtained knowledge of BC, and we found that the majority of women obtained BC  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  Hispanic college women revealed that the Internet was the most common information resource (75%) [17] , while a Spanish survey indicated that the main sources of information were television, press, family, and friends [18] . In Malaysia and Cameroon, television was still the major resource for obtaining knowledge [19,20] . It was worth noting that an increasing number of people are using the Internet to obtain information on diseases such as cancer.
However, in our country, women were not accustomed to employing the Internet to search for cancer related information, which may be a possible consequence of cultural and economic diversity.
In our study, more than two thirds (72.8%) of the women knew or had heard about BC, but their in depth knowledge of the early symptoms of BC and risk factors was insufficient.
Although more than half (52.7%) of the women knew that the presence of a lump in the breast was a BC symptom and family history was an important risk factor for BC, the proportion of women who knew other BC symptoms and BC risk factors was low (overall, less than 40%).
These results were consistent with our previous study [8] (Figure 1), which indicated that although most women in China knew or had heard about BC, their in depth knowledge of BC needed to be urgently improved. Studies from other countries showed consistent results [4,14,21,22] , especially those conducted in developing countries. In developed countries, the proportion  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   19 of women reporting cancer related symptoms was higher to a certain extent; this may be a possible consequence of their higher living standards, a greater consciousness of health, and more social publicity.
Many studies found a close relationship between age and awareness and knowledge of BC [9,21,23,24] . The results of a study by Mandelblatt J et al. revealed that the level of knowledge of BC decreased as the age of responders increased [25] . Several articles researching Indian women also yielded similar results [9,26] . However, a study completed by Sen et al. showed that older women were more interested in BC knowledge than younger women. Okobia et al.
also arrived at the same conclusion among Nigerian women [27] . In our study, we found that age was related to BC awareness as well (OR=0.843, 95% CI: 0.740-0.961). Younger women tended to have more awareness and knowledge of BC although older women were at higher risk of BC development. Thus, it is urgent to improve the awareness and knowledge of older women to decrease the incidence of malignant breast tumors in this population.
The relationship between occupation and the level of awareness and knowledge on BC were also demonstrated in this study. Workers and medical staff tended to be more aware of the symptoms and risk factors for BC, while the awareness of agricultural workers was significantly poorer. Similar results were observed in many other studies including our community based population survey [8,28,29] . Nonetheless, an Iranian study showed that the awareness of BC of rural women was moderate; this may indicate the existence of ethnic and population differences [10] . As is shown in both community and hospital based studies, medical personnel were more aware of BC than agricultural workers (OR: 4.774, 95% CI: 4.316 5.281), likely benefiting from easier access to relevant knowledge. In this study, the result of the reliability index was 0.902 and the validity was 0.883, while the results in our previous study were 0.910 and 0.870, respectively. We used the same questionnaire in these two studies, but we still did not have uniform standards and methods for measuring breast cancer related knowledge evaluation. Thus, we consider that a standard measurement of breast cancer related knowledge should be developed. Additionally, the reality and validity of our 5 items questionnaire for assessing levels of behavioral prevention was not good enough for this assessment (Cronbach's alpha 0.769 and KMO coefficient 0.780), which limited the results of the study in terms of BC practices.

CONCLUSIONS
Combined with our previous study, we concluded that most women were aware of BC as a disease entity, but their in depth knowledge of the disease was poor. BC awareness, increased publicity, and education programs are necessary and urgent, especially for older women and agricultural workers.      Conclusion: Most women were aware of breast cancer as a disease, but their in depth knowledge of breast cancer was still poor. Increased publicity and education programs to increase breast cancer awareness are necessary and urgent, especially targeting aging women and agricultural workers.

Eastern China
Keywords: Breast cancer, Awareness, Knowledge, Chinese women

Strengths and limitations of this study
This was a multicenter, matched case control study designed to investigate the breast cancer awareness of women in 21 hospitals in Northern and Eastern China.
We designed this hospital based study to investigate the level of knowledge of breast cancer in female breast cancer patients. Meanwhile, we compared the results with our previous community based study.
Although we used the same questionnaire in these two studies, we still did not have uniform standards and methods for measuring knowledge evaluation. Thence a standard measurement of breast cancer related knowledge should be developed.  [1] . Although in the past, China had a relatively low incidence of BC, recently BC incidence has been increasing much faster in China than globally on account of change of diet, lifestyle and unique one child policy [2,3] . According to latest statistics from the National Central Cancer Registry, breast cancer is the most frequently diagnosed cancer among Chinese women in all age and ethnic groups, accounting for nearly one fifth of all cancer types. Moreover, by 2011, BC incidence had increased to 32.43/100,000, which is higher than the average BC incidence in East Asia (27/100,000) [4] .
Many studies have shown that the early detection of BC plays a vital role in patient survival.
Further, a delay in the diagnosis and subsequent treatment can lead to worsening of morbidity and mortality [5,6] . From the time of onset to that of diagnosis, patients may experience disease progression, which could lead to tumor growth, and consequently, worse outcomes [7] .
Stages at diagnosis differ among countries with different incomes. A report showed that more than 70% of women with BC in developed countries had disease stage I or II, compared with 20% to 60% in low and middle income countries [8] . A commercial report showed that in China, nearly two thirds of BC patients were diagnosed with advanced disease, which was obviously higher than that in the USA [2] . It is well known that BC screening is an effective way to detect early stage disease; however, in contrast to developed countries, China has no such current nationwide BC screening program because of economic and demographic factors [9] . Additionally, people are not considered to consciously have the appropriate attitudes toward such BC screening programs [10] . Many studies have shown that the level of cancer awareness is a significant risk factor for early detection of BC [10,11,12] . Therefore, it is necessary to implement interventions aimed at increasing the comprehensive knowledge and awareness of BC symptoms and screening methods.
To promote BC awareness among Chinese women and build education programs to prevent delays in diagnosis and treatment, healthcare specialists must know their current level of understanding. Thus, we performed this cross sectional survey in China to assess the level of awareness and knowledge of breast cancer related symptoms and risk factors, and identify awareness related factors.

Study design
We designed a multicenter, case control hospital based study to investigate the awareness of

Study population
All participants were of the Han ethnic group. Cases and controls were matched 1:1 for age (±3 years) , diagnosis hospital (same hospital) and timing of examination (within 2 months).
The inclusion criteria for BC group were as follows: (1)   for the control group were as follows: (1) negative physical examination results; (2) negative ultrasound scans of breast and/or mammographic screening results; and (3) no evidence or history of cancer. The exclusion criteria for the control group were as follows: a neoplastic disease at any other site, history of cancer or other major chronic diseases. We collected data strictly according to the inclusion and exclusion criteria, and then excluded subjects whose data were incomplete or lacking. Finally, a total of 1489 case control sets were involved.

Data collection
A self designed structured questionnaire was previously developed to record information through person to person interviews. The theoretical basis of this interview questionnaire was numerous published articles and the opinions of a variety of experts in breast surgery, epidemiology, statistics, nutrition, and molecular biology. Several similar questions were asked in different sections of the questionnaire to minimize recall bias. Previously, we conducted an investigation to assess the practicality and effectiveness of the survey and by which we validated the questionnaire [9] . The final interviewer administered questionnaire was composed of the following parts: (1)

Scoring scheme
Awareness and knowledge of BC were assessed through 15 items on risk factors and early symptoms of BC included in the questionnaire (Table 2). For each item, if respondents gave a correct response ("yes"), they scored one (1) point; if a wrong response ("no" or "do not know") was given, the score for the item was zero (0). Total scores thus ranged from 0 to 15.
Then, we set a score to identify the status of respondents' awareness and knowledge of BC.
Respondents with scores ranging from 0 to 8 were considered to have poor awareness and knowledge, whereas those with scores ranging from 9 to 15 points were considered to have high awareness and knowledge. Behavioral prevention was scored cumulatively by five items: participation in BC screening, breast self examination (BSE), clinical breast examination (CBE), radiographic breast examination, and breast ultrasound examination. Its scoring rules were the same as the 15 item questionnaire, and the total scores ranged from 0 to 5. The overall life satisfaction score was cumulatively based on 12 items; high scores meant low life satisfaction, and low scores indicated high life satisfaction.

Quality control
Interviewers were selected by medical professionals and medical post graduate students. All interviewers had completed standardized training and were certified to conduct surveys independently. The questionnaires and forms were coded twice and were entered twice by different clerks. If there were inconsistent records, professionals would manually check and correct these. We also used computer software to check the logic and reasonable range of responses throughout the questionnaire to identify contradictory responses.

