Techniques for Determining the Treatment Costs of Cervical Cancer: A Systematic Review ()
1. Introduction
Uterine cervical cancer (UCC) is the fourth most common malignant neoplasia in women worldwide [1]. It has a heterogeneous distribution, with 84% of cases occurring in less developed regions [2]. Cancer treatment costs are increasing as a result of new medications, innovative surgical procedures, radiotherapy, and technologies related to diagnosis and treatment, thereby resulting in a substantial economic impact for those who pay for healthcare in most countries. The estimated 14 million new cases of cancer annually worldwide lead to significant economic costs, and the total annual economic cost of cancer in 2010 was estimated at approximately US$1.16 trillion. Although impressively high, this figure is underestimated and not equally distributed across all nations [3]. Even though low and middle-income countries represent 84.5% of the world population and 61.3% of new cancer cases globally, these areas account for only 6.2% of financial expenditures on cancer [4]. Middle-income countries (MICs) are defined as having a per capita gross national income (GNI) of US$1026 to $12,475 (2011) [5].
Since resources are limited, economic evaluation has emerged as an important tool in evaluating healthcare [6]. Complete or full economic evaluation refers to the costs and consequences associated with some medical conditions. Studies that do not consider the consequences and contain only cost descriptions represent partial economic evaluations. These studies are defined as cost analysis, cost descriptions, cost of illness, or burden of illness, and can be conducted using different costing methods [7].
Costs refer to all expenditures related to an intervention, including the treatment itself, adverse effects, and acute or late complications [8]. These costs may be direct or indirect. Direct costs are related to spending on medical (physician time, tests, and drugs) and non-medical treatment (food, transport, accommodation, and home aides). Indirect costs refer to lost productivity resulting from absence from work or loss of life due to disease or treatment [9].
The cost of an intervention can be estimated by means of a top-down (TD) or a bottom-up (BU) method. In a TD approach, an adequate population is identified and their aggregated economic or resource costs of a particular health intervention are extracted from the health services billing data. It is then adjusted by cost to charge ratios. In a BU method, estimates are obtained in a two-step process. First, the utilization frequency of individual resources is obtained, after which the frequencies are multiplied by each unit’s cost and summed to yield a total cost. A full assessment of the costs is described by some as micro costing, where details of additional costs are provided―such as the contribution of the care provided by health workers, supplies, or ancillary services [9]. However, this nomenclature is unclear, and some authors use BU as a synonym of micro costing and TD as a synonym of macro costing or gross costing [10] [11]. In addition, there are controversies about the ideal costing method. When different costing methods are used, different cost estimates are frequently reached [9].
Conducting cost description studies may be a difficult task, mainly for low and middle-income countries. Another major point of discussion is transferability of health economic data between jurisdictions; the results of cost evaluations may vary from place to place because of differences in the severity of the disease, the availability of health care resources, clinical practice patterns, and prices [12].
All over the world, especially in developing countries, strong economic health policies are necessary and economic evaluation studies are essential in order to identify the best action for each situation. Thus, the aim of this review is to identify the methods used to measure the cost of treating invasive UCC, and to search for correlations between cancer treatment expenditures and local economies, in articles published in the last ten years.
2. Methods
This systematic review incorporated studies where costing methods of invasive UCC treatment costs were described.
Studies were identified by searching MEDLINE, LILACS, and SciELO from January 01, 2007 to December 31, 2016. The search strategy included the descriptors: “uterine cervical neoplasms” or “cervical cancer” were combined with: “costs and cost analysis”, “drug costs”, “cost of illness”, “cost-benefit analysis”, “economics”, “direct service costs” and “hospital costs”. A free search was also conducted in the references of studies considered relevant. No language restriction was applied.
Studies were included if they met the following criteria: 1) was primary research; 2) described the annual direct cost of invasive UCC; and 3) described a detailed costing method. An initial selection of titles and abstracts was conducted by two independent researchers. Discordant cases were evaluated at a consensus meeting. Thereafter, articles were read in full to ensure that they met the inclusion criteria. Complete economic evaluations, which comprised “cost-effectiveness, cost-benefit, and cost-utility analysis”, were excluded.
