Abstract
Background. Breast cancer is the leading cause of cancer-related mortality among women in Ghana. Data are limited on the predictors of poor outcomes in breast cancer patients in low-income countries; however, prolonged waiting time has been implicated. Among breast cancer patients who received treatment at Korle Bu Teaching Hospital, this study evaluated duration and factors that influenced waiting time from first presentation to start of definitive treatment. Method. We conducted a hospital-based retrospective study of 205 breast cancer patients starting definitive treatment at Korle Bu Teaching Hospital between May and December 2013. We used descriptive statistics to summarize patient characteristics. Mann–Whitney U and Kruskal–Wallis tests and Spearman rank correlation were performed to examine the patients, health system, and health worker factors associated with median waiting time. Poisson regression was used to examine the determinants of waiting time. Results. The mean age of the patients was 51.1 ± 11.8 years. The median waiting time was 5 weeks. The determinants of waiting time were level of education, age, income, marital status, ethnicity, disease stage, health insurance status, study sites, time interval between when biopsy was requested and when results were received and receipt of adequate information from health workers. Conclusion. A prolonged waiting time to treatment occurs for breast cancer patients in Ghana, particularly for older patients, those with minimal or no education, with lower income, single patients, those with late disease, those who are insured, and who did not receive adequate information from the health workers. Time to obtain biopsy reports should be shortened. Patients and providers need education on timely treatment to improve prognosis.
Breast cancer is the most common cause of cancer morbidity and mortality among women worldwide (World Health Organization, 2012). The global breast cancer incidence increased from 641,000 cases in 1980 to 1.6 million cases in 2010 with an annual increase rate of 3.1% (Forouzanfar et al., 2011). The incidence of breast cancer is also increasing in many low- and middle-income countries due to increased life expectancy, urbanization, and adoption of Western lifestyles (World Health Organization, 2012). However, the diagnosis is made in late stages in low- and middle-income countries resulting in poor outcomes. The World Health Organization (2012) reported that about one third of breast cancer deaths worldwide could be saved every year if detected and treated early.
Although the reported estimates of breast cancer incidence are increasing in Ghana (Asumanu, Vowotor, & Naaeder, 2000), these estimates may be lower than the actual incidence partly due to underdiagnosis and underreporting of breast cancer cases considering the weak health care systems in country (Mayo & Hunter, 2003). Breast cancer is the second most common cancer among women in Ghana, with an estimated age standardized incidence rate of 25.8/100,000 and an age standardized mortality rate of 15.2/100,000 (GLOBOCON, 2008). Age standardized incidence rate in West Africa is 31.0/100,000 with an age standardized mortality rate of 10.9/100,000, while in East Africa and North America the corresponding incidence rates are 19.3 and 76.7/100,000 with mortalities of 11.4 and 14.8/100,000, respectively (Ferlay et al., 2010). In 2006, breast cancer was the commonest cause of cancer death among women in Ghana (Wiredu & Armah, 2006).
Not only is the incidence of breast cancer rising in Ghana, there is also late presentation, with 60% to 85% of patients presenting with stages III and IV disease (J. N. A. Clegg-Lamptey & Hodasi, 2007; Ohene-Yeboah & Adjei, 2012) and an average duration of symptoms before presentation is 10 months (J. N. A. Clegg-Lamptey & Hodasi, 2007). At Korle-Bu Teaching Hospital (KBTH), the mean duration of symptoms among patients reporting for the first time with breast cancer was 46 weeks and varied up to 5 years (J. Clegg-Lamptey, Dakubo, & Attobra, 2009; J. N. Clegg-Lamptey, Dakubo, & Attobra, 2009). In a study in Nigeria, it was found that many patients who had prolonged time to first presentation of breast cancer symptoms, also had prolonged time to start of definitive treatment (Sharma, Costas, Shulman, & Meara, 2012). Prolonged time interval of more than 3 months from onset of breast cancer symptoms to start of definitive treatment (total delay in treatment), is associated with decreased survival (Richards, Westcombe, Love, Littlejohns, & Ramirez, 1999; Unger-Saldaña & Infante-Castañeda, 2009). Although some of the evidence regarding the effect of treatment delay on survival is contradictory, timely diagnosis and treatment would ensure the best chance for a positive outcome in current practice (Reed, Williams, Wall, & Hasselback, 2004). For example, despite the high incidence rates in Western countries, about 89% of women diagnosed with breast cancer are still alive 5 years after diagnosis due to early detection and treatment (Parkin, Pisani, & Ferlay, 2008). In a worldwide population-based study on cancer survival in five continents, 5-year relative survival for breast cancer was >80% in North America, Sweden, Japan, Finland, and Australia; <60% in Brazil and Slovakia; and <40% in Algeria (Coleman et al., 2008).
