Funding of Oncology Benefits by Medical Schemes, South Africa: A Focus on Breast and Cervical Cancer

Abstract

Breast and cervical cancers are among the top five worldwide. The mortality rate for breast cancer is over 50%, when compared to cervical cancer, which is nearly 90%. Early breast and cervical cancer screening can reduce mortality risk. This study examined breast and cervical cancer rates among South African medical scheme members. The study’s secondary goal was to analyse how medical schemes funded these two cancers, including patient and/or out-of-pocket payments, to identify funding gaps. The study was a cross-sectional retrospective review of medical scheme claims data for oncology benefits, especially for breast and cervical cancers. The study used a multivariate logistic regression model to assess cancer rates. The results showed that the relative proportion of beneficiaries with breast cancer was higher in open schemes than restricted, in large schemes than medium and small schemes, in comprehensive plans, efficiency discount options (EDOs), hospital plans than in partial cover plans, in age groups older than 55, in an out-of-hospital setting than in in-hospital setting. The paper advises examining the funding mechanism of oncology benefits to reduce out-of-pocket payments (OOPs) for cancer patients, revising network arrangements, and using designated service provider (DSP) as a barrier to access against uneven oncology provider distribution.

1. Introduction

Cervical and breast cancers threaten the lives of many women, accounting for two million newly diagnosed cases and 800,000 cancer-related deaths annually [1]. The incidence rate of cervical and breast cancer is much greater in low- and middle-income countries (LMICs) than in industrialised nations, where screening facilities and other preventive treatments are readily available [2, 3, 4, 5, 6]. Compared to low-income countries, high-income countries have significantly raised screening rates [7, 8]. In the World Health Organization Regional Office for Africa WHO (AFRO) African area, screening rates are still relatively low [7, 8]. In addition, 90% of cervical cancer deaths occur in women residing in low- and middle-income countries (LMICs), with sub-Saharan Africa bearing the heaviest burden [6, 9]. In South Africa, cervical and breast screening uptake disparities are caused by a lack of education and access to information. A study in Korea also showed that breast and cervical cancer screening rates varied widely among women with higher household incomes and education levels [1]. Other variables, such as limited access to health care facilities, lack of knowledge and health promotion, and inadequate support and awareness programmes for women, which are more prevalent in less affluent rural areas, influence early detection, diagnosis, and screening uptake [10, 11]. Recent research has shown that metropolitan location remains a significant factor related to greater awareness of cervical cancer risk factors, resulting in lower screening rates in rural regions [11, 12, 13]. In addition, lack of access to facilities in rural locations adds to inadequate screening services, a barrier to an admission that results in delayed seeking behaviour. Studies have demonstrated that early screening helps in detecting cervical cancer early [14]. The availability and accessibility of early screening and treatment services contribute to the notable disparities in cervical and breast cancer incidence. Early detection through screening and surveillance should be linked to available resources for treatment, such as access to providers and facilities [15, 16]. The author further warns that careful consideration must be paid to ensure that all aspects of cancer control programmes are balanced to limit unintended harm [15].

Other factors linked to health systems are human resources challenges, the scarcity of specialist services, and the uneven distribution of specialists by sector. Studies have shown that most specialists for oncology-related services are concentrated in the private sector rather than in the public sector [17]. In South Africa, nearly 80% of oncology specialists (radiation oncologists) practice in the private sector, are mainly remunerated through the funding model of medical schemes (health insurance providers), and cater for 16% of the population [18, 19]. The remaining 20% of the specialists cater to 84% of the population. Figure 1 further illustrates inequities in the distribution of oncology service providers and how they have evolved over the last 3 years, showing a consistent trend toward an increasing share in the private sector.

Figure 1.

Distribution of radiation oncologists by sector-South Africa. Source: Adapted from Abratt [17].