Data were collected using a spreadsheet that contained information about the author, location, year of publication, costing method, total annual cost, and annual cost per patient. For comparison, we searched economic and demographic data referent to each country discussed in this analysis. Based on their GNI, countries were divided into MIC (between US$1026 and US$12,475) and high-income countries (HIC) (more than US$12,476) [5].
All costs were inflated for 2016 and then converted to 2016 International dollar (I$) using purchasing power parities (PPP) from the World Bank consumer prices (Table 1) [13] [14].
In order to test the differences in costs between MIC and HIC, statistical analysis was done using a non-parametric Kruskal-Wallis test in STATA, version 12.1 SE. A p value < 0.05 was considered statistically significant.
This project was approved by the Ethics Committee at IMIP (document
Table 1. Annual inflation rate between 2006 and 2016.
*Source: The World Bank, 2017 [15].
number 4026-14). As secondary published data were analyzed, no informed consent was needed.
The selection algorithm is described in Figure 1.
3. Results
The initial search retrieved 1581 references. Among these, 1546 articles were excluded by title and another 22 were excluded because of eligibility criteria. The remaining 13 articles were included in the analysis.
The economic characteristics of the countries referred to in this review are outlined in Table 2.
Ten studies used the payer perspective and included only direct costs. Four studies used the societal perspective, where direct and indirect costs were calculated. For these articles, indirect costs represented between 14.11% (only morbidity costs) and 80.50% (morbidity and mortality costs included) of the total costs. Nine studies did not name the costing method, although they detailed how costs were derived. In this review, the costing method (CM) of each article was classified into TD and BU; these methodologies were used in seven articles each, with one article using both (see Table 3). Only four studies specified that the cost referred to the first year of diagnosis [20] [21] [22] [23].
The annual per patient invasive UCC treatment cost, considering only direct costs, were inflated to the 2016 local currency and then converted to International dollar (I$). Results varied from I$2146.22 (Poland) to 34351.54 (Sweden), mean of I$16390.15 (±9566.20). For comparison, we obtained the ratio between gross domestic product (GDP) per capita (in 2016 I$) and the annual cost per
Table 2. Economic characteristics by country.
GDP: Gross Domestic Product; GNI: Gross National Income; PPP: Purchasing Power Parity; USD: United States Dollar (US$); I$: International Dollar; EUR: Euro (?; TND: Tunisian dinar; MAD: Moroccan dirham; BRL: Real (R$); JPY: Yen (¥); CND: Canadian Dollar; SEK: Swedish Krona; PLN: Zloty; MXN: Mexican Peso.
CM: Costing Method; NN: Not Named; TD: Top-Down; EUR: Euro; BU: Bottom-Up; NA: Not Applicable; USD: United States Dollars; JPY: Yen; CND: Canadian Dollar; PLN: Zloty; MXN: Mexican Peso. #Not described in the article, value calculated by this reviewer based on data in the article.
patient, results ranged from 7.71% to 241.63% (median 61.36%). Final costs are outlined in Table 4.
Based on 2011 per capita GNI, Tunisia, Morocco, Brazil, and Mexico are defined as MIC by the World Bank [5]. The Kruskal-Wallis test revealed that MIC spent proportionally more than HIC (p = 0.032), as illustrated in Figure 2.
However, when separated by costing method, no difference was found (p = 0.522) (Figure 3).
4. Discussion
This review compared results by relating the adjusted cost of invasive UCC in 2016 I$ in different countries with each country’s GDP per capita, thus accommodating for the huge economic differences across jurisdictions.
Most studies used the payer perspective, where only direct costs were included. Indirect costs were described when the societal perspective was used.
Figure 2. Costs percentage distribution according to gross national income. GNI: gross national income; HIC: high income countries; MIC: middle income countries; TC: total ICC treatment costs per patient per year; GDP: gross domestic product per patient.