In the United Kingdom, the National Health Service recommends a 2-month maximum wait from urgent General Practitioner referral for suspected cancers to first definitive treatment (Department of Health, 2011). The Canadian Society of Surgical Oncology recommends that patients be seen within 2 weeks of referral and that their treatment, including surgery be initiated within 2 weeks of completion of preoperative tests (Darling, Maziak, Clifton, & Canadian Association of Thoracic Surgery, 2004).
Although previous studies in Ghana have investigated the delay in time to presentation for patients with breast cancer (J. Clegg-Lamptey et al., 2009; J. N. A. Clegg-Lamptey & Hodasi, 2007), we are not aware of any study that has evaluated the time between presentation and start of definitive treatment in Ghana. Thus, the goals of this study are twofold. First, we determined the time interval between first presentation at KBTH with diagnosis of breast cancer and the start of definitive treatment. Next, we identified the patient-, health care provider-, and health care system factors that contribute to prolonged waiting time to treatment among breast cancer patients at KBTH.
Method
Study Design
This was a hospital-based retrospective study of patients who started treatment for breast cancer at the KBTH between May 2013 and December 2013.
Sampling Technique
Two hundred and five consecutive breast cancer patients starting treatment at the study site were recruited. This represented a 97% response rate. Patients were identified from the general surgical wards and outpatient clinics, and the breast clinic of the National Center for Radiotherapy and Nuclear Medicine at KBTH. At the study site (KBTH), cancer patients are first seen at either the surgical outpatient department or the National Centre for Radiotherapy, which is the oncology division of the hospital. Treatment subsequently starts either on the general surgical wards or at the radiotherapy center. KBTH being a tertiary referral center admits patients referred from all over the country for cancer treatment. Patients who first reported to the study sites and were starting definitive treatment were recruited.
Data were collected via a questionnaire and review of patients’ medical records to determine the waiting time between when they first reported to the hospital and when definitive treatment was started. Patients who qualified to be recruited for the study included all consenting patients who were histopathologically diagnosed with breast cancer and who had started definitive treatment. Those who had previous treatment for breast cancer and those unable to provide informed consent were excluded.
Dependent Variable
The primary outcome was waiting time from first presentation to treatment for breast cancer measured in weeks and obtained from patients’ medical records.
Independent Variables
The patients’ factors examined included the following: age, marital status, level of education, ethnicity, religion, income, disease stage, and health insurance status. The health worker factors included the following: the approachability of workers, whether the workers provided adequate information to patients, and if health workers were readily available to the patients. The health system factors examined were the ease with which patients navigated the hospital, site of patients’ recruitment (surgery or oncology), and time interval between when biopsy was requested and when results were received by the doctor. The health worker and health system factors were examined from the patients’ perspectives.
Data Analysis
Data were analyzed using descriptive statistics, Spearman rank correlation, Mann–Whitney U test and Kruskal–Wallis tests. Mean, median and frequencies were used to summarize the characteristics of the patients. Mann–Whitney U test and Kruskal–Wallis tests were performed to examine patients’ factors, health system, and health worker factors associated with median waiting time to initiation of definitive treatment. Specifically, Mann–Whitney U was performed for independent variables with two categories while the Kruskal–Wallis test was performed for variables with more than two categories. Spearman rank correlation analysis was used to test the association between time interval between when biopsy was requested and when results were received and waiting time to start of definitive treatment. At the multivariate level, Poisson regression was used to examine the determinants of waiting time. The significance level was set at .05.
Ethical Approval
Ethical approval for the study was obtained from the Ethical and Protocol Review Committee of University of Ghana Medical School. Informed consent was sought from the patients before the interviews were administered
Results
Characteristics of Patients
A total of 205 patients were recruited with almost equal proportions from the surgical and oncology practices. As shown in Table 1, almost all the patients were women and the mean age was 51.1 (±11.8) years. The largest proportion of patients was in the 50 and 59 years category and more than half were married. Slightly more than one third had middle/junior high school education; close to one fifth had no education; and more than 80% were Christians. Furthermore, about 9% earned less than GHS100 (USD27) monthly and 2.3% earned more than GHS2000 (USD540) monthly. Sixty percent of the patients had stage III breast cancer, 1% came with stage IV disease, and in 13.7% the disease stage was not known.
Characteristics of Patients.
Median Waiting Time (in Weeks) From Presentation to Start of Definitive Treatment
The median waiting time from presentation to start of definitive treatment was 5 weeks. As shown in Figure 1, more than 4 out of 10 of the patients started treatment before 4 weeks; close to one third started definitive treatment between 4 and 12 weeks, and slightly more than one fifth started definitive treatment after 12 weeks of presentation to the hospital. Generally, only 22.4% of the patients were of the opinion that there was a delay in treatment.
Waiting time (in weeks) from presentation to start of definitive treatment.
Factors Associated With Waiting Time From Presentation to Start of Treatment
Patient Factors
Table 2 shows that the disease stage and health insurance status were associated with median waiting time from presentation to start of definitive treatment. Specifically, patients with stages III and IV had shorter waiting time than those with stages I and II (4 weeks and 7 weeks, respectively). Also, the median waiting time for those who had health insurance (6 weeks) was higher than those without health insurance (4 weeks). The median waiting time did not differ by sex, age, marital status, level of education, ethnicity, religion, and income of the patients. However, those who were married or cohabiting had lower median waiting time compared with those who were single, although this was not statistically significant. Furthermore, the median waiting time reduced with higher level of education such that those with higher education had the lowest median waiting time to start of definitive treatment (also not statistically significant).
Factors Associated With Waiting Time From Presentation to Start of Treatment.
Health Care Worker and Health System Factors
The results show that—the health care worker factors were not significantly associated with median waiting time to treatment (see Table 3). However, two health system factors were associated with median waiting time. Specifically, there was significant difference in the waiting time depending on the sites of treatment. Patients recruited from the surgery site had higher median waiting time in the start of definitive treatment compared with patients who received care at the oncology site (6 weeks and 4 weeks, respectively; Table 3). Furthermore, the median time interval between when biopsy was requested and when results were received was 6.0 weeks. There was a significant positive correlation between waiting time for biopsy results and waiting time in the start of definitive treatment. The magnitude of the association between waiting time for biopsy results and start of treatment was r = .348 (p = .001).
Health Workers and Health System Factors Associated With Waiting Time.
Determinants of Waiting Time From Presentation to Start of Treatment
Table 4 shows the determinants of waiting time from presentation to start of definitive treatment. The tables show that level of education, age, income, marital status, ethnicity, disease stage, health insurance status, study sites, time interval between when biopsy was requested and when results were received, and receipt of adequate information from health workers were the determinants of waiting time from presentation to start of definitive treatment. However, religion was not significantly related to waiting time to start of definitive treatment.
Determinants of Delay in Breast Cancer Treatment.
Note. IRR = incident rate ratio; RC = reference category.
p < .05. **p < .01. ***p < .001.
Specifically, the results show that those with middle/junior high school, secondary/senior high school and postsecondary education had significantly shorter waiting time than those with no education (44.6%, 39.2%, and 35.4%, respectively). Also, those who were 50 to 59 years and 60 to 69 years had significantly longer waiting time than those who were aged 40 years and younger. However, there was no significant difference in waiting time between those who were 40 and 49 years and those who were aged 40 years and younger. Furthermore, those with no income and those who earned less than GHC 1,000 had significantly longer waiting time than those who earned GHC 2,000 and above. With regard to marital status, those who were single had 21% longer waiting time to start of definitive treatment compared with those who were married.
Furthermore, those who were Ewe and Ga had shorter waiting time to start of definitive treatment (incident rate ratio [IRR] = 0.38 and 0.41, respectively) compared with those who were Akan. With respect to the disease stage, those with stage I and II had significantly shorter waiting time (IRR = 0.588, p < .001) than those with stage III and IV. Also, breast cancer patients recruited from the oncology site had significantly longer waiting time than those recruited from the surgery site (IRR = 1.231, p < .05).
In addition, a unit increase in the time between when biopsy was requested and when results were received increases the waiting time in the start of definitive treatment by 2.2% (IRR = 1.022, p < .001). Also, the waiting time for those who received adequate information from the health workers was 61.1% shorter than that of those without adequate information from the health workers.
Discussion
In this study, we determined the waiting time (in weeks) between presentation and start of definitive treatment and the factors associated with it among patients with diagnosis of breast cancer. To the best of our knowledge, this is the first study in Ghana to address this question. Our study showed that the median waiting time from presentation to start of definitive treatment was 5.0 weeks. This waiting time is much higher than the median time of 15 days in Germany (Arndt et al., 2003) and 21.5 days in Kenya (Otieno, Micheni, Kimende, & Mutai, 2010). However, it is lower than 11.1 weeks reported by Jassem et al. (2013) for 12 countries in low- and middle-income countries excluding Africa. In the United Kingdom and in Canada, standard waiting times are recommended (Canadian Society for Surgical Oncology, 2013; Department of Health, 2011), but there is no national recommended standard waiting time currently in Ghana. This makes it very difficult to audit whether breast cancer patients are receiving timely treatment which can affect their survival (Reed et al., 2004)
Our study also showed that close to 50% of the patients started treatment by 4 weeks; this is higher than the one observed in Nigeria (17.0%; Ezeome, 2010). It is however lower than that of Kenya and Germany in which 59.4% and 73.0% of patients, respectively, started treatment within 4 weeks (Arndt et al., 2002; Otieno et al., 2010). Furthermore, our study showed that 29.3% of the patients started treatment between 1 and 3 months as compared with Nigeria (10.6%), Kenya (24.3%), and Germany (16.0%; Arndt et al., 2002; Ezeome, 2010; Otieno et al., 2010). A little less than a quarter (22.8%) waited for more than 3 months to start treatment in our study. This is higher than that observed in a German study in which only 11% waited for more than 3 months and that in Kenya with 16%, but lower than 72.4% in Nigeria (Arndt et al., 2002; Ezeome, 2010; Otieno et al., 2010). For the Nigerian study, however, the waiting period was from first contact with any health provider and not specifically a health provider within the center where treatment was started.
Our results showed that the median waiting time was significantly associated with age, marital status, level of education, level of income, insurance status, stage of disease, the median time taken for biopsy, and receipt of adequate information from health workers. A German population-based study indicated that 1 out of 6 women (17.4%) aged 18 to 80 years with symptomatic breast cancer waited 3 months or more before first consultation with a doctor. Although older age is a risk factor for delay by patients, younger age is a risk factor for delay by providers (Arndt et al., 2003; Ramirez et al., 1999).
From our study, those who were aged 50 and 59 years and 60 and 69 years had longer waiting time than those who were younger than 40 years. However, there was no significant difference in waiting time between those who were 40 and 49 years and those who were younger than 40 years. There are several potential explanations for this finding. Older adults may have access to less information about breast cancer, or health in general due to being less likely to use the Internet frequently and being less likely to be an active member of the workforce where health-related information may be disseminated. As of 2013, 95% of younger adults go online with a positive correlation between education attainment and Internet use (Fox & Duggan, 2013). Internet usage may positively affect health consciousness among younger women. This is coupled with the fact that, when retired or nearing retirement, there are a lot of social responsibilities and engagements that tend to be financially demanding and time-consuming; hence, not much emphasis may be placed on health. This increased waiting time to treatment in older patients was also found in a study in Louisiana on delays in breast cancer treatment (Williams, Tortu, & Thomson, 2000).
Marital status from our study was a significant predictor of waiting time from presentation to start of definitive treatment. This may be explained by spousal support which has been proven to significantly influence response to treatment. Aizer and colleagues in a study in Boston, United States of America found that married patients with cancer did significantly better than single patients—they lived longer, received better treatment, and were more likely to be diagnosed earlier (Aizer et al., 2013). Furthermore, our study showed that having no income was significantly associated with a higher median waiting time to treatment. This is similar to findings from a systematic review carried out on barriers to breast cancer care in developing countries (Sharma, Costas, Shulman, & Meara, 2012). Having no or low income places financial barriers on patients which may cause delays in examinations and diagnoses which may result in a delay in the time to treatment (Maly et al., 2011).
Our study further showed an association between education and the median waiting time and more educated respondents showed shorter waiting time. This may be because highly educated patients are likely to have more health information and hence be more health conscious. This is inconsistent with a study on provider delay in Germany (Arndt et al., 2003), where having higher education was significantly associated with longer provider delay.
