SEARCH

SEARCH BY CITATION

Keywords:

  • breast neoplasms;
  • early detection of cancer;
  • breast self-examination;
  • neoplasm metastasis;
  • psychosocial factors

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

BACKGROUND

Young women may experience delays in diagnosis of breast cancer, and these delays may contribute to poorer outcomes.

METHODS

In a prospective, multicenter cohort study, women recently diagnosed with breast cancer at age ≤40 years were surveyed regarding their initial signs or symptoms of cancer and delays in diagnosis. Self delay was defined as ≥90 days between the first sign or symptom and a patient's first visit to consult a health care provider. Care delay was defined as ≥90 days between that first visit and the diagnosis of breast cancer. In a medical record review, tumor characteristics were assessed, including disease stage. Univariate and multivariate models were used to assess for predictors of self delay, care delay, and advanced stage in the self-detected subset.

RESULTS

In 585 eligible participants, the first sign or symptom of cancer was a self-detected breast abnormality for 80%, a clinical breast examination abnormality for 6%, an imaging abnormality for 12%, and a systemic symptom for 1%. Among women with self-detected cancers, 17% reported a self delay, and 12% reported a care delay. Self delays were associated with poorer financial status (P = 0.01). Among young women with self-detected breast cancers, care delay was associated at trend level (P = .06) with higher stage in multivariate modeling.

CONCLUSIONS

Most young women detect their own breast cancers, and most do not experience long delays before diagnosis. Women with fewer financial resources are more likely to delay seeking medical attention for a self-detected breast abnormality. Cancer 2014;120:20–25. © 2013 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

Breast cancer is diagnosed in fewer than 1 in 200 women aged ≤40 years in the United States, yet it is the leading cause of cancer-related deaths in this age group.[1] Breast cancers in young women are more likely to be large and lymph node-positive at diagnosis, and younger women have an increased risk of recurrence and death from breast cancer compared with older women.[2, 3] These variations may be explained by biologic differences, although delays in diagnosis also may contribute to presentation with more advanced stage disease and, thus, have an impact on prognosis. Data are limited regarding diagnostic delays and their possible contributions to outcome.[4, 5] In an analysis of patient-reported data from a national sample of patients with cancer in the United Kingdom, younger and unmarried women with breast cancer experienced longer delays than older women.[4] A recent evaluation within the National Comprehensive Cancer Network Breast Cancer Outcomes Database Project revealed that women diagnosed with breast cancer at age ≤40 years were more likely to have diagnostic delays. Further analysis suggested that symptomatic presentation (ie, with a lump instead of a mammographic abnormality), which was more common in young women, was strongly associated with delays.[5]

Data on the impact of delays in diagnosis on breast cancer outcomes are mixed. In a meta-analysis by Richards and colleagues of >25,000 women, those who experienced a delay of at least 3 months between the onset of symptoms and treatment for breast cancer were 12% less likely to be alive 5 years later.[6] In a more recent study, Love and colleagues did not observe shorter survival in Asian women who experienced a delay of greater than 6 months between finding a lump and being diagnosed with breast cancer, but they did report that women who experienced the longer delays had larger and more lymph node-positive tumors than those who did not.[7] In a large, prospective cohort study, we sought to evaluate how young women present with breast cancer, the frequency of delay in diagnosis, the factors associated with delays, and the relationship between delays and disease stage in patient-detected cancers. This work is based on 2 related models: 1) a psychosocial model that assumes that certain psychosocial and clinical factors could predispose some young women to delay seeking medical care for signs or symptoms of breast cancer or to experience a delay between their initial visit to a health care provider and receipt of a breast cancer diagnosis, and 2) a biologic model that assumes that delays could allow more time for cancers to grow and spread before diagnosis, potentially adversely affecting prognosis.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

We used patient-reported data from a multicenter, prospective cohort of women who were diagnosed with breast cancer at age ≤40 years. We studied the methods of breast cancer detection and the delays in seeking medical attention and subsequent diagnosis. Women were enrolled from 11 sites in Massachusetts and from 1 site in Denver, Colorado, into a prospective cohort study: The Young Women's Breast Cancer Study (available at: http:// clinicaltrials.gov/ct2/show/NCT01468246, accessed July 23, 2013). Eligibility requirements included age ≤40 years and a diagnosis of stage I through IV breast cancer less than 6 months before enrollment. Surveys were mailed to all enrollees, and only those who responded within 9 months of diagnosis were included. These methods were detailed in a previous publication.[8] We evaluated initial signs or symptoms of cancer. The initial sign or symptom of cancer was categorized as either “self-detected” (ie, the patient or a family member/friend detected a lump, nipple change, or other breast abnormality), “exam-detected” (ie, a health care provider detected an abnormality), “imaging-detected” (ie, a mammogram, breast magnetic resonance image, or other imaging test indicated an abnormality), or “detected based on systemic symptoms” (eg, weight loss or fatigue).

