Background
There were 3525 cases of female breast cancer that were registered in the National Cancer Registry Malaysia in 2006, accounting for 16.5% of all cancer cases registered that year [
1]. The overall age-standardized rate was 39.3 per 100,000 populations [
1]. The cancer incidence in Malaysia is expected to increase because of increasing life expectancy, better socio-economic status and changes in lifestyle. Patients with breast cancer in Malaysia commonly present with advanced disease. The Kelantan Cancer Registry reported that 19.0% of patients presented in stage I, 25.5% in stage II, 20.7% in stage III and 34.9% in stage IV [
2]. It was previously reported that the 5-year survival rate in Kuala Lumpur was 59.1% [
3], whereas this rate was 25.8% in Kelantan [
4].
The delay in the presentation and detection of patients with breast cancer is partially responsible for the advanced stage at presentation and low survival rates in Malaysia. Early detection of cancer is important because delay is preventable and earlier treatment can lead to improved survival. Late detection has been associated with larger tumour size, increased involvement of the lymph nodes and organ metastases [
5], and negative implications on cost and choice of treatment. A study in Thailand reported that patients with stage III disease had a median delay of presentation of 2 months compared to those in stage IV (7 months) [
6]. In addition to lower survival rates, patients with delays of 12 weeks or longer had a significantly higher probability of local cancer spread or distant metastases compared to those with shorter delays [
7].
Studies in developed countries reported that the median time to consultation was 14-61 days [
8‐
10]. A delay of more than 3 months prior to physician consultation occurred in 14-53% of cases [
8‐
11]. Low socio-economic status, minority ethnicity and young age were associated with a longer duration of symptoms [
12]. Another study found that patient delay was associated with older age, lighter symptoms, not informing anyone, a negative attitude toward medical practitioners and fear of treatment [
13]. Failure of medical practitioners to act on clinical findings and false-negative mammogram and fine needle aspiration cytology (FNAC) were the main factors for system delay [
14,
15]. Lack of knowledge regarding risk factors, individual own risk and the variability of symptoms of breast cancer, was related to patient delay [
16].
This study was conducted to identify the magnitude of the delay in breast cancer diagnosis and the factors associated with this delay. To date, most of the research on this topic has been conducted in developed countries and among minorities, and very few studies have been conducted in less developed countries. No studies on this subject in Malaysia have been published. Research on this topic is important for clinicians to have a better understanding of how to manage patients and so that policy makers can implement strategies and activities to prevent delay in breast cancer diagnosis.
Methods
This study had a cross-sectional design. The respondents were women who were diagnosed with primary breast cancer by histo-pathological examination between 2005 and 2007. We excluded patients with cognitive problems, recurrent cancer and incomplete medical records. Respondents were from three referral medical centres in the East Coast of Malaysia and two government hospitals in Kuala Lumpur. All patients who attended surgical, oncology and radiotherapy clinics at the East Coast hospitals during our study period were included. Systematic sampling was conducted on every fourth eligible patient at Kuala Lumpur hospitals.
Face-to-face interviews were conducted using standardized questionnaires. The questionnaire was developed in the Malay language based on expert discussions and a literature review. It was pre-tested for face and content validity and reliability, which were satisfactory (Cronbach's Alpha 0.63-0.92). The content included socio-demography, medical and obstetric history, the date of all the chronological events (i.e., first recognition of symptoms, first consultation, referral, first hospital appointment, first meeting with the surgeon and oncologist, diagnostic tests and when the results were known) and the use of alternative therapy. We considered the symptoms' interpretation process spanned from the point at which the women started to notice abnormalities until a diagnosis was made [
17]. An agreement was decided after discussion with the respondents when there were conflicting dates of events. The respondents were reminded of events in the calendar year, such as festival celebrations, Independence Day, school holidays or birth dates, to help them remember important dates relative to their medical history.
We also included yes-no questions on the interpretation of symptoms, knowledge about presenting symptoms, aetiology and metastatic organs, beliefs about breast cancer and treatment, fear, denial, barriers, healthcare services, husband support, attitude on medical care and treatment, and health care practices. A 'yes' response for positive questions was given a score of one and a 'no' response was given a zero. The scoring was reversed for the negative questions. All scores for each group were totalled and then categorised using the median of the total score. We also collected data from the medical records including clinical presentation, diagnostic tests, histo-pathological reports and treatment information.
The Psycho-physiological Comparison Theory developed by Andersen, Cacioppo and Roberts was taken into consideration when we formulated the operational definition of delay [
18]. Consultation time was the time taken to visit the first general practitioner after the recognition of symptoms. The time to diagnosis was measured from the date of the recognition of symptoms to the date of final diagnosis based on histo-pathological examination of FNAC, trucut or excision biopsy. Breast cancer was staged according to the 6
th edition of Cancer Staging Manual published by American Cancer Joint Committee on Cancer [
19]. A family history of breast cancer was defined as having a first-degree relative, i.e. sister, mother or daughter who had breast cancer. The respondents were questioned about previous use of oral contraceptives, hormone replacement therapy or alternative therapy if these therapies were taken regularly for at least one month. Complementary alternative therapy was defined as any therapy using methods and products not included in conventional modern medicine. Chronic diseases assessed in this study included hypertension, diabetes, heart diseases, asthma and other diseases that require lifelong monitoring. Misdiagnosis occurred when there was a false-negative mammogram or FNAC.
The research had been approved by the Research and Ethical Committee from all the respective institutions, with the reference numbers: UKM1.5.3.5/244/SPP2, HKL/98/AM.882, USMKK/PKK/JK EP(M)-191 USM, Bil(43)HRPZ ll.71/20 Jld.8, HSNZ.KT.100-22/15(27) and (4)KKM/NIHSEC/08/0804/P07-13. Respondents were explained about the research and asked for consent prior the interview. All information was confidential and individual data had no identification of the respondents.
Statistical analyses
Data were analysed using SPSS for Windows (version 12.0.1, SPSS Inc., Chicago, IL, USA). Continuous data were summarised as mean (standard deviation (SD)) or median (interquartile range (IQR)) depending upon the normality of distribution, whereas categorical data were presented as frequency (percentage (%)). We divided the diagnosis time into a binary outcome, i.e., delay and non-delay, by using a six-month cut-off point. Six months was chosen instead of three month to allow balance number of respondents in each category. Multiple logistic regression was used to identify the factors associated with diagnosis delay. A stepwise backward selection procedure was used when selecting significant variables in the model. The interaction terms and multi-collinearity problem of the final model were checked. The final model was tested for fitness using the Hosmer-Lemeshow goodness of fit test. Results were presented as the crude and adjusted odd ratios (OR), 95% confidence interval (CI) and p value. The p value <0.05 was considered to indicate statistical significance.
Discussion
Our study found that consultation and diagnosis delay among breast cancer patients are very serious problems in Malaysia. In the present study, the median times before consulting a medical practitioner and before diagnosis were longer [
6,
8,
9,
20], and the prevalence of delay was higher, than in other studies in developed and developing countries [
5,
6,
9‐
11].
Our study found that the clinical presentation was associated with a delay, which is supported by the literature [
21]. Our study also found that patients who presented with breast ulcer and palpable axillary lymph nodes had significantly delayed diagnoses. The delay would eventually cause breast cancer symptoms to worsen. The first symptoms of breast cancer are usually not debilitating and can be ignored until the occurrence of new symptoms or worsening of symptoms. Patients with aggressive types of breast cancer usually had progressive symptoms in a short time, causing them to have a high index of cancer suspicion and seek early consultation and diagnosis. Patients who had a breast lump had earlier consultations than those with nipple discharge [
9] or pain [
22] or non-lump symptoms [
23].
False-negative FNAC or mammogram contributed to the delayed diagnosis of breast cancer in our study. Only 241 (73.5%) of respondents had FNAC and only 51.2% had positive results of breast cancer and 12.9% reported as benign. The false-negative FNAC rate in our study was higher than that reported in other studies, with only 1.0-1.9% [
15,
24,
25]. Most of the errors occurred during the performance of the FNAC procedure or the interpretation of the pathological report.
Only 136 (41.5%) of respondents in our study had mammogram and 8.8% of those were false negative, which is a higher rate than the previous study of only 3% [
15]. The acceptable false negative rate of mammogram was 10-15% [
26]. The failure of mammograms to detect breast cancer was related to the limitations of the film screen, poor radiographic technique, exceptional tumour characteristics and error in interpreting the film [
26]. In Malaysia, mammograms are not used for screening except for high-risk women who have had previous breast cancer or a strong family history of breast cancer. It is not used for screening because of the high cost and limited expertise in Malaysia. In Malaysia, there are clinical practice guidelines for breast cancer management, but the use and implementation of the guidelines are unknown. More widespread implementation of these guidelines may increase the quality of care for breast cancer patients and shorten the diagnosis delay.
In our study, 42.7% of the respondents had taken alternative therapy, compared to 14.8-73.1% in Europe [
27]. Most patients took alternative therapy as a way to avoid surgery or when they perceived modern medicine would not cure the disease, when the prognosis was fatal, when the disease caused suffering or in cases of chronic disease. Some patients believed that there were no effective treatments for breast cancer or that traditional medication is more effective than modern medicine. While pursuing alternative treatments, most patients experienced worsening symptoms, which led to them eventually presenting at a more advanced stage. A systematic review reported that people practiced complementary alternative medicine because of its benefits and because they wanted to be in control of their treatment, had strong beliefs and used it as the last hope [
27]. Some patients also have less trust in modern medicine because they had bad previous experiences or felt that the system was not as friendly as traditional healers or shaman. Complementary alternative therapy is also easily available and affordable.
Patients' interpretation of their symptoms as a sign of cancer had an important influence on whether they sought medical help immediately [
28]. The evaluation of breast symptoms is based on the pre-existing knowledge, experience, self-education and observation of individuals [
18]. Knowledge regarding the variation of symptoms in breast cancer enables patients to interpret the symptoms correctly and influences their assessment of symptoms as well as their decision to seek medical attention [
13].
Patients are more inclined to attribute new symptoms to less serious conditions instead to a life threatening disease [
18]. Patient delay has been reported for patients who assumed that symptoms were benign [
23] and would disappear without intervention. Perception of the seriousness of a symptom is dependent upon the first symptom and how fast the symptom changes and multiplies. Most of breast cancer symptoms are mild, not specific, unclear, confusing, do not require urgent attention and can be ignored temporarily.
Our study found that a negative perception of breast cancer treatment prevented patients from receiving early diagnoses, similar to a study in Nigeria [
29]. Negative information, such as the side-effects of chemotherapy, led to fear and caused some patients to refuse treatment. Some believed that the effects of chemotherapy were worse than breast cancer itself. Another negative perception of breast cancer treatment was related to the traditional woman's role in the family of taking care of children and the husband. Some believed that treatment would disrupt and burden their family because they could not perform their usual roles and might even have to rely on others to care for them. Because the women could not take care of the family and the husbands usually had difficulty in taking over the roles of the women, the husband might choose to separate or take another wife. Fear of divorce or the husband remarrying could lead some women to decide not to get their symptoms diagnosed if they suspected breast cancer. Some patients also believed that breast cancer could not be cured [
17], so there was no point of having it diagnosed and treated. Diagnosis delay was also related to a belief that mastectomy causes disfigurement and disability [
28].
The strength of our study lies in the fact that it was conducted at five large medical centres in Malaysia. One of the medical centres on the East Coast was the only centre that offered oncology and radiotherapy services in that area. The chances of patients receiving treatment elsewhere were minimal because most patients could not afford private services. However, a multicenter study leads to a considerable variation in the management of breast cancer patients because medical practitioners from different hospitals have different preferences.
We collected relevant dates based on the detailed approach described by Andersen, Cacioppo and Roberts [
18]. Dates of events were collected via interviews with support from the medical records. Previous report suggests that the collection of actual date is more accurate than asking about the duration of time [
30]. Our study used face-to-face interviews, in contrast to most other studies that used the existing medical records, postal or telephone interviews that had low response rates. Researchers have no control over the data if the medical records are used. There are variations in the definitions of variables, many missing data and a limited number of variables that can be collected when medical records are used.
There were differences in defining and categorising the delay. Most studies divided delays into patient and system delays [
8,
13], and many studies used a cut off point of three months for categorising delay [
5,
8,
9,
14,
20]. In our study, we use a six-month cut off point for delay because our respondents experienced longer delay.
There was a selection bias in recruiting respondents because they were included after the diagnosis of breast cancer was made. Our study did not include those who had died before the study was conducted or those who were lost to follow up. However, it was assumed that those who died early had an advanced stage of disease and had longer delay, suggesting that the diagnosis delay was under-estimated in our study.
Selection bias also occurred because our study was conducted in hospitals and some patients may not have presented to the hospital at all, instead preferring alternative treatment. A population-based study was not possible because of logistic problems in enrolling those patients.
Furthermore, there was information bias in this study. Our study method relied on the patients' recall of the events leading up to their diagnosis. Patients who delayed consultation and diagnosis needed to remember more distant events than those who did not delay. We interviewed respondents after the diagnosis was made instead of after the recognition of symptoms, so some of the variables could not be measured retrospectively. We also could not obtain detailed information on the type and quality of the diagnostic procedures because we were not involved in the management of the patients and procedural details were not included in the patients' medical records.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
BN participated in the conception and design of the study, obtaining of funding, acquisition, analysis and interpretation of the data and drafting the manuscript. KGR participated in the conception and design of the study, obtaining of funding and revising the manuscript critically. MAR participated in the conception and design of the study, analysis and interpretation of data and drafting the manuscript. NNN participated in the obtaining of funding, analysis and interpretation of data. BMB participated in the enrolment of patients and acquisition of data. All authors read and approved the final manuscript.