Return on Investment of a Breast Cancer Screening Program in Tanzania: Opportunity for Patient and Public Education

Abstract

Background:

Breast cancer is the second incident and second cause of cancer mortality among women in Tanzania due to late-stage presentation. The screening clinic at the Ocean Road Cancer Institute (ORCI), can help detect cases early and reduce cost of treatment. We calculated the return on investment (ROI) of the ORCI breast screening clinic.

Methods:

Screening and treatment data of all newly-diagnosed breast cancer patients seen at ORCI during 2016–2018 were abstracted from the medical records. Also, data on time, resources, and cost of screening and treatment were obtained. The cost of treating screened patients was compared to cost of treating unscreened patients and differences in cost of treatment were compared to cost of operating the screening program.

Results:

Of 730 total patients, 58 were screened prior to treatment and 672 were not. There was no significant difference between stage at diagnosis and treatments received by screened and unscreened patients (79.3% late- stage vs 72.2% late-stage diagnosis, respectively (p=.531), or cost of treatment between the two groups (cost, in Tanzanian Shillings, for screened (2,167,155.14 or $954.27) vs unscreened (1,918,592.28 or $844.52), (p=.355). There was also no significant difference in cost of treatment between the screened and unscreened groups and a slightly negative ROI (−0.05%) from implementing the program.

Discussion:

The breast screening clinic in Tanzania has not yet proven its cost-effectiveness in reducing stage with screening. The likelihood that patients have utilized the clinic for treatment rather than early detection is a possible reason for the lack of cost-effectiveness. Future studies should focus on educational initiatives to encourage screening at early disease stage. Public education should increase awareness about the clinic for early detection. The experience of this program is ideal for dissemination to other low-income countries that are initiating cancer early detection and cancer education programs.

Introduction:

According to the Global Burden of Disease Study, breast cancer is the most common cancer among women.¹ Data from cancer registries around the world indicated that in 2015 there were 2.4 million new cases of breast cancer and 523,000 deaths resulting from the disease_, including men and women.¹ Most cases of breast cancer occur in women, with male breast cancer rates being estimated at less than 1 per 100,000 man-years. This is further highlighted by a calculated an incident rate ratio of 122 between women and men.¹ Though the incidence rate of breast cancer is higher in high-income countries than middle or low-income countries, this has not correlated with lower mortality rates for those lower and middle-income countries.² In Tanzania, breast cancer is the second most common cancer among women after cervical cancer and the second leading cause of cancer mortality.³ Furthermore, recent research has predicted a 82% increase in breast cancer incidence in Tanzania by 2030.³ This is concerning because the survival rate of breast cancer is less than 45%⁴ that is attributed to late-stage diagnosis.⁴ Patients diagnosed with breast cancer at earlier stages generally have lower mortality rates than patients diagnosed at later stages.⁵ Moreover, delays in or lack of access to treatment leads to lower probability of survival, greater morbidity, and more expensive treatment.⁶ A study of breast cancer patients during the period of 2007–2009 at the Ocean Road Cancer Institute (ORCI) in Tanzania showed that 90.7% of the patients presented with late stages (III and IV).⁷ In Egypt, a study revealed that late stage patients were less likely to experience painful symptoms, indicating that even if patients have symptoms they may not seek treatment or screening unless prompted by painful symptoms.⁸ This is troubling given the increased likelihood of adverse health outcomes associated with late-stage diagnosis.

In addition to the negative effects on mortality, advanced-stage diagnosis can be more expensive. Treating late-stage breast cancer is associated with significant incremental costs compared to treatment of those with early-stage breast cancer.⁹ Studies have shown that screening for breast cancer in women over 50 is cost-effective in reducing mortality.^10,11 However, those findings were determined in the United States and it is unclear whether that holds true for low- and middle-income countries, such as Tanzania.

This assumption is further brought into question by a study examining the cost of treatment of cervical cancer at Ocean Road Cancer Institute (ORCI), the only specialized cancer treatment center in Tanzania, which found that treatment for early stages was more expensive, as late stage patients only received palliative care for short periods.¹² Because most breast cancer patients at ORCI are diagnosed at advanced stages⁷ and because of the success of the cervical early detection clinic of ORCI,¹³ a breast cancer early detection clinic has been established at ORCI. The program started alongside the ORCI cervical screening program in 2002 but on a much smaller scale until it became well-established with a clinic in 2014. Therefore, the aim of this study was to examine the effect of the screening clinic at ORCI to determine whether it has led to downstaging of breast cancer and/or lowered the cost of treatment. The aim of the study was to determine the return on investment, from the Tanzanian government standpoint as the payer, within the scope of the treatment provided at this location.