Statistical analysis
EpiData3.1 was used to create the database. Statistical methods, including Student's t test, Pearson's χ2 test, reliability analyses, exploratory factor analysis, and univariate and multivariate logistic regression analyses, were used to identify factors related to the knowledge of BC. Odds ratios (OR) with 95% confidence intervals were also calculated. All data analyses were performed using SPSS21.0. subjects had a monthly family income greater than 5000 RMB. Regarding the awareness of early symptoms of BC, more than half (52.7%) of the subjects were aware that a breast lump could be a symptom of BC. Awareness of other symptoms was higher when compared with our previous community based survey results (Figure 1). In summary, 595 women (20.0%) showed high BC awareness and 2383 (80.0%) poor BC awareness.  3.433-5.657) were independently correlated with BC awareness and knowledge (Table 3) through stepwise method. Occupation: "Other" includes occupations such as teacher, civil servant, individual business, driver, service, company employee, and housewife.

In
Household social status, reference: low social status.
History of breast hyperplasia, reference: absence of breast hyperplasia history.
Behavioral prevention score was a cumulative score of 5 items; reference: low scores.
RMB, Chinese Yuan Renminbi; BMI, body mass index; WHR, waist hip ratio Reliability and construct validity and internal consistency reliability estimates of the 15 item scale of awareness and knowledge of BC were calculated using Cronbach's alpha. The α coefficient for the total scale was 0.902, which was considered acceptable for internal consistency reliability [13] . Exploratory factor analysis was conducted to explore construct validity. The Kaiser-Meyer-Olkin measure produced a coefficient of 0.883, indicative of excellent sampling adequacy. Bartlett's test of sphericity produced a value of 23825.328 ( < 0.001), indicating that the correlation matrix was unlikely to be an identity matrix and was therefore suitable for factor analysis [14] .

DISCUSSION
In this hospital based study, we evaluated the level of BC awareness among Chinese women.
Results showed that most participants had poor awareness regarding BC (80.0%), which is similar to our previous community based study (81.4% of subjects had poor awareness). This seems to be a common phenomenon both in developing and developed countries, although the proportion of women with high awareness in developed countries was reportedly higher than that in developing countries or regions [15,16] .
It has also been reported that different approaches for obtaining cancer related knowledge influenced the level of awareness of BC [17] . We investigated the resources by which women obtained knowledge of BC, and we found that the majority of women obtained BC information through traditional media such as TV broadcasts (30.6%) and their friends or relatives with BC (29.6%). The Internet has developed rapidly and become widespread, but in the present study, women failed to benefit from it in terms of BC awareness (8.6%). Less women participated in special lectures about BC (2.5%), which indicated that the effort to publicize such events for BC were likely insufficient. A study comparing non Hispanic and Hispanic college women revealed that the Internet was the most common information resource (75%) [18] , while a Spanish survey indicated that the main sources of information were television, press, family, and friends [19] . In Malaysia and Cameroon, television was still the major resource for obtaining knowledge [ 20,21] . It was worth noting that an increasing number of people are using the Internet to obtain information on diseases such as cancer.
However, in our country, women were not accustomed to employing the Internet to search for cancer related information, which may be a possible consequence of cultural and economic diversity.
In our study, more than two thirds (72.8%) of the women knew or had heard about BC, but their in depth knowledge of the early symptoms of BC and risk factors was insufficient.
Although more than half (52.7%) of the women knew that the presence of a lump in the breast was a BC symptom and family history was an important risk factor for BC, the proportion of women who knew other BC symptoms and BC risk factors was low (overall, less than 40%).
These results were consistent with our previous study [9] (Figure 1), which indicated that although most women in China knew or had heard about BC, their in depth knowledge of BC needed to be urgently improved. Studies from other countries showed consistent results [5,15,22,23] , especially those conducted in developing countries. In developed countries, the proportion of women reporting cancer related symptoms was higher to a certain extent; this may be a possible consequence of their higher living standards, a greater consciousness of health, and more social publicity.
Many studies found a close relationship between age and awareness and knowledge of BC [10,22,24,25] . The results of a study by Mandelblatt J et al. revealed that the level of knowledge of BC decreased as the age of responders increased [26] . Several articles researching Indian women also yielded similar results [10,27] . However, a study completed by Sen et al. showed that older women were more interested in BC knowledge than younger women. Okobia et al.
also arrived at the same conclusion among Nigerian women [28] . In our study, we found that age was related to BC awareness as well (OR=0.843, 95% CI: 0.740-0.961). Younger women tended to have more awareness and knowledge of BC although older women were at higher risk of BC development. We speculated that young women were more likely to focus on self health conditions and be active learners and to access available information. From the above, it is urgent to improve the awareness and knowledge of older women to decrease the incidence of malignant breast tumors in this population.
The relationship between occupation and the level of awareness and knowledge on BC were also demonstrated in this study. Workers and medical staff tended to be more aware of the symptoms and risk factors for BC, while the awareness of agricultural workers was significantly poorer. Similar results were observed in many other studies including our community based population survey [9,29,30] . Nonetheless, an Iranian study showed that the awareness of BC of rural women was moderate; this may indicate the existence of ethnic and population differences [11] . As is shown in both community and hospital based studies, and BSE remains controversial [9] , but most scholars still believe that self examination probably improves awareness and might play an important role in nationwide programs for earlier stage detection in China [37,38] . From this point of view, breast examination is helpful for early diagnosis and decreased mortality.
We compared the results with our earlier study [9] . Results showed that the proportion of women with poor awareness regarding BC was lower than that of community based sample (80.0% vs 81.4%). As to the awareness of breast cancer awareness, we also found higher percentages in hospital based women (Figure 1 advanced countries, the level of BC awareness in China was still lower, which was caused by many reasons. Financial obstacle was important one of reasons [8] . In China, average household income level could not catch up with that in advanced countries in a short time.
Although there are no major differences socially, culturally, and economically, the heightened awareness might be due to better infrastructure, advanced technology, and educational facilities available in the advanced regions [39,40] . Developing countries provided women with limited access to professional knowledge and affordable quality healthcare treatment.
Meantime, breast cancer screening was neither cost effective nor feasible [41] .
In this study, the result of the reliability index was 0.902 and the validity was 0.883, while the results in our previous study were 0.910 and 0.870, respectively. We used the same questionnaire in these two studies, but we still did not have uniform standards and methods for measuring breast cancer related knowledge evaluation. Thus, we consider that a standard measurement of breast cancer related knowledge should be developed, which will be a part of important content of our further study. Additionally, the reality and validity of our 5 items questionnaire for assessing levels of behavioral prevention was not good enough for this assessment (Cronbach's alpha 0.769 and KMO coefficient 0.780), which limited the results of the study in terms of BC practices.