We focused our analysis on patients who had self-detected tumors to minimize bias in the results and because opportunities for future delay-reducing interventions may be greatest in this group. Among those who had self-detected cancers, we separately analyzed delays of at least 90 days either between the first sign or symptom and seeking medical attention (defined as “self delay”) or between seeking medical attention and receiving a diagnosis of breast cancer (defined as “care delay”). We used a 90-day definition for delays based on published research supporting a possible clinical relevance of this time period.[6, 7] To assess our psychosocial model, univariate models were used to explore differences between women with and without delays of at least 90 days in either timeframe with regard to the following variables: age at diagnosis, college education, employment (fulltime vs not), married or living as married (vs not), finances (more vs less financially comfortable according to the measure described below), first-degree relative with breast or ovarian cancer (vs none), previous pregnancy (vs never pregnant), race (white vs other), and pregnancy at diagnosis (vs no pregnancy at diagnosis).

Financial status was determined with the following measure: How would you describe your household's financial situation right now? (please circle only 1):

  1. After paying the bills, you still have enough money for special things that you want.
  2. You have enough money to pay the bills, but little spare money to buy extra or special things.
  3. You have money to pay the bills, but only because you cut back on things.
  4. You are having difficulty paying the bills, no matter what you do.

Responses were dichotomized for analysis; participants were characterized as more financially comfortable if they chose the first answer and as less financially comfortable if they chose the second, third, or fourth answer. Those factors that were significant at the P < .20 level were tested in multivariate models. To assess our biologic model, univariate and multivariate logistic regression models also were used to evaluate associations between stage IV disease (vs stage I-III disease) and the following variables: self delay, care delay, age at diagnosis, college education, employment (fulltime vs not), married or living as married (vs not), finances (“money for special things” vs not), first-degree relative with breast or ovarian cancer (vs none), previous pregnancy (vs never pregnant), race (white vs other), pregnancy at diagnosis (vs no pregnancy at diagnosis), grade (1/2 v 3), estrogen receptor status, progesterone receptor status, and human epidermal growth factor receptor [Her2] status. In a secondary analysis, we excluded women for whom self delay or care delay was ≥365 days (n = 17) to evaluate the impact of the outliers on the analysis of predictors of delays in women with self-detected cancers.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

Between November 1, 2006 and October 31, 2012, 1408 women were invited to participate in a cohort study, and 843 women (60%) enrolled. Seven hundred twenty women had completed at least part of the baseline survey at the time of this analysis. Forty-three women were excluded from the current analysis because they had stage 0 disease, 49 were excluded because their stage information was not available at the time of analysis, and 43 were excluded because they did not answer both questions regarding delays in diagnosis. There were 585 remaining eligible participants. Four hundred seventy women (80%) reported that the first evidence of their breast cancer was a self-detected breast abnormality. Fewer women reported that the first evidence of their cancer appeared during a clinical breast examination (6%), on a mammogram or magnetic resonance image (12%), or as a systemic symptom (1%) (see Table 1). Table 2 presents patient and tumor characteristics for the entire cohort and for those who had self-detected cancers. The median age at diagnosis was 37 years (range, 17-40 years). In those who self-detected their cancers, the median time between the initial sign and seeking medical attention was 14 days (25th to 75th percentile, 3-45 days). Seventeen percent of women with self-detected cancers recalled self delay between the initial sign and medical attention. The median time between seeking medical attention and diagnosis was 16.5 days (25th to 75th percentile, 7-30 days). Twelve percent of women with self-detected cancers recalled care delay. Three percent of women with self-detected cancers recalled self delay and care delay (for data regarding the timing of cancer presentation and diagnosis, see Table 3).

Table 1. Presentation of Breast Cancer in Young Women
Primary Method of Cancer DetectionNo. of Patients (%)
  1. Abbreviations: MRI, magnetic resonance imaging.

Self-detected470 (80)
Exam-detected33 (6)
Imaging-detected, eg, by mammogram or MRI68 (12)
Detected based on systemic symptoms7 (1)
Missing7 (1)
Table 2. Patient and Tumor Characteristics
 No. of Patients (%)
CharacteristicTotal, N = 585Self-Detected, N = 470
  1. Abbreviations: ER, estrogen receptor; Her2, human epidermal growth factor receptor 2; PR, progesterone receptor.