Methods:

Study Setting:

This study was conducted at the Ocean Road Cancer Institute (ORCI), the main cancer center in Tanzania. ORCI is located in the capital city of Dar es Salaam and is funded and operated by the Tanzanian government through the Tanzania Ministry of Health. Since 1996, ORCI has been Tanzania’s main cancer center for treating patients with chemotherapy and radiotherapy. ORCI’s budget pays for treatment for patients who cannot afford to pay, once their diagnosis is confirmed. All diagnostic testing prior to diagnosis is the financial responsibility of the patients.

Data Collection:

This study used data of newly-diagnosed breast cancer patients treated at ORCI from 2016 through 2018. These years were utilized because they were the most recent years with the best available data as ORCI is in the process of digitizing paper medical records to an EMR system. Information regarding patients’ screening status, demographic information including age, gender, marital status, religion, and occupation, as well as patients’ stage at diagnosis, region of residency, and detailed course of treatment were extracted from a combination of the paper and electronic medical records of breast cancer patients. The data was de-identified before the aggregate analysis. The breast screening clinic at ORCI receives two main types of women, those who are referred to the breast screening because they had symptoms that required examination and those who walked into the clinic asymptomatically. However, ORCI does not distinguish between these groups in their screening records. There were 730 patients diagnosed with breast cancer at ORCI from 2016–2018 for whom their screening status was available. Twenty-eight patients’ screening status were unavailable and therefore excluded from the cohort. Additionally, 30 patient files were missing or unavailable and were therefore excluded from this study. Of the cohort, only 628 patients had initial stage diagnosis data available.

Physicians and nurses working both in the ORCI treatment and screening clinics were interviewed qualitatively to obtain a full perspective on all the resources used and procedures followed in the course of screening and treatment of breast cancer, revealing an administrative fee for each treatment. Once the itemized lists of the various treatment methods, medications, and other administrative resources used for patient care was compiled, five accountants were interviewed to obtain the cost of these resources and procedures. Price lists and annual budgets were also acquired to precisely compute the cost of chemotherapy and radiation therapy as well as the direct and indirect costs of operating the screening clinic. The costs of the screening clinic were categorized as advertising and education, which comprises all the brochures, radio and television advertisements and other promotions offered by and at ORCI, employee salaries for the 13 employees working in the screening clinic, direct costs of screening, which includes all resources utilized for screening such as gloves or office supplies, all of these variables were extracted from the budgets provided by accountants at ORCI. Additionally, the cost of the screening clinic included estimated overhead costs were derived from the accountant interviews and incorporated the value of the space allocated towards the screening clinic, the estimated cost of the utilities such as water and electricity as well as labor such as maintenance and security. The final cost sheets were reviewed and validated by the executive director of the ORCI.

The cost of treatment was defined as the cost of administering each chemotherapy and radiation treatment to the patient. The costs included individual drug prices, the cost of individual radiation treatments, pre-chemo or pre-radiation drugs, and administrative fees. The total cost of direct treatment was calculated for each patient based on the course of treatment listed in their medical record.

Once the cost of each treatment was calculated, the patients were stratified based on their ever screening status at ORCI during the study period. The average total cost of direct treatment was calculated by dividing the total cost of treatment for those groups by the number of patients in them. Hormone therapy was excluded from this study as specific hormonal therapy based on receptor status was introduced in 2016. Surgeries such as mastectomies and lumpectomies were also excluded from this study as surgery is not offered at ORCI.

Data Management and Statistical Analysis:

The demographic and clinical information as well as the cost of treatment for the screened patients were compared to those who were unscreened. Chi square test was used to examine the association between the stage of breast cancer, age, and duration of symptoms prior to seeking treatment.

Finally, in order to calculate the return on investment the annual cost of investment and net profit were needed. The budgets and interviews were used to determine the annual cost of operating the screening clinic in terms of utilities, resources and space. The net profit was calculated by deducting the average cost of treatment for the screened group from the average cost of treatment for the unscreened group. All calculations were made in IBM SPSS Statistics version 25.