CONCLUSIONS
Combined with our previous study, we concluded that most women were aware of BC as a disease entity, but their in depth knowledge of the disease was poor. BC awareness, increased publicity, and education programs are necessary and urgent, especially for older women and agricultural workers.

Competing interests
The authors declare that they have no competing interests.

Conclusion:
Most women were aware of breast cancer as a disease, but their in depth knowledge of it was poor. More publicity and education programs to increase breast cancer awareness are necessary and urgent, especially for the aging women and agricultural workers.
Keywords: Breast cancer, Awareness, Knowledge, Chinese women

Strengths and limitations of this study
This was a multicenter, matched case control study designed to investigate the breast cancer awareness of women in 21 hospitals in Northern and Eastern China.
We designed this hospital based study to investigate the level of knowledge of breast cancer in female breast cancer patients. Meanwhile, we compared the results with our previous community based study.

INTRODUCTION
Breast cancer (BC) is one of the most common cancers and the leading cause of cancer related death among women worldwide [1] . Despite that the incidence of BC was low in China, it has been increasing much faster than globally recently on account of change of diet, lifestyle and unique one child policy [2,3] . According to latest statistics from the National Central Cancer Registry, breast cancer is the most frequently diagnosed cancer among Chinese women in all age and ethnic groups, accounting for nearly one fifth of all cancer types. Moreover, by 2011, BC incidence had increased to 32.43/100,000, which is higher than the average BC incidence in East Asia (27/100,000) [4] .
Many studies have shown that the early detection of BC plays a vital role in patient survival.
Further, a delay in the diagnosis and subsequent treatment can lead to worsening of morbidity and mortality [5,6] . From the time of onset to that of diagnosis, patients may experience disease progression, which could lead to tumor growth, and consequently, worse outcomes [7] .
Stages at diagnosis differ among countries with different incomes. Salih AM et al. reported that more than 70% of women with BC in developed countries had disease stage I or II, compared with 20% to 60% in low and middle income countries [8] . A commercial report showed that nearly two thirds of BC patients were diagnosed with advanced disease in China, which was obviously higher than that in the USA [2] . It is well known that BC screening is an effective way to detect early stage disease. However, the current nationwide BC screening program is not available in China because of economic and demographic factors [9] .
To promote BC awareness among Chinese women and build education programs to prevent delays in diagnosis and treatment, healthcare specialists must know their current level of understanding. Thus, we performed this cross sectional survey in China to assess the level of awareness and knowledge of breast cancer related symptoms and risk factors, and identify awareness related factors.

Study design
We designed a multicenter, case control hospital based study to investigate the awareness of

Study population
All participants were of the Han ethnic group. Cases and controls were matched 1:1 for age (±3 years), diagnosis hospital (same hospital) and timing of examination (within 2 months).
The inclusion criteria for BC group were as follows: (1) 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   6 for the control group were as follows: (1) negative physical examination results; (2) negative ultrasound scans of breast and/or mammographic screening results; and (3) no evidence or history of cancer. The exclusion criteria for the control group were as follows: a neoplastic disease at any other site, history of cancer or other major chronic diseases. We collected data strictly according to the inclusion and exclusion criteria, and then excluded subjects whose data were incomplete or lacking. Finally, a total of 1489 case control sets were involved.

Data collection
A self designed structured questionnaire was previously developed to record information through person to person interviews. The theoretical basis of this interview questionnaire was numerous published articles and the opinions of a variety of experts in breast surgery, epidemiology, statistics, nutrition, and molecular biology. Several similar questions were asked in different sections of the questionnaire to minimize recall bias. Previously, we conducted an investigation to assess the practicality and effectiveness of the survey and by which we validated the questionnaire [9] . The final interviewer administered questionnaire was composed of the following parts: (1) demographic characteristics, physiological and reproductive factors, such as current age, age at menarche, age at menopause, menopausal  diagnostic tests results. Additionally, for BC patients, the histological and immunohistochemical diagnosis were also evaluated.

Scoring scheme
Awareness and knowledge of BC were assessed through 15 items on risk factors and early symptoms of BC included in the questionnaire (Table 1). For each item, if respondents gave a correct response ("yes"), they scored one (1) point; if a wrong response ("no" or "do not know") was given, the score for the item was zero (0). Total scores thus ranged from 0 to 15.
Then, we set a score to identify the status of respondents' awareness and knowledge of BC.
Respondents with scores ranging from 0 to 8 were considered to have poor awareness and knowledge, whereas those with scores ranging from 9 to 15 points were considered to have high awareness and knowledge. Behavioral prevention was scored cumulatively by five items: participation in BC screening, breast self examination (BSE), clinical breast examination (CBE), radiographic breast examination, and breast ultrasound examination. Its scoring rules were the same as the 15 item questionnaire, and the total scores ranged from 0 to 5. The overall life satisfaction score was cumulatively based on 12 items; high scores meant low life satisfaction, and low scores indicated high life satisfaction.

Quality control
Interviewers were selected by medical professionals and medical post graduate students. All interviewers had completed standardized training and were certified to conduct surveys independently. The questionnaires and forms were coded twice and were entered twice by different clerks. If there were inconsistent records, professionals would manually check and correct these. We also used computer software to check the logic and reasonable range of responses throughout the questionnaire to identify contradictory responses.

Statistical analysis
EpiData3.1 was used to create the database. Statistical methods, including Student's t test, Pearson's χ2 test, reliability analyses, exploratory factor analysis, and univariate and multivariate logistic regression analyses, were used to identify factors related to the knowledge of BC. Odds ratios (OR) with 95% confidence intervals were also calculated. All data analyses were performed using SPSS21.0.

RESULTS
In this survey, 2978 women were included in our final analysis. The demographic characteristics are shown in subjects had a monthly family income more than 5000 RMB. Regarding the awareness of early symptoms of BC, more than half (52.7%) of the subjects were aware that a breast lump could be a symptom of BC. Awareness of other symptoms was higher when compared with our previous community based survey results ( Figure 1). In summary, 595 women (20.0%) showed high BC awareness and 2383 (80.0%) poor BC awareness. 3.433-5.657) were independently correlated with BC awareness and knowledge (Table 3) through stepwise method.  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 [13] . Exploratory factor analysis was conducted to explore construct validity. The Kaiser-Meyer-Olkin measure produced a coefficient of 0.883, indicative of excellent sampling adequacy. Bartlett's test of sphericity produced a value of 23825.328 ( < 0.001), indicating that the correlation matrix was unlikely to be an identity matrix and was therefore suitable for factor analysis [14] .