Patient characteristics  
Age: Median [range], y37 [17–40]36 [17–40]
White518 (89)416 (89)
College educated484 (83)389 (83)
Pregnant at diagnosis25 (4)24 (5)
Never pregnant before diagnosis164 (28)132 (28)
Employed fulltime253 (43)205 (44)
Married or living as married441 (75)348 (74)
Financially comfortable286 (49)225 (48)
Medically insured584 (99)469 (99)
First-degree relative with breast or ovarian cancer78 (13)56 (12)
Tumor characteristics  
Disease stage  
I210 (36)147 (31)
II258 (44)226 (48)
III80 (14)73 (16)
IV37 (6)24 (5)
Tumor grade  
137 (6)21 (4)
2187 (32)142 (30)
3356 (61)303 (64)
Missing5 (1)4 (1)
ER positive399 (68)302 (64)
ER missing1 (0.2)1 (0.2)
PR positive359 (61)265 (56)
PR missing1 (0.2)1 (0.2)
Her2 positive185 (32)149 (32)
Her2 missing5 (1)3 (0.6)
Table 3. Delays Among Women With Self-Detected Cancers
DelayTotal No. of Patients RespondingMedian Delay [25th to 75th percentile], d
Delay from noticing symptom to seeking care from a health care provider45514 [3–45]
Delay from seeing health care provider to receiving diagnosis46416.5 [7–30]
DelayTotal No. of PatientsNo. With Delay (%)
Delay ≥90 d to medical attention: Self delay45577 (17)
Delay ≥90 d to diagnosis: Care delay46457 (12)
Self delay ≥90 d and care delay ≥90 d44912 (3)
Delay ≥180 d to medical attention45530 (7)
Delay ≥180 d to diagnosis46430 (6)
Delay ≥365 d to medical attention4557 (2)
Delay ≥365 d to diagnosis46410 (2)

In univariate and multivariate modeling among patients with self-detected cancers, those who were less financially comfortable (ie, did not have “money for special things”) were more likely to experience self delay (univariate odds ratio [OR], 2.08; 95% confidence interval [CI], 1.25-3.57; P = .005; multivariate OR, 2.0; 95% CI, 1.18-3.33 P = .01). In those who had a first-degree relative with breast or ovarian cancer, there was also a trend toward less self delay (univariate OR, 0.39, 95% CI, 0.14-1.11; P = .08; multivariate OR, 0.39; 95% CI, 0.13-1.12; P = .08). College education also met criteria for inclusion in the multivariate model for self delay based on a P value ≤ .2 (univariate OR, 0.63; 95% CI, 0.35-1.16; P = .14), but this was not significant in the multivariate model (multivariate OR, 0.76; 95% CI, 0.41-1.43; P = .39). Multivariate regression revealed that patients with self-detected cancers who had a first-degree relative with breast or ovarian cancer were more likely to report care delay (OR, 2.76; 95% CI, 1.36-5.59; P = .005). Other covariates that met criteria for inclusion in the multivariate model for care delay based on a P value ≤ .2 in univariate models, but that were not significant in multivariate analysis, included age at diagnosis and race (see Table 4 for models).

Table 4. Relationships Between Delays and Characteristics Among Women With Self-Detected Cancers
 No. of PatientsPNo. of PatientsP
CharacteristicSelf Delay, N = 77No Self Delay, N = 378UnivariateMultivariateCare Delay, N = 57No Care Delay, N = 407UnivariateMultivariate
Age: Median (range), y36 (22–40)37 (17–40).42 35 (17–40)36 (22–40).11.13
Race  .36   .15.12
White66336  48363  
Other1036  938  
Education  .140.39  .34 
≥College59318  45340  
<College1758  1265  
Pregnant at diagnosis     .53 
Yes023  222  
No75348  55376  
Ever pregnant  1.00   .36 
Yes22108  19112  
No55270  38295  
Employment  .57   .75 
Full time32170  26176  
Less than full time45207  31230  
Marital status  .59   .58 
Married or living as married55279  40305  
Other2296  16100  
Financial status  .005.01  .75 
Financially comfortable26191  29193  
Not financially comfortable51178  28204  
Family history  .08.08  .008.005
First-degree relative with breast or ovarian cancer447  1342  
No first-degree relative with breast or ovarian cancer73331  44365  

Although univariate modeling suggested a possible association between stage IV disease and self delay, care delay, financial discomfort, and positive Her2 status, none of those variables reached statistical significance in multivariate modeling (see Table 5). However, there was a possible, statistically nonsignificant (P = .06) link between care delays and disease stage. When we restricted the analysis (data not shown) to include only women who did not have a self delay ≥365 days (N = 7) or a care delay ≥365 days (N = 10), the results were similar, indicating that data from outliers were not driving the findings.

Table 5. Relationships Between Disease Stage and Patient/Tumor Characteristics in Women With Self-Detected Cancers
 No. of PatientsP
CharacteristicStage I-III, N = 446Stage IV, N = 24UnivariateMultivariate
  1. Abbreviations: Dx, diagnosis; ER, estrogen receptor; Her2, human epidermal growth factor receptor 2; PR, progesterone receptor.