Patients who did not have national health insurance and so paid cash for their treatment experienced shorter waiting time. This may be because of the protracted and bureaucratic nature of the processes involved in processing one’s National Health Insurance card in hospital. Anecdotal evidence has also shown that patients with National Health Insurance are given appointments at longer intervals because of their very large numbers. In Ghana, three types of insurance schemes exist. These are the district-wide National Health Insurance, the private Mutual Health Insurance, and Private Commercial Health Insurance. The district-wide National Health Insurance scheme is the most widely used as it is supported financially by the government to ensure access to basic health for all Ghanaians. Comparing our results with those from other countries may reflect differences in the functionality between health insurance systems which vary widely by country. For example, in a German study there was no significant difference in waiting time between patients who had health insurance and those without it (Arndt et al., 2003).
Majority of the patients in our study presented with stage III/IV disease just like in a study in India (Saxena et al., 2005). This finding is contrary to a study in the United States in which majority of the patients presented with stage I/II disease (Maly et al., 2007). The stage of disease at presentation is significantly associated with waiting time to start of treatment. In our study, patients with late stages of breast cancer at presentation experienced longer waiting time. This may be because in advanced stages of breast cancer, more extensive and expensive investigations such as CT scans and bone scans may be required to accurately stage the disease. This may contribute to the longer waiting time to treatment. In a study in Louisiana, USA, patients with significantly smaller tumor sizes experienced diagnostic delays compared with those who did not experience delays. Conversely, older women with larger tumor sizes experienced delay in start of treatment compared with those with smaller tumors which is similar to what we also found (Williams et al., 2000). Also in a German study, a complex association was found between delay and breast cancer stage where the highest proportion of metastasized breast cancer was found among women with either very short or very long delay (Arndt et al., 2002).
With respect to health care worker factors, almost all the patients stated that health care workers were approachable, willing to direct patients, readily available, and gave adequate information about their condition. However, these factors were not significantly associated with median waiting time to treatment. The patients who said they had adequate information from the health workers had shorter waiting time than those who had inadequate information. This is similar to what was found in the United States where patients with greater self-efficacy in interacting with health care providers had significantly shorter delays in treatment than patients who self-detected their tumors (Maly et al., 2011). For health care system factors, the time in getting biopsy results correlated positively with median time to the start of treatment. This may be due to the constraints involved in getting a biopsy performed and read, which may be either patient, health care worker, or health care system related. This kind of delay was also found in Kenya (Otieno et al., 2010) where the median time to getting a biopsy reported from the time of request was 17 days.
Our study showed that only about one fifth of patients were of the opinion that there was a delay in their breast cancer treatment. The fact that almost 80% of the patients were satisfied with this waiting time suggests that they do not recognize the need for reduced waiting time for treatment of breast cancer and this may in itself add to the delay. Hence, patients should be educated on the need for early detection and treatment of breast cancer to improve survival. This could contribute to reducing provider delay caused by patient factors, as was alluded to in a study done in Kaduna, Nigeria (Ukwenya, Yusufu, Nmadu, Garba, & Ahmed, 2008).
Study Limitations
Our study had a few limitations: First, it was a hospital study; thus, there may be selection bias of patients and may not be generalizable to the whole of Ghana, especially primary care practices and district hospitals. Second, being a retrospective study, the ability of the respondents to accurately recall the events may have introduced some bias.
Conclusion
The median waiting time from presentation to start of definitive treatment was 5 weeks. This was significantly associated with age, level of education, marital status, level of income, insurance status, stage of disease, the median time taken for biopsy, and receipt of adequate information from health workers. Although this time interval is similar to those in low-income countries, it is relatively longer than those noted for patients in high-income countries. The processes involved in obtaining a biopsy should therefore be clearly mapped out and investigated to determine the cause of the delay so as to take the necessary steps to reduce this time interval. Longer total delays of 3 to 6 months from detection of symptoms to start of definitive treatment have been found to be associated with worse survival and hence the need to work on reducing these waiting times especially as the patient delay is also quite prolonged among our patients. Education of both providers and patients is needed to help reduce the total waiting time.
Furthermore, a nationwide campaign on breast cancer awareness may encourage early presentation and treatment of breast cancer. Also, the bureaucracies involved with processing patients on the health insurance scheme may need to be looked at again to reduce time to treatment to improve prognosis. More research needs to be carried out to explore what the factors are which account for the delay in receiving biopsy results so effective interventions may be carried out to reduce the time.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors declared receipt of the following financial support for the research, authorship, and/or publication of this article: The authors received funding from the Fogarty International center grant number 5D43TW009140 through the Cardiovascular Research Training Institute.