2. Breast cancer

Breast cancer is the most common cancer among women and the top cause of death in more than 100 countries worldwide [20, 21, 22]. Over 2 million new breast cancer cases have been detected worldwide, representing a quarter of all cancer cases in women [20, 23, 24]. The incidence of breast cancer in the WHO’s African region (AFRO) accounts for 27.7 percent of all cancer cases, and there are 29,593 breast cancer patients in South Africa [20, 22, 25]. The high incidence rates in affluent and developing nations are driven by non-genetic risk factors associated with menstruation, reproduction, exogenous hormone replacement, alcohol consumption, and weight gain [20, 26, 27, 28]. Age and gender are risk factors associated with breast cancer; the incidence of breast cancer increases and is more prevalent in women aged 50 or older than in men. Madeira et al. [29] stated that men account for 1% of all breast cancer cases [30, 31]. The average age of a breast cancer diagnosis varies according to gender. In a systematic analysis of 1,201 male breast cancer patients from 27 African nations, Ndom et al. [32] showed that the average age for men was 54.6 years, and for women, it was 47.7 years. Baudouin Kongolo Kakudji et al. [33] created an epidemiological, clinical, and diagnostic profile of breast cancer patients treated at the regional hospital in Potchefstroom, South Africa. The study’s majority of patients (98.6 percent) were female, with a mean age of 56.2 years (standard deviation: 14.4) (95 percent confidence interval: (54.6–59.7) [33]. The average cost of breast cancer therapy varies by location or geographic area; sector, setting (in and out of hospital), treatment modality and level of care, disease severity, and disease stage [34]. The expected cost of chemotherapy in the South African public sector is R15,740 [34]. The average cost of medical treatment for breast cancer in each episode in South Africa is imprecise, and the available data contain methodological flaws. Finestone et al. [27] demonstrate that the average cost of breast cancer in the private sector was more than three times that of the public sector. During the first year after a breast cancer diagnosis, Discovery Health Medical Scheme (DHMS) estimated the average cost to treat breast cancer to be in the region of R207 561 [35]. Finestone et al. [27] estimate that the cost varies with stage, with the estimated cost for stage 1 in the private sector being R352 495 and the projected cost for stage 4 being significantly higher at R522 553. Indicating a variation between health plans or benefit options, the author confirmed that the cost was considerably lower for low-cost benefit alternatives [27]. The costs vary based on diagnosis and therapy, as shown in Table 1.

Diagnosis	Treatment
Mammogram - R1 800 Ultrasound - R1 400 MRI scans and CT scans - R3 000 – R15 000 Stereotactic biopsy - R4 000 to R6 000 Oncotype DX (one type of genetic testing) - R30 000	Surgical; Lumpectomy - R30 000 Mastectomy of both breasts with reconstruction - R150 000 – R250 000. Non-surgical; Hormone treatments such as Tamoxifen can be reasonable at R100 a month. A drug such as Herceptin can cost around R20 000 for one dose, which can be prescribed monthly. A session of radiation therapy and/or chemotherapy could cost from a few thousand to over R10 000

Table 1.

Estimated cost of medical treatment of breast cancer (Diagnosis and Treatment).

Source: Adapted from 1Life [36]: data excludes additional and aftercare: ongoing follow-up doctors’ visits and health checks and tests

3. Cervical cancer

In 2020, cervical cancer was the fourth most prevalent cancer in women worldwide, with a projected incidence of more than 600,000 cases and over 350,000 deaths [21, 22]. Cervical cancer incidence and mortality rates are the greatest in Africa, notably in Southern, Eastern, and Western Africa [25]. Cervical cancer is the most common in many sub-Saharan African countries (22 percent of all cancers), and its prevalence in poor developing countries is potentially 15 times higher. After breast cancer, cervical cancer accounts for 18.7 percent of all cancer incidences in South Africa’s female population [22, 27]. The incidence of cervical cancer increases with age, with the average age of diagnosis ranging from 35 to 44. The average age of cervical cancer diagnosis in South Africa is 45 [37]. Due to the greater prevalence of cervical cancer in women with HIV, which is more prevalent in younger women, it is suggested that younger women get cervical cancer screening at an early age. Cervical cancer is diagnosed at an average age of 50 to 53 years [38]. A study conducted in South Africa revealed a significantly lower average age of 40.8 years (SD 18.6, range 15–95 years); nevertheless, the analysis concentrated on 5,903 females (15–49 years) [39]. Similarly to breast cancer, the average cost of medical therapy for cervical cancer varies by sector, setting (in and out of hospital), treatment modality, level of care, disease severity, and disease stage [37, 39]. In the private sector, the average cost of cervical cancer was 9 times greater for stage 1 and 13 times greater for stage 4 [27]. The limitations of the study by Finestone et al. [27] were that it contrasted one medical scheme to the public sector, excluding closed schemes like the Government Employees Medical Scheme (GEMS), which primarily serves public sector personnel; other medical schemes were also excluded in the analysis. Similarly, non-included plans may have a specific risk and age profile.

4. Legislative requirements

The level of care for breast and cervical (oncology) services (Table 2) is outlined in the Council of Medical Schemes Act’s Prescribed Minimum Benefit (PMB) list [40]. PMBs are defined benefits designed to ensure that all members of medical plans have access to some fundamental health care, regardless of the benefit option selected. The PMB list consists of 25 Chronic Disease List (CDL) diseases and an additional 271 Diagnosis and Treatment Pair (DTP) conditions. Regarding breast and cervical cancer services, medical schemes must cover the diagnosis, treatment, and care for these disorders. However, medical schemes are not required to pay for diagnostic tests to determine that a patient does not have a PMB illness. Cervical cancer screening is a PMB level of service under DTP Code 960M. The treatment component of breast cancer screening includes PMB-level periodic breast examinations. However, members are entitled to specific screening intervals, instruments, and HPV versus cervical screening. Table 2 demonstrates the optimal amount of care recommended by the regulator [40].

	Cervical cancer	Breast cancer
Screening	Pap smears	Mammography
Diagnosis	Consultations with GPs (1) and specialists (4) for a diagnosis, staging and risk assessment of cancer. Radiology Pathology (Full blood count, liver/renal function, Creatinine, HIV)	Consultations with a nurse or GP (1) and specialists (4) for a cancer diagnosis, staging and risk assessments. Radiology (mammogram, ultrasound) Pathology (Full blood count, liver/renal function)
Staging	Additional pathology where necessary Imaging radiology (ultrasound, X-ray, CT, MRI*, PET*)	Imaging radiology (chest X-ray, liver ultrasound, CT, MRI)
Treatment	Stage 1: Biopsy, Hysterectomy; Stage 2: Hysterectomy, lymphadenectomy; Stage 3: Chemoradiotherapy, chemoradiation; Stage 4: Radiotherapy	Breast-conserving surgery; Systemic therapy; Radiotherapy; Management of advanced disease (hormone therapy, chemotherapy)

Table 2.

Level of care for breast and cervical (oncology) services.

Source: CMS [40]

5. Funding models (benefit design)

Most oncology treatments are provided in-hospital and subject to network hospitals and designated service providers (DSPs), with some entry-level plans using the state as a DSP. Entry-level coverage plans typically cover oncology care at the PMB level. Oncology coverage includes PMBs in their entirety. Schemes provide different cancer treatment limits and extended benefits for more complete benefit options. However, after the annual maximum is reached, patients may be required to pay co-payments for treatment exceeding PMB level care (which costs more than the scheme rate). The oncology benefit does not cover hospital admissions; these are paid by the hospital benefit of the benefit choice the patient is enrolled in. After this hospital benefit has been exhausted, patients may be required to continue treatment at an entry-level network institution or a state facility. Most medical schemes include cancer benefits, annual limits for oncology treatment, and entry-level benefit alternatives for PMB-level care. Comparatively, comprehensive plans cover R400,000 or more [41, 42, 43]. ICON Oncology is the major designated service provider with regard to out-of-hospital benefits [19]. Nonetheless, in-hospital benefits are accessible through the state as a DSP (for some entry-level alternatives) and private institutions (subject to annual hospital limits).

6. Objectives

This study’s primary purpose was to investigate the rates of breast and cervical cancer among South African medical scheme beneficiaries. The secondary purpose of the study was to analyse the funding model of these two types of diagnoses by medical schemes, including measuring the level of pocket payments made by these patients to identify funding gaps, and lastly, to execute a logistic regression model to identify the factors that contribute to the greater exposure rates of the two types of cancer.

7. Methods

The study design was a retrospective cross-sectional investigation of medical schemes’ claims data associated with oncology benefits, primarily breast and cervical cancer. The review period was 2019, and the secondary data came from the annual submissions of aggregated CMS statutory returns data. During the evaluation period, the analysis comprised claims data from 59 medical schemes with comprehensive expenditure data. There were 15 open schemes and 44 closed schemes. Breast cancer and cervical cancer were represented in the study by 47 886 and 4 116 participants, respectively. 97% more women than men had breast cancer. Based on study and calculation, 46,571 female beneficiaries older than 20 were diagnosed with breast cancer. We counted and proportioned categorical variables. Unadjusted comparisons were statistically significant at p value <0.05. The study used a multivariate logistic regression model to examine cancer risk variables. Table 3 shows model dependent and independent variables. The analyses were done in STATA and SAS 9.4.

	Description
Dependent Variable (DV)	Rate of breast and cervical cancer. The rate was higher than a specified cut-off points for each type of cancer: Cancer of the Cervix (0.45 per 1000) beneficiaries and cancer of the breast (2.74 per 1000).
Independent Variables (INDV)	Demographic characteristics Age bands
	A medical scheme is a non-profit organisation with a board of trustees that must be registered with the Council for Medical Schemes. In exchange for a monthly contribution or premium, medical schemes in South Africa provide members with coverage for their medical expenses.; Scheme Type: (Open Schemes, Restricted Schemes). The Medical Schemes Act of 1998 defines open schemes as open membership. As a result, they accept anyone who wants to become a member and pay the premium (Medical Schemes Act 131 of 1998). Closed or restricted medical schemes are limited to an employer or union (Medical Schemes Act 131 of 1998).
	Scheme Size: A large scheme has more than 30,000 beneficiaries. A medium scheme has less than 30,000 beneficiaries and more than more than 6,000. A small scheme has fewer than 6,000 members [44].
	Geographic distribution of beneficiaries in the nine provinces in South Africa: Gauteng, Western Cape, KwaZulu Natal, Eastern Cape, Northern Cape, Limpopo, Free State, North West.
	Benefit design: Benefits options (Health plans) were reclassified into the following categories to assess the effect of benefits option richness [45].; Comprehensive Plans: Provide comprehensive cover for almost all medical costs, including unlimited hospital cover and generous benefits for day-to-day expenses. EDOs: Efficiency Discounted Options (or “EDOs”) offer an appealing value proposition to medical scheme members. Hospital Plans: Supplementary in-hospital benefits relative to PMB; no out-of-hospital (OOH) benefits Partial Cover Plans: Partial cover for OOH benefits from risk, savings account, and no above-threshold benefits (ATB).
	Setting: In- and out-of-hospital benefits
Benefits paid/expenditure	Expenditure reported in South African Currency: 1$= ZAR17
Out-of-pocket (OOP)	The proxy measures are determined as the difference between what the medical service provider claimed and what the medical scheme paid.; Out-of-pocket is the maximum amount that you could pay for covered medical expenses in a year. This amount includes deductibles, copays (co-payments), and coinsurance.

Table 3.

Description of variables of interest.

8. Results

The average age of female beneficiaries with breast cancer was 59 years, whereas the age profile of female industry beneficiaries was substantially younger at 34 years. The breast cancer rate was, therefore, 14 per 1000 female beneficiaries. In the age range of 20 to 24 years, the rates were fewer than 1 per 1000 female beneficiaries, which was significantly lower than those for older age groups. In age groups 65–69, 70–74, 75–79, and 80–84, the breast cancer incidence rate was greater than 30 per 1000 female beneficiaries (Figure 2).

Figure 2.

Number of female beneficiaries vs cancer of breast – treatable beneficiaries.

In addition, the analysed schemes accounted for 4,103 female beneficiaries diagnosed with cervical cancer who were 20 or older. Thus, the rate of breast cancer was 1.25 per 1,000 female beneficiaries. The average age of female beneficiaries diagnosed with breast cancer was 50 years, whereas the age profile of female industry beneficiaries was substantially younger at 34 years. Less than one per thousand female beneficiaries were between 20 and 24 years of age. The cervical’s cancer rate was significantly greater in women aged 45 to 49, exceeding 3.3% per 1000 female beneficiaries (Figure 3).

Figure 3.

Number of female beneficiaries vs cancer of breast – treatable beneficiaries.

Figure 4 shows the distribution of beneficiaries diagnosed with breast cancer and cancer of the cervix. The analysis shows that there were more beneficiaries with breast cancer than cervical cancer in Gauteng (41% vs 33%) and Western Cape (21% vs 9%). There were more beneficiaries with cancer of the cervix than those with cancer of the breast, 17% and 12%, respectively. However, in other provinces, such as KwaZulu Natal, a more notable difference was in Limpopo province (13% vs 5%) and Mpumalanga (8% vs 4%) (Table 4).

Figure 4.

Distribution of beneficiaries diagnosed with breast cancer and cancer of the cervix.