Table 4. Uterine cervical cancer: annual cost per patient treatment in 2016 I$ and percentage ratio between per capita gross domestic product and treatment cost.
*Not explicit in the article; data were calculated by the reviewer using data from the article. GDP: Gross Domestic Product; GNI: Gross National Income; PPP: Purchasing Power Parity; USD: United States Dollar (US$); I$: International Dollar; EUR: Euro (?; TND: Tunisian Dinar; MAD: Moroccan Dirham; BRL: Real (R$); JPY: Yen (¥); CND: Canadian Dollar; SEK: Swedish Krona; PLN: Zloty; MXN: Mexican Peso.
Figure 3. Costs percentage distribution according to costing method. TD: top-down; BU: bottom-up; TC: total ICC treatment costs per patient per year; GDP: gross domestic product per patient.
Although the societal perspective has the stronger claim to be the basis for comparison across studies [33] , only one article incorporated the ratio of the mortality cost to the total cost, besides the morbidity cost. The mortality cost was calculated by summing the income that could have been earned in future if death had not occurred; the indirect cost in this article represented 80.5% of the total costs [27]. Because of this discrepancy, and to make values comparable, we have excluded indirect costs from the statistical analysis.
We found no significant difference in the results derived by different costing methods. Actually, most articles did not name the costing method. However, they did describe how costs were calculated, and could therefore be separated into studies using TD and BU methods. Although BU is the theoretically correct way to estimate service cost, this approach may not be practical in all cases; the resources required for the BU costing could outweigh the benefit of more accurate costing [11]. Experience shows that the TD approach could be useful and reasonably accurate in cases where marketed health technologies (expensive drugs and medical devices) are responsible for most of the resource costs. In these cases, the BU approach may yield a very similar result, but could be more expensive and time consuming. For international comparative studies, the use of the TD approach is recommended [34].
In this review, MIC had a relatively higher expenditure as percentage of per capita GDP than HIC on the treatment of invasive UCC. This is in some respects different from what was demonstrated in earlier studies, where Latin America spent 7.7% of its GDP on health, while the USA and Canada spent approximately 18% and 12%, respectively [4]. This difference can be explained by the lower incidence of invasive UCC in HIC when compared to MIC [1]. Even though HIC had a superior percentage of GDP expenditure on health as a whole, it probably was not relevant to invasive UCC.
This study has limitations. First, only four studies specified that the treatment costs were related to the first year after diagnosis. It has been demonstrated that costs may change depending on the time since diagnosis, and costs are the lowest in the period between the initial and end-of-life phase, following a “U-shaped” curve [35]. In fact, costs can be higher when only the cost of the first year of treatment is included, compared to the inclusion of the costs for all phases. In this review, we found two articles referring to Morocco, in which the cost in one study, based on data from the first year of treatment, was more than twice the cost calculated by the other [20] [21]. Second, two articles did not refer to the entire country’s population and data were extrapolated [26] [28]. Finally, the lack of clarity in cost concepts and differences in the inclusion and exclusion of cost items made these comparisons problematic. Consequently, the findings should be interpreted carefully.
However, because all costs were adjusted to one common currency and correlated to each country’s GDP, we were able to determine the cost of invasive UCC treatment relative to the health budget of each evaluated country. Economic evaluations are important tools to guide decision makers with respect to appropriate resource allocation. The methodology of these studies is heterogeneous, but forms of equivalency should always be sought.
Although we recognize many caveats, this is to our knowledge the only study that has tried to determine a parallel between invasive UCC treatment costs across different regions. The intent is that this would contribute to the continued efforts of policy makers to standardize and develop reproducible economic studies, especially in low and middle-income counties.
5. Conclusion
We found that, for the treatment costs of invasive UCC, the percentages of GDP per capita were statistically higher in MIC than in HIC. However, no significant difference was found between costing methods, and the top-down approach could be used.