DISCUSSION
In this hospital based study, we evaluated the level of BC awareness among Chinese women.
Results showed that most participants had poor awareness regarding BC (80.0%), which is similar to our previous community based study (81.4% of subjects had poor awareness). This seems to be a common phenomenon both in developing and developed countries, although the proportion of women with high awareness in developed countries was reportedly higher than that in developing countries or regions [15,16] .
It has also been published that different approaches for obtaining cancer related knowledge influenced the level of awareness of BC [17] . We investigated the resources by which women  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  Hispanic college women revealed that the Internet was the most common information resource (75%) [18] , while a Spanish survey indicated that the main sources of information were television, press, family, and friends [19] . In Malaysia and Cameroon, television was still the major resource for obtaining knowledge [ 20,21] . It is worth noting that an increasing number of people are using the Internet to obtain information on diseases such as cancer.
However, in our country, people were not accustomed to employing the Internet to search for cancer related information, which may be a possible consequence of cultural and economic diversity.
In our study, more than two thirds (72.8%) of the participants knew or had heard about BC, but their in depth knowledge of the early symptoms of BC and risk factors was insufficient.
Although more than half (52.7%) of them knew that the presence of a lump in the breast was a BC symptom and family history was an important risk factor for BC, the proportion of women who knew other BC symptoms and BC risk factors was low (overall, less than 40%).
These results were consistent with our previous study [9] (Figure 1), which indicated that although most women in China knew or had heard about BC, their in depth knowledge of BC needed to be urgently improved. Studies from other countries showed consistent results [5, 15, 22, 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 19 23] , especially those conducted in developing countries. In developed countries, the proportion of women reporting cancer related symptoms was higher to a certain extent; this may be a possible consequence of their higher living standards, a greater consciousness of health, and more social publicity.
Many studies found a close relationship between age and awareness and knowledge of BC [10,22,24,25] . The results of a study by Mandelblatt J et al. revealed that the level of knowledge of BC decreased as the age of responders increased [26] . Several articles researching Indian women also yielded similar results [10,27] . However, a study completed by Sen et al. showed that older women were more interested in BC knowledge than younger women. Okobia et al.
also arrived at the same conclusion among Nigerian women [28] . In our study, we found that age was related to BC awareness as well (OR=0.843, 95% CI: 0.740-0.961). Younger women tended to have more awareness and knowledge of BC although older women were at higher risk of BC development. We speculated that young women were more likely to focus on self health conditions and be active learners and to access available information. From the above, it is urgent to improve the awareness and knowledge of older women to decrease the incidence of malignant breast tumors in this population.
The relationship between occupation and the level of awareness and knowledge on BC were also demonstrated in this study. Workers and medical staff tended to be more aware of the symptoms and risk factors for BC, while the awareness of agricultural workers was significantly poorer. Similar results were observed in many other studies including our community based population survey [9,29,30] . Nonetheless, an Iranian study showed that the awareness of BC of rural women was moderate; this may indicate the existence of ethnic 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   20 population differences [11] . As is shown in both community and hospital based studies, medical personnel were more aware of BC than agricultural workers (OR: 4.774, 95% CI: 4.316 5.281), likely benefiting from easier access to relevant knowledge.
We compared the results with our earlier study [9] . Results showed that the proportion of women with poor awareness of BC was lower than that of community based sample (80.0% vs 81.4%). As to the awareness of breast cancer awareness, we also found higher percentages in hospital based women ( Figure 1). It can be implied that medical community played an important role on it. Some studies showed the significance of medical staff on increasing level of cancer awareness. High levels of awareness about common health issues such as breast cancer was shown in Gökay Terzioğlu's study based on Turkish population, as a consequence of direct communication with professional doctors [31] . Further study on importance of medical community in China should be conducted. When compared with the advanced countries, the level of BC awareness in China was lower, which was caused by many reasons. Financial obstacle was important one of reasons [8] . In China, average household income level could not catch up with that in advanced countries in a short time. Although there are no major differences socially, culturally, and economically, the heightened awareness might be due to better infrastructure, advanced technology, and educational facilities available in the advanced regions [39,40] . Developing countries provided women with limited access to professional knowledge and affordable quality healthcare treatment.
Meantime, breast cancer screening was neither cost effective nor feasible [41] .
In this study, the result of the reliability index was 0.902 and the validity was 0.883, while the results in our previous study were 0.910 and 0.870, respectively. We used the same questionnaire in these two studies, but we still did not have uniform standards and methods for measuring breast cancer related knowledge evaluation. Thus, we consider that a standard measurement of breast cancer related knowledge should be developed, which will be a part of important content of our further study. Additionally, the reality and validity of our 5 items questionnaire for assessing levels of behavioral prevention was not good enough for this assessment (Cronbach's alpha 0.769 and KMO coefficient 0.780), which limited the results of the study in terms of BC practices.

Competing interests
The authors declare that they have no competing interests.