	Cancer of the breast (N = 47,886) n (%)	Cancer of the cervix (N = 4,116) n (%)	% Difference
Scheme type
Open	30 111 (62.88)	1 698 (53.55)	−17%
Restricted	17 775 (37.11)	2 418 (46.44)	20%
Scheme size
Large	44 912 (93.78)	3 919 (89.86)	−4%
Medium	1 978 (4.13)	130 (6.29)	34%
Small	996 (2.07)	67 (3.84)	46%
Benefit design strata			-
Comprehensive Plans	20 913 (43.67)	1 613 (39.18)	−11%
EDOs	3 588 (7.49)	268 (6.511)	−15%
Hospital Plans	11 199 (23.38)	648 (15.74)	−49%
Partial Cover Plans	10 906 (22.77)	1 482 (36.00)	37%
Unknowns/Not classified	1 280 (2.67)	105 (2.55)	−5%
Gender
Female	46 608 (97.33)	4 111 (99.69)	2%
Male	1 278 (2.66 )	5 (0.30)
Age bands
< 20 years	68 (0.14)	9 (0.51)	72%
20–24 years	55 (0.11 )	10 (0.51)	78%
25–29 years	221 (0.46 )	56 (2.56)	82%
30–34 years	1 123 (2.34)	188 (6.14)	62%
35–39 years	2 036 (4.25)	445 (11.16)	62%
40–44 years	3 779 (7.89)	572 (14.74)	46%
45–49 years	4 360 (9.10)	1 138 (14.38)	37%
50–54 years	5 541 (11.57)	543 (12.23)	5%
55–59 years	6 271 (13.09)	355 (9.98)	−31%
6–64 years	6 108 (12.75)	279 (8.34)	−53%
65–69 years	5 973 (12.47)	205 (6.70)	−86%
70–74 years	5 576 (11.64)	146 (5.47)	−113%
75–79 years	3 831 (8.00)	96 (3.89)	−106%
8–84 years	1 960 (4.09)	46 (2.09)	−95%
85 years+	984 (2.054)	28 (1.22)	−67%
Province
Eastern Cape (EC)	2 593 (5.41)	270 (8.96)	40%
Free State (FS)	1 677 (3.50)	152 (6.09)	43%
Gauteng (GP)	19 821 (41.39)	1 344 (32.46)	−28%
KwaZulu Natal (KZN)	5 838 (12.19)	702 (17.46)	30%
Limpopo (LP)	2 195 (4.58)	542 (4.30)	−7%
Mpumalanga (MP)	1 718 (3.58)	340 (6.70)	47%
Northern Cape (NC)	658 (1.37)	55 (2.66)	48%
North West (NW)	3 290 (6.87)	322 (5.73)	−20%
Not classified	205 (0.42)	12 (0.61)	30%
Outside South Africa	11 (0.02)	5 (0.25)	91%
Western Cape (WC)	9 880 (20.63)	372 (14.74)	−40%
Hospital setting
Out of Hospital (OOH)	32 891 (68.68)	2 109 (48.23)	−42%
In Hospital (IH)	14 995 (31.31)	2 007 (51.76 )	40%

Table 4.

Demographic characteristics, number of beneficiaries (%).

8.1 Distribution of oncology specialists

Independent practice specialists’ radio oncology services are available in all provinces. Figures 5 and 6 show the distribution of oncology specialists by region. The results depict that many oncology specialists are concentrated in urban and more affluent areas such as Gauteng and KwaZulu Natal, as the proportion of providers was higher than that of beneficiaries within the province. Medical oncology services are based mainly in Gauteng province, which accounts for 92% of medical oncology services. At the same time, there are no medical oncology services in two regions, Western Cape and KwaZulu Natal, where each accounted for 4%. When adjusting for beneficiaries, there are significantly more independent practice specialists in radiation oncology than beneficiaries in Gauteng province. Slightly more beneficiaries than providers in KwaZulu Natal, Western Cape and the Free State province. There is a high scarcity of independent practice specialists in radiation oncology in Limpopo province, where the relative ratios were 1% vs 5%.

Figure 5.

Distribution of oncology specialists by province.

Figure 6.

Distribution of independent practice specialist radiation oncology providers by province- adjusted for utilising beneficiaries.

8.2 Benefits paid by setting

Table 5 shows benefits paid per beneficiary by setting. The average amount paid differed by setting. The average amount spent per beneficiary was nearly twice that in-hospital setting compared to the out-of-hospital setting, R53 680 vs R30 984 for breast cancer, respectively. Similarly, the average amount spent on cervical cancer was R54 760 vs R31 044 for in-hospital and out-of-hospital settings, respectively. On average, cervical cancer was more expensive (nearly R10 000 more) to treat than breast cancer, R46 905 vs R38 114. The maximum amount paid per beneficiary with cervical cancer was R523 695 and R962 103 out-of-hospital and in-hospital settings, respectively, as shown in Figure 7. The maximum amount paid per beneficiary for breast cancer was much higher for the in-hospital setting than for cancer of the cervix at R682 364. In contrast, the amount paid for breast cancer in the in-hospital setting was R 910 431.

Setting	Cancer of breast	Cancer of cervix	Consolidated
Out of hospital	R30 985	R31 926	R31 044
In hospital	R53 680	R62 645	R54 760
	R38 114	R46 905	R38 827

Table 5.

Benefits paid per beneficiary by cancer type and setting.

Figure 7.

Distribution of benefits paid per beneficiary (Cancer of the cervix and the breast).

8.3 Level of OOP by setting

Table 6 shows the levels of OOP by setting; the data indicate that the in hospital setting was twice that and out-of-hospital setting at 2 and 4% for breast cancer, respectively. The same phenomenon was notable in cervix cancer, where the OOP was 2% and 4% for out-of-hospital and in-hospital settings, respectively.

Setting	Cancer of breast	Cancer of Cervix	Combined
Out of hospital	2%	2%	2%
In hospital	6%	4%	5%
	4%	3%	4%

Table 6.

Levels of OOP by setting.

8.4 Benefits paid and OOP benefit design

The analysis of comprehensive and hospital plans attracted higher expenditure levels on cancer of the cervix at R57 205 and R53 037, respectively. However, for breast cancer, benefits paid per beneficiary were higher for EDOs and hospital plans at R42 740 and R41 797, respectively. Benefit design groupings further stratified the analysis. When adjusting for the funding of cancer of the breast, the data show higher levels of OOP in EDOs (5.3% OOP levels) and hospital plans (5.9% OOP levels) compared to comprehensive (3.3 % OOP levels) and partial cover type of plans (3.4% OOP level). A slightly different phenomenon emerges when adjusting for cancer of the cervix, where, hospital plans (4.9% of OOP levels), EDOs (3.8% of OOP levels) and partial cover plans (3.6% of OOP levels) accounted for higher levels of OOP compared to comprehensive plans (2.6% of OOP levels) (Figure 8).

Figure 8.

Average amount paid per beneficiary and OOP levels by benefit design.

8.5 Scheme type: Sector

Table 7 shows that open scheme beneficiaries were exposed to slightly higher co-payment levels than those in restricted schemes. This was prevalent in both beneficiaries with cancer of the breast and those with cancer of the cervix. Similarly, with the average benefit paid on average, open schemes paid R41 797 for breast cancer compared to restricted schemes that paid R30 127. Again, open schemes paid even more for cervix cancer than restricted schemes, at R52 108 vs R43 160, respectively. Despite being less than 5%, OOP for open schemes was twice that of restricted schemes for cancer of the cervix beneficiaries, at 4% and 2%, respectively.

Scheme type	No of the DTP beneficiaries	Paid per DTP beneficiary	OOP (%)
Cancer of breast - treatable	47 886
Open	30 111	R41 797	4%
Restricted	17 775	R30 127	3%
Cancer of Cervix - treatable	4 117
Open	1 699	R52 108	4%
Restricted	2 418	R43 160	2%

Table 7.

Proportion of oncology benefits incurred by members by sector.

8.6 Scheme size

There were no significant differences in breast cancer funding by scheme size. The average expenditure per beneficiary for cervix cancer was higher for medium schemes at R57 911, followed by large schemes at R46 566 and small schemes at R42 176. However, large schemes paid slightly higher than medium and small schemes at R37 801, R35 713 and R30 733 per beneficiary for breast cancer, respectively (Table 8).

Scheme size	No of the DTP beneficiaries	Paid per DTP beneficiary	OOP (%)
Cancer of breast - treatable
Large	44 912	R37 800,76	4%
Medium	1 978	R30 732,57	4%
Small	996	R35 713,29	3%
Cancer of Cervix-treatable
Large	3 920	R46 565,81	3%
Medium	130	R57 911,83	4%
Small	67	R42 175,78	3%

Table 8.

Proportion of oncology benefits incurred by members by scheme size.

8.7 Multivariate regression analysis

Table 9 shows the results of the multivariate regression analysis; the results showed higher cervical cancer rates were significantly associated with the hospital setting, geographic distribution of beneficiaries, sector, and benefit design. At the same time, breast cancer was significantly associated with the geographical distribution of beneficiaries, sector, and scheme type. The odds ratio is 1.24, which indicates that the odds that the cervix’s cancer rate was 1.24 times higher in an in hospital setting than in out-of-hospital. The odds ratio is 6.482, which suggests that the odds of higher rates of cancer of breast cancer are seven times higher in Gauteng than in Limpopo. The odd ratio of 8.521 indicated that the odds of higher rates were 9 times higher in Gauteng than in the Northern Cape. The cancer of the cervix was also significantly associated with scheme types. The odds ratio of 1.494 indicated that the odds of higher rates were nearly twice higher in open schemes than in restricted schemes. The results show the odds ratio of 1.165 for breast cancer, which indicated that the odds were nearly twice higher in open schemes than in restricted schemes. Similarly, North West, where the odds are 1.476, shows breast cancer rates are two times higher in Western Cape than in the Northern Cape province. Our study also found the effect of benefit design on the cervix’s cancer rates, where the odds of 1.594 indicated that higher rates were in comprehensive plans than in hospital plans. The multivariate analysis results for breast cancer revealed that higher rates were significantly associated with the geographical distribution of beneficiaries in the Northern Cape and Gauteng provinces. The odds ratio of 1.587 indicated that Gauteng province had the odds of nearly twice higher rates than the Northern Cape province. The effect of the sector was also prevalent as this was statistically significant in both models. Benefit design and hospital setting did not affect the higher breast cancer rates.

Effect	Cancer of the cervix		Cancer of the breast
	Odds ratio	95 % CI	Odds ratio	95 % CI
In Hospital 1 vs 0	1.24	(1.016, 1.514) *	1.038	(0.963, 1.118)
Province Code EC vs GP	2.508	(1.73, 3.636)	1.201	(1.041, 1.385)
Province Code FS vs GP	2.331	(1.513, 3.592)	1.19	(1.018, 1.391)
Province Code KZN vs GP	1.709	(1.283, 2.278)	1.06	(0.939, 1.195)
Province Code LP vs GP	6.482	(3.493, 12.026) **	1.077	(0.891, 1.301)
Province Code MP vs GP	2.097	(1.386, 3.171)	1.197	(1.017, 1.409)
Province Code NC vs GP	8.521	(3.83, 18.954) **	1.587	(1.286, 1.959) *
Province Code NW vs GP	3.415	(2.141, 5.446)	1.256	(1.064, 1.483)
Province Code OTH vs GP	0.577	(0.168, 1.982) *	1.191	(0.826, 1.718)
Province Code OUT vs GP	1.18	(0.17, 8.172)	0.548	(0.164, 1.825)
Province Code WC vs GP	2.313	(1.693, 3.161)	1.052	(0.943, 1.173)
Type Rest vs Open	1.494	(1.195, 1.868) **	1.165	(1.069, 1.269) *
Size Medium vs Large	0.01	(0.002, 0.042)	0.833	(0.737, 0.941)
Size Small vs Large	0.056	(0.026, 0.119)	0.89	(0.75, 1.057)
Benefit Design; EDOs vs Comprehensive Plans	0.575	(0.408, 0.81)	0.98	(0.866, 1.109)
Benefit Design; Hospital Plans vs Comprehensive Plans	1.594	(1.187, 2.141) **	0.99	(0.888, 1.104)
Benefit Design; Partial Cover Plans vs Comprehensive Plans	0.769	(0.6, 0.985)	1.028	(0.933, 1.133)
Benefit Design; Unknowns vs Comprehensive Plans	0.093	(0.043, 0.199) **	0.978	(0.812, 1.178)

Table 9.

Multivariate logistic model assessing the association between demographic, scheme characteristics and setting variables as predictors of cancer proportion.