Age: Median (range), y36 (17–40)36.5 (23–40).68 
Self delay  0.110.30
Yes707  
No36117  
Care delay  .06.06
Yes516  
No38918  
Race  .76 
Caucasian39422  
Other452  
Education  .58 
≥College37019  
<College735  
Pregnant at Dx  .81 
Yes231  
No41323  
Ever pregnant  .30 
Yes1239  
No32315  
Employment  .30 
Full time1978  
Less than full time24816  
Marital status  .37 
Married33216  
Other1118  
Financial status  .08.13
Financially comfortable2187  
Other21916  
Family history   
Yes560  
No39024  
Grade  .98 
1 or 21558  
328815  
ER status  .81 
Positive28616  
Negative1598  
PR status  .28 
Positive25411  
Negative19113  
Her2 status  .14.14
Positive13811  
Negative30513  

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

Breast cancer is the leading cause of cancer-related death in young women.[9] Although prior studies have raised concern that delays in diagnosis may be responsible for poorer prognoses in younger patients, our study indicates that most young women do not experience long delays after they detect a breast abnormality themselves. Our study confirms that the vast majority of women who are diagnosed with breast cancer at age ≤40 years detect their own breast abnormalities, which is consistent with data from prior studies and with the lack of effective routine breast cancer screening for young women.[10]

We observed that some women with self-detected cancers experienced delays of at least 90 days either before they sought medical attention for a symptom (self delay) or between the time of first consultation with a health care professional and receiving the diagnosis of breast cancer (care delay). These delays may result from inadequate awareness of the disease or low suspicion for malignancy among both patients and health care professionals.

In our study population, women who were less financially comfortable were more likely to experience delays in seeking medical attention. This suggests an economic disparity that deserves further attention and supports our psychosocial model. It is possible that copays and “hidden” costs of seeking medical care (eg, parking charges, child-care expenses, lost wages) may be obstacles to timely evaluation of a new symptom for women who feel financially insecure. Although we recognize that tumor subtype may have an impact on clinical presentation,[11] we chose not to include tumor features like estrogen receptor status, progesterone receptor status, Her2 status, and tumor grade in our models for delays because we were seeking to identify characteristics that would allow targeted prediagnostic interventions to reduce delays in the future.

Counter to expectations, we observed that a family history of breast or ovarian cancer predicted a higher likelihood of care delay. Although a relevant family history would be expected to trigger more rapid evaluation and diagnosis of a self-detected breast complaint, the opposite was reported by women in this study. This finding may be spurious, or it could be the result of recall bias in a population that may have more baseline anxiety based on lived experience and family history. Another explanation is that some women who have first-degree relatives with breast or ovarian cancer may be hypervigilant, presenting so quickly for medical attention that clinicians wait for a few menstrual cycles before they perform definitive diagnostic testing. Additional research is needed to explore the potential relationship between family history and diagnostic delays.

Our finding that higher stage may be associated with care delay at trend level (P = .06) supports the hypothesis (founded on a biologic model) that diagnostic delays may contribute to higher stage and, thus, potentially may have an impact on outcomes in young patients. Delays of at least 180 or 365 days were very rare (too rare to allow statistical testing for associations with stage in this study), but it is possible that women who experience these longer delays are the most detrimentally affected. Because we did not evaluate breast cancer recurrence or survival rates in the current study, it remains unclear whether delays contribute substantially to the poorer prognoses of young women overall, which may be driven more by tumor and host biology.[12-14]

Limitations of this work include that we did not prospectively collect data regarding reasons for delays and that delay data relied solely on patient self-report. Because the vast majority of our study population was white, highly educated, married or living as married, and insured, it is possible that a lack of diversity prevented identification of other sociodemographic factors that predispose young women to delays in the general population. Our findings are also limited by multiple testing with no adjustments for nominal P values and by recall bias (ie, participants may be inaccurate in their reporting of delays). Our findings may not be generalizable because this multicenter cohort study identified participants by pathology review at participating sites and was not population-based; furthermore, because participation was strictly voluntary, our sample may not be representative of the population at large.

Future research is needed to investigate whether delays can be reduced by: 1) educating young women about the importance of seeking prompt medical attention for new breast abnormalities and 2) increasing awareness in the medical community of appropriate evaluation and follow-up for breast abnormalities in young women. If education of young women and/or health care providers about the signs and symptoms of breast cancer facilitates more rapid diagnosis and treatment, more cancers might be excised before they metastasize outside the breast, possibly resulting in fewer deaths. Relevant interventions may be most effective if they are targeted at young women who are less financially comfortable and, thus, are more at risk of a delay in seeking medical attention.

FUNDING SUPPORT

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

This study was supported by a grant from Susan G. Komen for the Cure (Ann H. Partridge, principal investigator).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES