Abstract

Background: Delay in diagnosis of breast cancer can occur at several points on the diagnostic pathway. We examined characteristics of women with breast cancer who before diagnosis actively refused recommended follow-up of tests or symptoms suggestive of breast cancer. Methods: We identified women aged 50 years or older diagnosed with late-stage (metastatic disease or tumors ≥ 3cm at diagnosis) and a matched sample of women with early-stage (tumors < 3 cm) breast cancer from 1995 to 1999. Using medical records, we investigated clinical characteristics, use of health care, and documentation of care refusal during the 3 years before diagnosis. We used logistic regression models to compare refusers to nonrefusers. Results: Of the 2694 women studied, 7.2% refused provider follow-up advice during the 3 years. These women were more likely to have late-stage breast cancer at diagnosis than were nonrefusers (odds ratio [OR] = 1.9, 95% confidence interval [CI] = 1.4 to 2.6). They were more likely to be aged 75 years or older (OR = 1.9, 95% CI = 1.4 to 2.7 compared with age 50–64) or to have six or more children (OR = 2.3, 95% CI = 1.3 to 4.2 compared to women with one to two children). Clinical factors associated with refusal included low use of mammography, high use of clinical breast exam, and missed appointments. A minority of women who refused had a reason documented in the medical record; the most frequent reasons were avoidance–denial–fatalism, fear of diagnostic tests, and fear of surgery or disfigurement. Conclusions: Our results suggest that certain demographic and clinical characteristics are associated with women's refusal of diagnostic testing for breast cancer. Further study is needed on refusers' characteristics and on how such refusals affect outcomes. Efforts aimed at identifying and counseling women with abnormal results who refuse follow-up are warranted.

Delay in timely diagnosis of breast cancer can occur at several points on the diagnostic pathway. A growing body of literature has investigated demographic and personal characteristics associated with low use of breast cancer screening (15) and factors associated with delay in seeking medical care for breast cancer symptoms (610). Also, diagnosis can be delayed after an abnormal mammogram or clinical breast examination (CBE) or after breast cancer symptoms were reported to medical providers (1122), but less is known about patient characteristics associated with this delay. Delay during this period can be categorized into three main types: patient delay, provider delay, and delay due to system factors (11). Of these, patient delay appears to be a major contributor to untimely follow-up in some populations (1214), and explicit patient refusal to consent to timely follow-up is one cause of patient delay (13,15,16).

Patient characteristics reportedly associated with delayed follow-up of a positive mammogram or CBE include low estimated household income (17), poor health status (12,18), no history of previous screening mammograms (17), few mammograms in the 5 years before diagnosis (12), absence of a palpable lump and atypical symptoms at initial medical consultation (19,20), patient underestimation of the importance of following up a positive test (18), travel or holiday plans (18), patient too busy (18), patient fearful (18), and transportation problems (12). However, little is known about patient characteristics associated with explicit refusal of follow-up.

In a case–control study investigating reasons for late-stage breast cancer in seven health maintenance organizations, our research team found that women with medical record–documented evidence of patient noncompliance with or refusal of breast care before breast cancer diagnosis had higher odds of late-stage disease than women who did not refuse care (23). We conducted a secondary analysis in this population to obtain more information on women who refused one or more recommendations for follow-up of tests or symptoms suggestive of breast cancer during the 3 years before breast cancer diagnosis. We examined the characteristics of these women who refused important aspects of patient care, despite having obtained comprehensive prepaid healthcare coverage. We also investigated the reasons the women gave for their actions and the settings in which the refusals occurred.

METHODS

Setting

This study was conducted within the Cancer Research Network (CRN), a consortium of research organizations affiliated with nonprofit integrated healthcare delivery systems and the National Cancer Institute. The CRN consists of the research programs, enrollee populations, and databases of 11 integrated healthcare organizations that are members of the HMO Research Network. The health care delivery systems participating in the CRN are Group Health Cooperative, Harvard Pilgrim Health Care, Henry Ford Health System/Health Alliance Plan, HealthPartners Research Foundation, the Meyers Primary Care Institute of the Fallon Healthcare System/University of Massachusetts, and Kaiser Permanente in six regions: Colorado, Georgia, Hawaii, Northwest [Oregon and Washington], Northern California, and Southern California. The 11 health plans have nearly 10 million enrollees. The CRN conducts collaborative research on variations in cancer prevention and treatment policies and practices.Seven health plans (Kaiser Permanente Northwest, Kaiser Permanente Northern California, Kaiser Permanente Southern California, Kaiser Permanente Colorado, Kaiser Permanente Hawaii, Henry Ford Health System, and Group Health Cooperative) participated in a study to evaluate the reasons for late-stage breast cancer in prepaid health plans. These health plans provide comprehensive primary care, including cancer screening, diagnosis and treatment, to about 1.5 million women aged 50 years and older. Details of the study design and methods are described elsewhere (23). The study was approved by the institutional review boards of all seven health plans; all IRBs waived informed consent for this data-only study.

Study Subjects

In brief, the study included 1347 late-stage breast cancer cases (women with metastatic cancer or tumors ≥ 3 cm.) diagnosed between 1995 and 1999 inclusive in women aged 50 years and older and 1347 control subjects with early-stage breast cancer individually matched to cases on health plan, age within 1 year, and diagnosis within 6 months. Women with a prior history of breast cancer were excluded, but women with a history of breast procedures (such as implants, reductions, and biopsies) were not excluded.

Data Sources and Measures

The study collected data from medical records, cancer registries, health plan membership databases, pharmacy databases, and U.S. Census records. Data items collected included demographic variables (age, ethnicity, race, marital status, and census block group), data on the breast cancer (date of diagnosis, tumor characteristics), and information on all breast-related medical visits in the 3 years up to diagnosis (breast cancer screening services provided, any breast cancer symptoms reported, examination findings, and follow-up recommendations).

The medical record abstraction collected information on refusal of breast care during two periods prior to breast cancer diagnosis. For each woman reporting breast-related symptoms during a clinical encounter or those with an abnormal mammogram or CBE during the 3-year study period, medical record abstractors obtained information about follow-up of the symptoms/abnormal result. Abstractors would choose the option “patient refused provider advice” if clinical notes documented that the woman explicitly refused recommended follow-up cancer care at the time of the recommendation. They recorded the patient's stated reason for refusing recommended services as documented by the provider in the medical record and copied the page from the medical record, so that the study analyst could confirm the abstractor's interpretation of refusal and further analyze the exact wording of the notation. The word “refusal” did not have to specifically occur in the clinical notes, but it had to be clear to the abstractor and the analyst that the patient did refuse recommended diagnostic tests at the time they were offered. We did not attempt to validate this information on refusal by contacting the patient or family. The three settings at which refusal during the study period could occur included 1) a clinical visit, 2) the mammography appointment, and 3) a follow-up telephone encounter. In addition to refusal information collected in the 3-year period before diagnosis, notes made in the chart during the 3-year study period might indicate that the patient was noncompliant or refused breast care prior to the study, and abstractors recorded this prior refusal in a separate variable.

In this analysis, we used the term “refusers” to describe women whose medical records documented at least one refusal of provider recommendations for follow-up of abnormal screening results or breast cancer symptoms during the study period. The term “nonrefusers” is used to describe the study subjects with no chart notation of refusal of follow-up during this period. We did not consider any of the following to be refusal of follow-up: refusal of screening before a suspicious finding was noted, delayed initial visit for obvious signs or symptoms, request for a second opinion, request for a temporary delay where diagnosis occurred within 3 months of the request, or refusal of one diagnostic procedure but acceptance of another (e.g., refusal of additional mammographic views but agreement to ultrasound); these subjects are classified as nonrefusers.

We used education and income information obtained from the 2000 U.S. Census for census tract of residence as surrogates for subject education and family income. We calculated the proportion of refusers and nonrefusers living in a census tract where less than 25% of women had a college education (median for nonrefusers). We also calculated the proportions of refusers and nonrefusers with geocoded family income less than $40 000 per year, $40 000 to $54 999 per year, $55 000 to $71 999 per year, and $72 000 per year or more (quartiles for nonrefusers).

Statistical Analyses

We used descriptive analysis to evaluate the demographic and clinical characteristics of refusers, overall and by stage (late versus early). We enumerated the reasons documented in the medical record for refusal and described the medical settings where refusal occurred. We also calculated the length of time between first refusal of follow-up in the study period and diagnosis date. Since the convention for selecting the diagnosis date varied slightly from one health plan to another, we maintained consistency in our analysis across health plans by using the date of the pathology procedure at which cancer was confirmed.

We compared refusers with nonrefusers on demographic and clinical factors. We fit unconditional logistic-regression models, adjusting for health plan, stage, and other covariates to calculate odds ratios for the refuser–nonrefuser comparisons. We initially stratified the analysis on stage (late/early), since stage was the basis for the original study's control group selection (23). Results within each stage were similar, so only the combined results are presented.

RESULTS

Of the 2694 women in the study population, 195 (7.2%) refused provider advice on follow-up at least once during the 3-year observation period according to medical record documentation. This proportion varied by health plan with a range of 4.9%–10.3%. More refusers (64%) had late-stage breast cancer at diagnosis than did nonrefusers (49%) (odds ratio [OR] = 1.9, 95% confidence interval [CI] = 1.4 to 2.5). Since late-stage cases represented 50% of the study population but only 17.6% of the incident invasive breast cancers in these health plans during 1995–1999 (based on reporting from four sites), the overall proportion of refusers of follow-up among subsequently diagnosed breast cancer cases in the seven health plans during this period was estimated to be 6.0%.

Refusers were more likely than nonrefusers to be elderly (35% aged 75 or older compared with 22% of the nonrefusers) (OR = 1.9, 95% CI = 1.4 to 2.7 for age 75 or older compared with ages 50–64) (Table 1). Refusers and nonrefusers did not differ materially by race, menopause status, family history of breast cancer, or number of years of health plan enrollment. Refusers were more likely than nonrefusers to have six or more children (OR = 2.3, 95% CI = 1.3 to 4.2 for six or more children compared with one to two children).

Table 1.

Demographic characteristics of breast cancer cases, by evidence of refusal of follow-up*

 Evidence of refusal of follow-up (N = 195)
 
 No evidence of refusal of follow-up (N = 2499)
 
  
Characteristic OR (95% CI) 
Race      
    White, non-Hispanic 157 80.5 1913 76.6 Reference 
    Non-white or Hispanic 36 18.5 578 23.1 0.86 (0.57 to 1.3) 
    Unknown 1.0 0.3 — 
Age group      
    50–64 82 42.1 1302 52.1 Reference 
    65–74 44 22.6 638 25.5 1.1 (0.72 to 1.6) 
    ≥75 69 35.4 559 22.4 1.9 (1.4 to 2.7) 
Postmenopausal      
    Yes 155 79.5 2030 81.2 Reference 
    No 19 9.7 190 7.6 1.2 (0.84 to 1.8) 
    Unknown 21 10.8 279 11.2 — 
Marital status      
    Married at diagnosis 96 49.2 1450 58.0 Reference 
    Not married at diagnosis 86 44.1 902 36.1 1.3 (0.91 to 1.7) 
    Unknown 13 6.7 147 5.9 — 
Parity      
    0 31 15.9 335 13.4 1.4 (0.86 to 2.2) 
    1–2 63 32.3 977 39.1 Reference 
    3–5 58 29.7 788 31.5 1.2 (0.81 to 1.7) 
    ≥6 16 8.2 107 4.3 2.3 (1.3 to 4.2) 
    Unknown 27 13.9 292 11.7 — 
Family history of breast cancer      
    Any 54 27.7 790 31.6 0.88 (0.63 to 1.2) 
    None 117 60.0 1475 59.0 Reference 
    Unknown 24 12.3 234 9.4 — 
Total years of health plan enrollment      
    3–7 31 15.9 463 18.6 0.91 (0.59 to 1.4) 
    8–12 31 15.9 470 18.8 0.84 (0.55 to 1.3) 
    ≥13 133 68.2 1566 62.7 Reference 
<25% of women in census tract with college education 92 47.2 1187 47.5 0.98 (0.73 to 1.3) 
Median income in census tract      
    <$40 000 59 30.3 679 27.2 Reference 
    $40 000–$54 999 52 26.7 623 24.9 0.97 (0.65 to 1.4) 
    $55 000–$71 999 42 21.5 614 24.6 0.81 (0.52 to 1.3) 
    ≥$72 000 42 21.5 583 23.3 1.1 (0.68 to 1.7) 
 Evidence of refusal of follow-up (N = 195)
 
 No evidence of refusal of follow-up (N = 2499)
 
  
Characteristic OR (95% CI) 
Race      
    White, non-Hispanic 157 80.5 1913 76.6 Reference 
    Non-white or Hispanic 36 18.5 578 23.1 0.86 (0.57 to 1.3) 
    Unknown 1.0 0.3 — 
Age group      
    50–64 82 42.1 1302 52.1 Reference 
    65–74 44 22.6 638 25.5 1.1 (0.72 to 1.6) 
    ≥75 69 35.4 559 22.4 1.9 (1.4 to 2.7) 
Postmenopausal      
    Yes 155 79.5 2030 81.2 Reference 
    No 19 9.7 190 7.6 1.2 (0.84 to 1.8) 
    Unknown 21 10.8 279 11.2 — 
Marital status      
    Married at diagnosis 96 49.2 1450 58.0 Reference 
    Not married at diagnosis 86 44.1 902 36.1 1.3 (0.91 to 1.7) 
    Unknown 13 6.7 147 5.9 — 
Parity      
    0 31 15.9 335 13.4 1.4 (0.86 to 2.2) 
    1–2 63 32.3 977 39.1 Reference 
    3–5 58 29.7 788 31.5 1.2 (0.81 to 1.7) 
    ≥6 16 8.2 107 4.3 2.3 (1.3 to 4.2) 
    Unknown 27 13.9 292 11.7 — 
Family history of breast cancer      
    Any 54 27.7 790 31.6 0.88 (0.63 to 1.2) 
    None 117 60.0 1475 59.0 Reference 
    Unknown 24 12.3 234 9.4 — 
Total years of health plan enrollment      
    3–7 31 15.9 463 18.6 0.91 (0.59 to 1.4) 
    8–12 31 15.9 470 18.8 0.84 (0.55 to 1.3) 
    ≥13 133 68.2 1566 62.7 Reference 
<25% of women in census tract with college education 92 47.2 1187 47.5 0.98 (0.73 to 1.3) 
Median income in census tract      
    <$40 000 59 30.3 679 27.2 Reference 
    $40 000–$54 999 52 26.7 623 24.9 0.97 (0.65 to 1.4) 
    $55 000–$71 999 42 21.5 614 24.6 0.81 (0.52 to 1.3) 
    ≥$72 000 42 21.5 583 23.3 1.1 (0.68 to 1.7) 
*

Logistic regression models adjusted for site, age group, and stage. CI = confidence interval; — = odds ratios not calculated for subjects classified as “unknown”.

Quartiles based on distribution among nonrefusers.

We observed no association between refusal status and median family income of the census tract of residence or between refusal status and living in a census tract with less than 25% of women with a college education. These variables did not alter the results of any of the other demographic or clinical variables.

Medical record documentation of refusal or noncompliance before the study period was more common for refusers during the study period (17%) than nonrefusers during the study period (5%) (OR = 3.9, 95% CI = 2.5 to 6.2) (Table 2). Among refusers, 82 (42%) first refused care in the study period between 1 and 3 years before breast cancer diagnosis, whereas 113 (58%) refused care in the year before breast cancer diagnosis but not during the 2 years previously. We calculated a median of 363 days from first refusal of follow-up in the study period to the date of the pathology procedure at which cancer was confirmed. For 25% of refusers, fewer than 32 days passed between first refusal and pathology confirmation. Another 25% of refusers had more than 746 days (slightly more than 2 years) between refusal and pathology confirmation.

Table 2.

Clinical characteristics and medical care use, by evidence of refusal of follow-up

 Evidence of refusal of follow-up (N = 195)
 
 No evidence of refusal of follow-up (N = 2499)
 
  
Characteristic OR*(95% CI) 
Late-stage breast cancer 124 63.6 1223 48.9 1.9 (1.4 to 2.6) 
Symptoms of breast cancer      
    Yes 119 61.0 1449 58.0 Reference 
    No 17 8.7 206 8.2 1.0 (0.72 to 1.7) 
    Unknown 59 30.3 844 33.8 — 
Method of diagnosis      
    Screening mammogram 37 19.0 762 30.5 Reference 
    Screening CBE 16 8.2 130 5.2 2.3 (2.3 to 4.3) 
    Opportunistic CBE 15 7.7 82 3.3 3.2 (1.6 to 6.2) 
    Nonscreening mammogram 43 22.1 635 25.4 1.3 (0.8 to 2.1) 
    Nonscreening CBE 78 40.0 853 34.1 1.1 (1.1 to 2.6) 
    Other 3.1 37 1.5 2.5 (0.97 to 6.6) 
No. of outpatient visits      
    0 11 5.6 108 4.3 1.4 (0.67 to 3.1) 
    1–10 57 29.2 664 26.6 1.2 (0.85 to 1.7) 
    ≥11 127 65.1 1727 69.1 Reference 
Any mammogram 160 82.1 2363 94.6 0.34 (0.22 to 0.54) 
Any screening mammogram 100 51.3 1812 72.5 0.45 (0.33 to 0.62) 
No. of clinical breast exams      
    0 0.5 38 1.5 0.48 (0.06 to 3.5) 
    1–2 37 18.9 698 27.9 Reference 
    3–4 69 35.4 919 36.9 1.6 (1.1 to 2.5) 
    ≥5 88 45.1 844 33.8 2.5 (1.7 to 3.8) 
No. of missed appointments      
    0 144 73.9 2103 84.2 Reference 
    1 29 14.9 236 9.4 1.9 (1.2 to 3.0) 
    2 10 5.1 91 3.7 1.9 (0.91 to 3.9) 
    ≥3 12 6.2 69 2.7 2.5 (1.3 to 5.1) 
Refusal or non-compliance more than 3 y before diagnosis 33 16.9 112 4.5 3.9 (2.5 to 6.2) 
 Evidence of refusal of follow-up (N = 195)
 
 No evidence of refusal of follow-up (N = 2499)
 
  
Characteristic OR*(95% CI) 
Late-stage breast cancer 124 63.6 1223 48.9 1.9 (1.4 to 2.6) 
Symptoms of breast cancer      
    Yes 119 61.0 1449 58.0 Reference 
    No 17 8.7 206 8.2 1.0 (0.72 to 1.7) 
    Unknown 59 30.3 844 33.8 — 
Method of diagnosis      
    Screening mammogram 37 19.0 762 30.5 Reference 
    Screening CBE 16 8.2 130 5.2 2.3 (2.3 to 4.3) 
    Opportunistic CBE 15 7.7 82 3.3 3.2 (1.6 to 6.2) 
    Nonscreening mammogram 43 22.1 635 25.4 1.3 (0.8 to 2.1) 
    Nonscreening CBE 78 40.0 853 34.1 1.1 (1.1 to 2.6) 
    Other 3.1 37 1.5 2.5 (0.97 to 6.6) 
No. of outpatient visits      
    0 11 5.6 108 4.3 1.4 (0.67 to 3.1) 
    1–10 57 29.2 664 26.6 1.2 (0.85 to 1.7) 
    ≥11 127 65.1 1727 69.1 Reference 
Any mammogram 160 82.1 2363 94.6 0.34 (0.22 to 0.54) 
Any screening mammogram 100 51.3 1812 72.5 0.45 (0.33 to 0.62) 
No. of clinical breast exams      
    0 0.5 38 1.5 0.48 (0.06 to 3.5) 
    1–2 37 18.9 698 27.9 Reference 
    3–4 69 35.4 919 36.9 1.6 (1.1 to 2.5) 
    ≥5 88 45.1 844 33.8 2.5 (1.7 to 3.8) 
No. of missed appointments      
    0 144 73.9 2103 84.2 Reference 
    1 29 14.9 236 9.4 1.9 (1.2 to 3.0) 
    2 10 5.1 91 3.7 1.9 (0.91 to 3.9) 
    ≥3 12 6.2 69 2.7 2.5 (1.3 to 5.1) 
Refusal or non-compliance more than 3 y before diagnosis 33 16.9 112 4.5 3.9 (2.5 to 6.2) 
*

Logistic regression models adjusted for site, age group, and stage. OR = odds ratio, CI = confidence interval; — = odds ratios not calculated for subjects classified as “unknown”.

During period 3 years before diagnosis of breast cancer.

More than half the refusers (61%) had breast cancer symptoms noted in the medical chart in the 3 years before breast cancer diagnosis; a slightly lower proportion of nonrefusers (58%) had symptoms before diagnosis (Table 2). The two groups did not differ materially in the distribution of breast cancer symptoms reported (i.e., lump, pain, nipple discharge, visual change, odor). Fewer refusers (19%) than nonrefusers (31%) were diagnosed using screening mammography; refusers were more likely than nonrefusers to be diagnosed using CBE.

The refusers used the medical care system; almost all women had at least one outpatient visit in the study period, and all but one had at least one CBE in the study period (Table 2). Refusers were less likely to have mammography in the study period than nonrefusers (OR = 0.34, 95% CI = 0.22 to 0.54), but they were more likely to have five or more CBEs (OR = 2.5, 95% CI = 1.7 to 3.8), compared to 1 or 2 CBEs. Refusers were more likely than nonrefusers to have missed one or more appointments (OR = 2.0, 95% CI = 1.4 to 2.9).

For the women who refused care in the study period, we reviewed the reasons for refusal noted in the medical record (Table 3). Although in most cases (68%) no reason was recorded, the most frequently documented reasons for refusal included avoidance–denial–fatalism (11%), fear of further diagnostic tests (4%), fear of surgery and/or disfigurement (3%), and discomfort–pain of mammogram (3%).

Table 3.

Timing, setting, service type, and documented reasons for refusal of breast cancer follow-up among women with breast cancer

 All subjects with evidence of refusal of follow-up (N = 195)
 
 
Characteristic 
Time period of first refusal*   
    1–3 years before diagnosis 82 42.1 
    <1 year before diagnosis but not 1–3 years before diagnosis 113 57.9 
Service setting where refusal occurred*,   
    Mammography or ultrasound visit 14 6.6 
    Other clinical encounter 165 77.5 
    Telephone contact 34 16.0 
Medical service refused   
    Immediate clinical evaluation 93 47.7 
    Return for additional examination 46 23.6 
    Repeat examination in 4–6 mo 19 9.7 
    Other or recommendation undetermined 37 19.0 
Stated reason for refusal according to notation in medical record   
    Avoidance–denial–fatalism 22 11.3 
    Discomfort/pain of mammogram 2.6 
    Fear of further diagnostic tests 3.6 
    Fear of surgery and/or disfigurement 3.1 
    Fear of other treatment 0.5 
    Belief that cancer treatment would cause or hasten death 2.1 
    Psychiatric problems 1.5 
    Family problems 1.0 
    Transportation problems 0.5 
    Claimed to be followed by another provider 2.1 
    Planned to seek nonallopathic alternative care 1.0 
    Advanced age 1.0 
    Too ill 0.5 
    Religious beliefs discouraging medical treatment 0.5 
    Unhappy with health care 0.5 
    Undetermined 133 68.2 
 All subjects with evidence of refusal of follow-up (N = 195)
 
 
Characteristic 
Time period of first refusal*   
    1–3 years before diagnosis 82 42.1 
    <1 year before diagnosis but not 1–3 years before diagnosis 113 57.9 
Service setting where refusal occurred*,   
    Mammography or ultrasound visit 14 6.6 
    Other clinical encounter 165 77.5 
    Telephone contact 34 16.0 
Medical service refused   
    Immediate clinical evaluation 93 47.7 
    Return for additional examination 46 23.6 
    Repeat examination in 4–6 mo 19 9.7 
    Other or recommendation undetermined 37 19.0 
Stated reason for refusal according to notation in medical record   
    Avoidance–denial–fatalism 22 11.3 
    Discomfort/pain of mammogram 2.6 
    Fear of further diagnostic tests 3.6 
    Fear of surgery and/or disfigurement 3.1 
    Fear of other treatment 0.5 
    Belief that cancer treatment would cause or hasten death 2.1 
    Psychiatric problems 1.5 
    Family problems 1.0 
    Transportation problems 0.5 
    Claimed to be followed by another provider 2.1 
    Planned to seek nonallopathic alternative care 1.0 
    Advanced age 1.0 
    Too ill 0.5 
    Religious beliefs discouraging medical treatment 0.5 
    Unhappy with health care 0.5 
    Undetermined 133 68.2 
*

During period 3 years before diagnosis of breast cancer.

18 subjects refused in multiple settings.

One category per person, in priority order top to bottom.

Most documented refusals occurred at a clinical encounter (Table 3). The primary medical services refused by the women involved immediate clinical evaluation and included biopsy, immediate mammogram, and surgical referral. Only two of the 14 women who refused care at a mammography visit also refused care at a clinic visit. Of the 34 subjects who refused care at a telephone contact, 16 (47%) had also refused care at a clinic visit, whereas 18 (53%) had no documentation of previous refusal of care at a clinic visit or mammography department visit. The main procedures refused at a telephone contact involved immediate clinical evaluation (surgical referral and immediate mammogram).

DISCUSSION

This report describes a group of women who had prepaid healthcare coverage but actively refused recommended diagnostic tests for breast cancer. Most had frequent contact with the medical system during the 3-year period before breast cancer diagnosis, and most had received at least one mammogram and CBE during this period. Most of these women were later diagnosed with late-stage breast cancer.

Among demographic factors, advanced age and high parity were most strongly associated with refusal of follow-up care in our study. Concerns about poor general health status may have influenced the decisions of the elderly women; for two, advanced age was mentioned as the reason for refusal. Otherwise, refusal reasons mentioned by elderly women and women with high parity were similar to reasons given by other refusers.

Among clinical and medical-care factors, refusers were more likely than nonrefusers to have a history of refusal of care before the study period. Refusers were also about twice as likely as nonrefusers to have a history of missed medical appointments. Refusers were less likely than nonrefusers to have screening mammograms in the study period or to be diagnosed using mammography. Several women cited the discomfort and pain of mammograms as the reason for refusing follow-up mammography and for refusing mammograms prior to the study period. Dislike or fear of mammography also may have been the reason that refusers were more likely than nonrefusers to have more CBEs in the study period; women may have requested CBE as an alternative procedure. Two previous studies conducted in prepaid health plans also found a deficit of mammography among women with delayed follow-up of a positive mammogram (12,17).

In our study, the reasons for refusal were not well documented in the health plan medical records. Records contained no information on the reason for refusal for more than two-thirds of refusers. Avoidance and denial were reported for about 11%. Fear of diagnostic tests and surgery and fear of disfigurement were also expressed by some. Few subjects in these prepaid health plans claimed that financial or family concerns prevented follow-up.

Future efforts encouraging providers to accurately record reasons for refusal of follow-up would be valuable for later intervention and to better understand this behavior.

To our knowledge, few other studies have ascertained patient reasons for refusal of or noncompliance with diagnostic testing for breast cancer. In a study of elderly African American women attending a public clinic for treatment of chronic illnesses (13), 27% of the women with abnormal mammograms refused diagnostic follow-up. Reported reasons for refusal included nonspecific “fears” and the belief that they were too old for treatment Other reasons cited by these women included the desire to “not want to know if something is wrong” and the perception that “nothing was bothering them.” A study investigating patient reports of processes of care related to breast cancer screening follow-up (24) in four prepaid health plans found high adherence with recommendation for immediate evaluation but lower compliance with request for additional imaging or 4- to 6-month follow-up. Overall, for 7% of subjects with abnormal mammograms, one or more recommended tests were not done. Subjects reported that lack of adherence was partially due to confusing and conflicting information provided by medical personnel.

Limitations of our study include using medical records to ascertain the reasons for refusal, which resulted in incomplete documentation and could have been affected by documentation bias, since we collected provider notation of the reasons for refusal, not the women's self-report. Providers' perceptions of these reasons may not have been accurate, and providers may have selectively recorded some reasons but not others. We did not have information on the level of education or family income of study subjects and instead used estimates derived from the U.S. Census block groups. Despite these limitations, the large study size and the ethnically and geographically diverse population were major study strengths. Other study strengths included excellent access to medical records to ascertain patterns of health care. Almost all selected subjects were included in the analysis. The study collected detailed information on breast-related clinical encounters and mammography visits and specifically attempted to collect as much information as possible on patient refusal of care.

In conclusion, we found patient refusal of follow-up in a proportion of this population of health plan members with breast cancer. Patient characteristics associated with refusal of follow-up included advanced age, high parity, low use of mammography, high use of CBE, and missed appointments. Although all study subjects were eventually diagnosed with breast cancer, our data suggest that refusal of follow-up is associated with delay in diagnosis and treatment of same. Further study is needed on the characteristics of women who refuse follow-up tests for breast cancer and the impact of such refusal on outcomes. Also, efforts aimed at identifying and counseling women with abnormal results who refuse follow-up are warranted.

We acknowledge the DETECT Study team members who, over several years, worked to help us better understand screening implementation. Team members include (GHC) Deb Casso, Laura Ichikawa, Sarah Parkhurst; (KPNC) Wendy Leyden, Noelle Blick, Rowena Allison, Reggie Jackson, Barbara Rowe; (KPSC) Carmen West; (KPNW) Weiming Hu, Deborah Reck, Jill Mesa, Terry Kimes; (KPH) Denise S. Williams, Mark M. Schmidt; (HFHS) Karen Wells, Susan McGuinness, Lisa May, Patricia Baker, Cheryl Spoutz; (KPCO) Jennifer Ellis. We thank Jill Mesa, KPNW, and Weiming Hu, KPNW, for assistance with data preparation and analysis for the manuscript.

References

(1)
Harris TJ, Cook DG, Shah S, Victor CR, DeWilde S, Beighton C, et al. Mammography uptake predictors in older women.
Fam Pract
 
2002
;
19
:
661
–4.
(2)
Romans MC. Utilization of mammography: social and behavioral trends.
Cancer
 
1993
;
1475
–7.
(3)
Banks E, Beral V, Cameron R, Hogg A, Langley N, Barnes I, et al. Comparison of various characteristics of women who do and do not attend for breast cancer screening.
Breast Cancer Research
 
2002
;
4
:
R1
.
(4)
Meyerowitz BE, Richardson J, Hudson S, Leedhan B. Ethnicity and cancer outcomes: behavioral and psychological considerations.
Psychol Bull
 
1998
;
123
:
47
–70.
(5)
Glasgow RE, Whitlock EP, Vogt TM. Barriers to mammography and pap smear screening among women who recently had neither, one or both types of screening.
Ann Behav Med
 
2000
;
22
:
223
–8.
(6)
Caplan LS. Patient delay in seeking help for potential breast cancer.
Public Health Rev
 
1995
;
23
:
263
–74.
(7)
Facione NC. Delay versus help seeking for breast cancer symptoms: a critical review of the literature on patient and provider delay.
Soc Sci Med
 
1993
;
36
:
1521
–34.
(8)
Rameriz AJ, Westcombe AM, Burgess CC, Sutton S, Littlejohns P, Richards MA. Factors predicting delayed presentation of symptomatic breast cancer: a systematic review.
Lancet
 
1999
;
353
:
1127
–31.
(9)
Timko C. Seeking medical care for a breast cancer symptom: determinants of intentions to engage in prompt or delay behavior.
Health Psychol
 
1987
;
6
:
305
–28.
(10)
Katz SJ, Hislop TG, Thomas DB, Larson EB. Delay from symptom to diagnosis and treatment of breast cancer in Washington State and British Columbia.
Med Care
 
1993
;
31
:
264
–8.
(11)
Kerlikowske K. Timeliness of follow-up after abnormal screening mammography.
Breast Cancer Res Treat
 
1996
;
40
:
53
–64.
(12)
McCarthy BD, Ulcickas-Yood MU, Janz NK, Boohaker EA, Ward RE, Johnson CC. Evaluation of factors potentially associated with inadequate follow-up of mammographic abnormalities.
Cancer
 
1996
;
77
:
2070
–6.
(13)
Mandelblatt J, Traxler M, Lakin P, et al. Breast and cervical cancer screening of poor, elderly, black women: clinical results and implications.
Am J Prev Med
 
1993
;
9
:
133
–8.
(14)
Pal S, Ikeda DM, Birdwell RL. Compliance with recommended follow-up after fine-needle aspiration biopsy of nonpalpable breast lesions: a retrospective study.
Radiology
 
1996
;
201
:
71
–4.
(15)
Robertson CL, Kopans DB. Communication problems after mammographic screening.
Radiology
 
1989
;
443
–4.
(16)
Cardenosa G, Eklund GW. Rate of compliance with recommendations for additional mammographic views and biopsies.
Radiology
 
1991
;
181
:
359
–61.
(17)
McCarthy BD, Ulcickas-Yood MU, Boohaker EA, Ward RE, Rebner M, Johnson CC. Inadequate follow-up of abnormal mammograms.
Am J Prev Med
 
1996
;
12
:
282
–8.
(18)
Caplan LS, Helzlsouer KJ, Shapiro S, Wesley MN, Edwards BK. Reasons for delay in breast cancer diagnosis.
Prev Med
 
1996
;
25
:
218
–24.
(19)
Caplan LS, Helzlsouer KJ, Shapiro S, Freedman LS, Coates RJ, Edwards BK. System delay in breast cancer in whites and blacks.
Am J Epidemiol
 
1995
;
142
:
804
–12.
(20)
Adam SA, Horner JK, Vessey MP. Delay in treatment for breast cancer.
Community Med
 
1980
;
2
:
195
–201.
(21)
GIVIO Interdisciplinary Group for Cancer Care Evaluation. Reducing diagnostic delay in breast cancer.
Cancer
 
1986
;
58
:
1756
–61.
(22)
Caplan LS, May DS Richardson LC. Time to diagnosis and treatment of breast cancer: results from the National Breast and Cervical Cancer Early Detection Program, 1991–1995.
Am J Public Health
 
2000
;
90
:
130
–4.
(23)
Taplin S, Ichikawa L, Ulcickas-Yood M, Manos MM, Geiger AM, Weinmann S, et al. Reason for late-stage breast cancer: absence of screening or detection, or breakdown in follow-up?
J Natl Cancer Inst
 
2004
;
96
:
1518
–27.
(24)
Zapka JG, Puleo E, Taplin SH, Goins KV, Ulcickas-Yood M, Mouchawar J, et al. Processes of care in cervical and breast cancer screening and follow-up – the importance of communication.
Prev Med
 
2004
;
39
:
81
–90.