Return on Investment Calculation

The return on investment analysis was completed using the costs of treatment and operating expenses of the breast cancer screening clinic. The formula for ROI was the net profit divided by the cost of the investment, or the operating cost of the clinic (Net Profit/Cost of Investment). Net profit was defined as the average cost of treatment for those in the non-screened group minus the average cost of treatment for the screened group. This calculation was performed in the local currency of Tanzanian Shillings and then converted to 2018 value of US dollars.

Chi square analyses were run to test the association between various factors, such as number of treatments and duration of symptoms, and the screening status of the patient. An independent sample t-test was run to test the significance of the difference between the cost of treatment of both the screened an unscreened groups. Finally, regression analyses were run to test the sensitivity of the results where the screening group was the independent variable and dependent variables were the cost of treatment and the duration of patients’ symptoms prior to seeking treatment.

Results:

Cohort Demographics

A summary of patient demographics is displayed below in Table 1, which shows that most patients presented with late stage breast cancer. A majority of the patients were of Christian faith, followed closely by Islam. Most of the patients were between 35–64 years old, worked in agriculture, and were married, as shown in Table 1. The age distribution for the screened and unscreened groups were very similar as displayed in Figure 1. There was an association between screening status and the occupation of the patients, if they were in agriculture, business or economically-dependent, meaning dependent on others in their household to provide, (among those who were screened 24.1% worked in business, 31% worked in agriculture and 29.3% were economically dependent, compared to the unscreened group where the occupation percentages of the three occupations were 10.6%, 49.1% and 17.1% respectively, p=.001, p=.005 and p=.38, respectively). The distribution of time between patients’ symptoms prior to their visit to ORCI and the stage of their cancer is illustrated in Table 2, showing that most patients whose symptom duration was available waited between 3–6 or 9–12 months before coming to ORCI for treatment.

Table 1.

Characteristics of 730 Breast Cancer Patients Treated at the Ocean Road Cancer Institute during the period of 2016–2018.

	Screening Status		P
	Screened	Not Screened
Total Number of Patients	58	672
Average Age	50.9	52.6	0.364
Percentage Female	100%	97%	0.183
Percentage Married	51.7%	50.4%	0.643
Percentage Christian	48.3%	55.3%	0.304
Percentage Muslim	48.3%	42.3%	0.377
Percentage Working in Business	24.1%	10.6%	0.005*
Percentage Working in Agriculture	31%	49.1%	0.001*
Percentage Economically Dependent	29.3%	17.1%	0.038*
Valid Percentage of Early Stage Cases	12.1%	13.4%	0.420
Valid Percentage of Late Stage Cases	79.3%	72.2%	0.420
Residing in Dar es Salaam	43.1%	26.2%	0.006*
Average Number of Chemo Treatments Received	7	6.5	0.484
Average Number of Radiation Treatments Received	21.4	22	0.295
Average Total Cost of Treatment	2,167,155.14 TSH or $954.27 USD	1,918,592.28 TSH or $844.82 USD	0.415

Figure 1.

Age Distribution by Screening Status

Table 2.

Duration of Symptoms Prior to Visit by Screening Status with Percentages

Duration of Symptoms	Screened Early (% Screened)	Not screened Early (% Unscreened)	Screened Late (% Screened)	Not screened Late (% Unscreened)	Sig
Missing	0 (0%)	26 (3.9%)	9 (15.5%)	87 (12.9%)	.174
0–3 Months	3 (5.2%)	17 (2.5%)	8 (13.8%)	69 (10.3%)	.572
3–6 Months	3 (5.2%)	19 (2.8%)	10 (17.2%)	97 (14.4%)	.607
6–9 Months	0 (%)	8 (1.2%)	4 (6.9%)	56 (8.3%)	.590
9–12 Months	0 (%)	10 (71.5%)	12 (20.7%)	101 (15%)	.349
12+ Months	1 (1.7%)	10 (1.5%)	3 (5.2%)	75 (11.2%)	.234

Associations with Screening Status

Additionally, there was no significant association between the time elapsed between diagnosis and receiving the first round of treatment and the screening status of the patients. The region of patients’ residence was significantly different between both groups, with those from Dar es Salaam more likely to be screened with 43.1% of the screened group residing in Dar es Salaam compared to 26.2% of the unscreened group (p=.006). There was no significant difference between the stage at diagnosis and the screening status with 72.2% of the unscreened group and 79.3% of the screened group being diagnosed at late stages, respectively (p=.420).

There was no significant difference between the average number of chemo treatment cycles received by those in the screening group (7 cycles), compared to those who were not screened (6.5 cycles) (p=.531).

Cost of Treatment and Screening

The average cost of treatment for those who were screened amounted to 2,167,155.14 Tanzanian Shilling (TSH) compared to the cost of those not screened at the clinic of 1,918,592.28 TSH. In 2018 US dollars that equates to $954.27 compared to $844.82, which was not statistically significantly different (p=.355). Figure 2, shows the cost of screening in millions of TSH based on screening status, and how the cost distribution was very similar across screening status.

Figure 2.

Total Cost of Treatment by Screening Status

The annual cost of operating the screening clinic was extracted from the annual budgets and reviewed and confirmed by the executive director of ORCI. The total amounted to 496,539,362 TSH or $218,643.49 2018 US dollars. A summary of the cost components is shown in Table 3. A majority of the cost of the screening clinic is attributed towards the salaries of clinic employees.

Table 3.

Annual Cost of Operating Screening Clinic at the Ocean Road Cancer Institute

Cost Category	Annual Screening Clinic Cost
Advertising and Education Materials	45,907,749 TSH
Direct Screening Cost	29,079,613 TSH
Budgeted Salaries for Clinic Employees	367,512,000 TSH
Estimated Clinic Overhead	54,040,000 TSH
Total Annual Clinic Costs	496,539,362 TSH or $218,643.49
Average Cost of Screening Per Person	125,961.28 or $55.47 USD

$$\begin{gathered} \frac{\mathbf{Average}\mathbf{Cost}\mathbf{of}\mathbf{Treatment}\mathbf{for}\mathbf{Unscreened-Average}\mathbf{Cost}\mathbf{of}\mathbf{Treatment}\mathbf{for}\mathbf{Screened}}{\mathbf{Annual}\mathbf{Cost}\mathbf{of}\mathbf{Operating}\mathbf{Screening}\mathbf{Clinic}} \\ \frac{(\mathbf{1},\mathbf{918},\mathbf{592}.\mathbf{28}-\mathbf{2},\mathbf{167},\mathbf{155}.\mathbf{14})}{\mathbf{496},\mathbf{539},\mathbf{362}} \\ \mathbf{ROI}=-\mathbf{0}.\mathbf{05}\mathbf{\%} \end{gathered}$$

The results show a slightly negative but miniscule return on investment of the breast cancer screening clinic.

Regression models showed that the variance in the cost is not explained by the screening status (adjusted R squared=0.00) or the duration of symptoms (adjusted R squared=0.002).

Discussion:

Although this study found a slightly negative though insignificant ROI, we believe the results underline the need to enhance breast cancer education in Tanzania. This study revealed three key findings to bolster this observation. First, the direct cost of screening breast cancer was not high, and there was no significant difference in cost in treatment based on screening status, likely due to the late stage at which patients were screened. Second, there was no significant difference in the time elapsed between onset of symptoms to the time of seeking treatment with or without screening. Third, the screening program is more utilized by residents of Dar es Salaam than those living outside the city.

The low cost of screening at ORCI including all overhead ($55.47) is less than the United States-regulated Medicare expenditure for breast cancer screening for women aged 66–74, totaling $84¹⁴ or the cost of screening per person in an Egyptian screening clinic amounting $112.10.¹⁵ This may be due to ORCI’s utilization of clinical breast exam instead of mammography which is both more expensive and common in developed countries. Clinical breast exams have been shown to be effective in downstaging breast cancer in low- and middle-income countries in Asia and Africa.¹⁶ Despite the low cost of screening, our study showed a slightly negative ROI and the cost of treatment was similar across the screened and the un-screened groups. However, even if the program does not save money in terms of cost of treatment, it can improve quality of life and eventually lead to downstaging of cancer and increased longevity. A cost-effectiveness study of ORCI’s cervical cancer screening program found it cost-effective when considering life years gained from treatment.¹² The cost-effectiveness of screening holds true for breast cancer as shown by in a study from Ghana that found clinical breast exams and mass media awareness were both cost effective using DALYs.¹⁷ An earlier study in India found that clinical breast exams were cost-effective¹⁸, and another study found that a breast cancer screening program in Egypt succeeded in yielding a positive ROI.¹⁵ This study highlights both the importance of screening and education and their benefits.

The similarity between the age distribution and the duration of enduring symptoms prior to seeking treatment with or without screening shows opportunities for patient and public education. This study’s showed that the period between the initial appearance of symptoms and healthcare seeking behavior is consistent with other studies of Sub Saharan Africa, which showed an average gap of 11.2 months for breast cancer patients to seek treatment in Nigeria.¹⁹ This finding shows an opportunity for improvement to encourage more women to recognize symptoms and seek screening or treatment early as soon as symptoms are recognized. The need for improved education within Tanzania is evident through the results of a recent study of a rural area outside of Dar es Salaam that showed lack of information as a barrier to screening.²⁰ Patient and public education for cervical cancer early detection in Tanzania resulted in downstaging¹³ and this experience could be replicated to downstage breast cancer.

Finally, the higher utilization of the screening program by women from Dar es Salaam is consistent with other studies. A 2010 study of cervical cancer patients at ORCI showed that screening clinic patients were more likely to be from Dar es Salaam and that the cost of travel and diagnostic testing proved to be a barrier for patients.²¹ Similarly, a study in South Africa reported that those living further away from a medical center were more likely to be diagnosed with more advanced tumors.¹⁹ Our study is in agreement with those studies that show the cost of transportation and diagnostic testing serves as a barrier to screening.²² This underlines the importance of making screening more accessible for those who live in other areas of the country.

Increased education and screening outreach are crucial to improving outcomes. A study from Sudan where local volunteers screened for breast cancer in rural communities was successful at detecting breast cancer.²³ Another study from 2009–2011 showed that local screening programs in Tanzanian villages were able to detect and downstage breast and cervical cancer, eventually referring patients for treatment at ORCI.²⁴ These study results highlight the importance of localizing screening and education that should be continued and expanded with ORCI’s screening program.

Screening clinics in developing countries, and most likely in Tanzania as showed in this study, see patients who are symptomatic and may seek the screening clinics because of their symptoms and to potentially avoid the long wait time to receive treatment³. In fact, a study in Canada found that those participating in screening programs did have a higher likelihood of being treated sooner.²⁵

ORCI currently has initiated a mobile screening program which ventures to more rural areas of the country a few times per year to provide screening and education to those unable to travel to Dar es Salaam. This program is important in increasing access to screening services, educating the public, and encouraging screening of asymptomatic women.

There are several strengths to our study. Since ORCI is the only cancer center for chemotherapy and radiotherapy in Tanzania and patients travel from all over the country for treatment there, the patient sample is likely representative of the country as a whole. Furthermore, the information regarding the cost of treatment and operating the screening clinic was obtained from formal government budgets since ORCI is run by the Tanzanian government, and as such is a complete and accurate depiction of the cost of treatment. Moreover, there is not a plethora of research examining the economics of screening programs in sub-Saharan Africa, so this contributes towards building a body of information to explore screening programs in developing countries in this region. Finally, the results reveal a potential area for improvement by increasing education and screening outreach to reach populations outside Dar es Salaam and lower the time between the emergence of symptoms and health seeking behavior. Progress in this area might lead to decrease mortality from breast cancer.

There are a few limitations to this study. First, the cost of hormone therapy received at ORCI was not included in the cost of treatment because of its recent introduction. Also, surgical treatment received at other hospitals was not included in our analysis. However, stage at diagnosis was not statistically different between screened and unscreened patients and therefore the cost between the two groups are not expected to be different from a surgical treatment standpoint.

The results of the ROI analysis show that there were no cost savings due to screening, and in fact there was a very slight increase in the average cost of treatment for those screened. These results, however, should not be interpreted to mean that the screening clinic should be closed. Rather, these results highlight that not enough women are being screened at an early enough stage with either minor or no symptoms. It is this population that truly benefits from screening clinics and will contribute to a positive return on investment. It is most likely that the screening clinic operates as the first step of treatment rather than a true screening clinic in order to jumpstart their treatment process. More time and resources should be dedicated to increasing public and patient education of the Tanzanian population to underline the importance of screening.

In summary, this study showed that the breast screening clinic in Tanzania has not yet proven its cost-effectiveness. The likelihood that patients are utilizing the screening clinic for diagnosis rather than early detection is the possible reason for the lack of cost-effectiveness. Value per Statistical Life Year would be a valuable perspective to measure the impact of the screening clinic. Unfortunately we do not have the requisite survival information to perform such a calculation but such important analysis should be performed once the necessary survival data is collected and available. Future studies should also focus on patient education to seek screening at early stage of the disease. Public health education should also increase awareness about the availability of the clinic for early detection and better outcomes of breast cancer management. This experience of this screening program is ideal for dissemination to other low-income countries that are on the path to initiating cancer early detection and cancer education programs.