** p < 0.001; *p < 0.05; CI: Confidence Interval

9. Conclusion

This study finds a higher number of beneficiaries diagnosed with breast cancer, nearly ten times more than those diagnosed with cervix cancer. These findings are consistent with the literature, where breast cancer is the most common cancer in women [20, 21, 22, 27]. The study also found that 97% of breast cancer was diagnosed in females than in males, who accounted for 3 % of breast cancer, slightly higher than another study conducted in the public sector the study found the rate of 1.4% [33]. The findings, however, were still within range when compared to international norms ranging between 1–3% [29, 30, 31, 46]. The weighted average of women diagnosed with breast cancer was much older at 59 years; however, it was within the range of systematic review and meta-analysis, which showed the average age of female breast cancer in Africa ranged between 30.6 to 60.8 years [47]. There is evidence of a much younger mean age of beneficiaries diagnosed with cervical cancer. The findings of this study were consistent with a study conducted in the public sector in South Africa, which found a mean age of 56.2 years [33]. However, this study shows the early stage of diagnosis in medical schemes in the age band of 20–24 years, thus denoting risk exposure in much younger age profiles. The study found that cervix cancer was diagnosed ten years earlier than breast cancer, and the weighted average age of cervical cancer beneficiaries was 50 years. This is consistent with global trends reporting the average age range (50–53 years). Condition-specific findings show that the proportion of beneficiaries with breast and cervical cancer was higher in Gauteng (41% vs 33%). Though much lower, the Western Cape had a similar phenomenon where the proportion of breast cancer beneficiaries was more than twice that of cervical cancer beneficiaries (21% vs 9%). Other provinces showed a higher proportion of cervical cancer than breast cancer. The study discovered significant differences in the distribution of oncology specialists relative to covered lives in affluent urban provinces like Gauteng and Western Cape and rural provinces like the Eastern Cape and Limpopo. These disparities were more pronounced in KwaZulu Natal and Limpopo (17% vs 12% and 15% vs 5%), respectively. The Gauteng, Western Cape, and KwaZulu Natal provinces accounted for 85% of independent practice specialist oncologists, while the other provinces accounted for only 15%, with other provinces showing less than ten specialists. The distribution of oncologists was less represented in other provinces than Gauteng, which accounted for 93% of the claiming medical oncologists. The two other provinces (Western Cape and KwaZulu Natal) accounted for the balance, with only one medical oncology service provider each. In rural provinces, there was no claiming medical oncologist. These findings further describe higher inequalities geographically, which are also prevalent in the private sector. The study also found a higher proportion of covered lives relative to the balance of independent practice specialist radiation oncologists in Limpopo (Rural) at 1% vs 5%, depicting an urgent need to develop and attract specialists in the province. This finding conforms to previous studies (e.g., [7, 48]) in those urban residences increase the access and uptake of cancer screening. A study by van Eeden et al. [49] further confirmed challenges with oncology services for lung cancer, mainly radiotherapy units primarily located in larger cities, limiting access to rural-based areas. The implication of practical recommendations for medical schemes is a scarcity of medical service providers as they relate to designated service providers and specialist network contracting. The significant shortfall of oncology specialists in poorer provinces directly leads to long treatment delays. High patient volumes in this province affect optimal treatment care irrespective of the sector, as this study shows [50]. The findings revealed a relatively higher proportion of beneficiaries with breast cancer compared to cervical cancer beneficiaries in open schemes than in restricted schemes; in large schemes than in medium and small schemes; in comprehensive plans, EDOs, and hospital plans than in partial cover plans; and in age bands older than 55; in provinces such as the Gauteng, Limpopo, and North West and Western Cape provinces; and in out-of-hospital setting than in an in-hospital setting. The multivariate analysis further supported these findings, which found that higher cancer rates of cervical cancer were significantly associated with the hospital setting, geographic distribution of beneficiaries, sector, and benefit design. Furthermore, this study provides critical insights for the National Health Insurance as they address human resources and relative socio-economic challenges. The regression model’s findings for breast cancer revealed that higher rates were significantly associated with the geographical distribution sector. The industry also affected the higher levels of cancer in the breast proportion. The odds ratio of 1.165 indicated that the odds were nearly twice as high in open schemes than in restricted schemes. Benefit design and hospital setting did not affect the higher breast cancer rates. The average expenditure for the two types of cancers differed by scheme type and was much higher in open schemes than in closed schemes, thus indicating the effect of setting. Cervical cancer was +/-R 10,000 more expensive than breast cancer per beneficiary. Partial cover plans paid around R25 000 for breast cancer compared to other benefit options, which paid around R40 000 per beneficiary region. EDOs and partial cover plans paid just under R40 000 for cervical cancer per beneficiary, while comprehensive and hospital plans paid just over R50 000. Finestone et al. [27] found that low-cost or less comprehensive benefit options paid much less for breast and cervical cancer treatments [27]. The authors found that the average cost for cervical cancer in the public sector ranged between R28 666 and R33 021 for stages 1–4.

The level of OOP for the two cancers was insignificant in the region of 2–4%; however, it still presents a financial burden to beneficiaries and could be detrimental for those rural-based provinces where the barrier to accessing specialist oncology services is even higher. The study recommends support programs (family support, government, medical service providers, private sector and government) for cancer patients and integrated into managed care services. Due to inequalities between and within the two-tiered health system in South Africa, the study proposes a multidisciplinary approach to address the scarcity of resources. Public-private partnerships on cancer treatment and support programs should be the critical feature to help move South Africa closer to Sustainable Development Goal (SDG) 3.4

To reduce, by one-third, premature mortality from NCDs+ through prevention and treatment and promote mental health and well-being by 2030 [51, 52].

10. Limitations

This study has the following methodological limitation:

• Expenditure per beneficiary did not distinguish between diagnosis and treatment (surgical and non-surgical), furthermore does not include additional and follow-up costs.

• Prevalence and incidence were not considered due to a lack of access to primary data

• The data do not consider the long-term treatment of the conditions analysed, including ongoing management of the diseases that could potentially have a cost implication.

• The study also did not adjust for the severity of the condition, nor did it consider demographic characteristics such as race which are potential risk factors.

• The study also did not consider co-morbidities which could exacerbate the conditions even further.

The analysis of the aggregated transaction data restricted the study because it did not consider patient or provider perspectives and experience; future studies should consider qualitative aspects such as patient experience. Future studies should include drivers of co-payment in PMB level of care conditions and an effort to develop approaches and interventions to minimise these.

Acknowledgments

The authors are grateful to Mr Phakamile Nkomo, Mr Martin Moabelo and Mr Sibusiso Ziqubu for their support in concluding this research work.

Conflict of interest

The authors declare that no financial or personal relationships may have influenced them inappropriately in writing this article.

Ethical considerations

The data were assessed and only reported at the consolidated level for privacy and confidentiality. No clinical or patient-specific information was accessed nor reported while conducting this research.