Abstract
Objectives. This cross-sectional study assesses resource use and costs in different states of prostate cancer (PCa) in a real-life setting. Costs were estimated as incremental costs due to cancer for a six-month period and they included direct medical costs, productivity costs and costs of informal care.
Methods. Resource use and cost data, irrespective of who the payer was, were retrieved from the registries for 611 PCa patients in the Helsinki area in Finland. In addition, patients answered background questions concerning informal care, work capacity and educational status. Patients were divided into four mutually exclusive groups based on disease state and time from diagnosis: primary (local disease, first six months after diagnosis; n = 47), rehabilitation (local disease, 0.5–1.5 years after diagnosis or recurrence; n = 158), remission (local disease, more than 1.5 years after diagnosis; n = 317) and metastatic (after detection of metastases; n = 89).
Results. Costs differed markedly between the states of disease. Mean direct health care costs for the six-month periods were: primary treatment state €2750, rehabilitation state €1143, remission state €760 and metastatic state €7423. Productivity costs were also highest (€4277) in the metastatic state. Overall, the average share of indirect costs was around one third of the total costs. However, when including informal care, their combined share of the total costs increased to around half or more.
Conclusions. The results provided state-specific estimates of the direct health care and indirect costs of PCa in Finland. The treatment of metastatic disease is significantly more costly than treatment of early stage PCa. Although direct medical costs were higher compared to productivity costs, they should be taken into consideration when evaluating the costs of PCa.
Prostate cancer (PCa) is the most common non-skin cancer in men and the third most common cause of cancer deaths in men in Western Europe. In 2012, about 1.1 million men worldwide were diagnosed with PCa, accounting for 15% of the cancers diagnosed in men. Most cases (around 70%) are diagnosed in more developed regions and the burden of PCa is remarkable in regard to both economic and clinical aspects [Citation1,Citation2]. PCa primarily affects men over the age of 65 years, with median age at diagnosis being 72 years. Therefore, the majority of those with PCa will die of other causes without suffering significantly from PCa, due to their advanced age at diagnosis.
There is less variation in mortality rates worldwide than in the observed incidence. This is because PSA testing has a much greater effect on incidence than on mortality [Citation1]. The increased use of PSA testing, combined with an ageing population, will burden health care systems with an increasing number of PCa patients requiring treatment.
For (health) economic evaluations it is essential to estimate both direct and indirect costs to avoid any possible underestimation of the true cost of a disease. Particularly, disregarding productivity losses may result in undervaluation of the disease burden. Informal care provided by family members or relatives and friends is an important element of care for many cancer patients and, therefore, recommended to be included in economic evaluations.
This study estimates the costs in different states of PCa in a real-life situation. Costs and resource use were analysed as direct health care or medical costs, productivity costs and costs of informal care.
Patients and methods
Study design
This study is part of a larger project investigating the health-related quality of life and economic impact of breast, colorectal and prostate cancers in Finland [Citation3,Citation4]. It is a cross-sectional registry and survey study and was approved by the local ethics committee. Patients were recruited in the Helsinki and Uusimaa Hospital District between September 2009 and December 2010. The Helsinki and Uusimaa Hospital District serves about 1.5 million citizens of Southern Finland and is responsible for about one third of all specialist care in oncology in Finland.
For every study patient, two control subjects were extracted from the Social Insurance Institution's (SII) electronic records. The control group subjects were standardised for age, gender and place of residence. Costs other than those incurred by specialist care were compared against a control group and reported as incremental costs related to PCa for a six-month period. The sample extracted from SII registries covered outpatient medication, sickness allowances and use of private health care. All costs are presented at the 2010 price level.
Patients
Patients were identified from hospital records by date of diagnosis. Adult patients who fulfilled the inclusion criteria were invited, and those agreeing to sign a written consent form were enrolled. The patients’ clinical background information regarding disease state and treatments given were collected from hospital records. Enrolled patients were divided into four mutually exclusive groups based on disease state: less than six months from diagnosis (primary), 6–18 months from diagnosis (rehabilitation), subsequent years of remission (remission) and metastatic disease (metastatic). A similar grouping was also used previously when health-related quality of life results were reported for the same patients [Citation3].
Direct medical costs
Specialist care
Specialist care data include inpatient episodes and outpatient visits to secondary health care. These retrospective data also included inpatient medicines. All secondary care visits that were not deemed, by an experienced clinician, to be related to cancer treatment were thus excluded. All the data, including cost data, came from the hospital's electronic records, and can be considered comprehensive. Costs of travel to treatment were available from the SII records when they exceeded the maximum co-payment for the patient (€14).
Primary health care
Primary health care in Finland is funded and organised by municipalities. Primary health care services are available for all residents in the municipality and services are free for patients with the exception of some user fees. Primary care data were collected from the three largest cities in the catchment area of the hospital, Helsinki, Espoo and Vantaa, covering more than 80% of study patients. The primary care data included information on general practitioner and nurse visits, home hospice care and primary care hospitalisation. Data concerning the reasons for the visits, however, remained unavailable. To estimate the proportion of visits that were related to PCa, a background questionnaire was used to identify the share of primary care visits that were cancer-related. For patients whose home municipality was different to any of those mentioned earlier (n = 100; 15.9% of the patients), missing primary care costs were imputed by using average cost from the same disease state. Travelling costs related to PCa visits were also included.
Private health care
Data on the use of private health care services were available from SII's registries. The costs of visits related to PCa were estimated as those that exceeded the control population's private health care usage and travel costs. Resource use and unit cost data were also included from the local private hospice care unit.
Medicines
Outpatient medicine costs related to PCa were extracted from SII's electronic records and compared to that of the control population. The cost of medication for the treatment of PCa dispensed in the hospital was extracted from hospital records.
Productivity costs
The work status of the patients and the potential retirement from work due to PCa were obtained from a patient survey. Patients were asked whether they were working, retired due to cancer, retired due to other reasons or not working due to some other reason. The registries of SII were used to calculate the number of days patients were on sick leave and absent from work due to PCa during the six-month observation period.
The human capital approach was used to value loss of productivity [Citation5]. To assess the loss of productivity, wages were translated to average labour costs by including employer's social security payments of, on average, 38.6% in addition to pre-tax salary [Citation6]. Actual annual wages were used when calculating productivity losses due to early retirement. The valuation of productivity losses due to sick leave followed the same approach, however, we used average daily cost of labour based on actual wages. Daily average labour cost was €213.96.
Informal care
Using a background questionnaire, the patients estimated the number of weekly hours of informal care received. Informal care was defined as care that patients had received free of charge from family members and friends. These estimates were used to extrapolate the costs to the six-month observation period in each disease state.
There is no established approach for calculating the value of informal care. We chose to use the proxy good method in which the value of informal care is calculated by multiplying the number of hours of informal care by the value per hour for each care task performed. In the proxy good method the value per hour is based on the shadow price of a market substitute. This provides valuable insight into the costs of replacing informal care with formal care [Citation7].
To value informal care, a practice nurse's mean hourly pre-tax salary of €13.63 for the year 2010 was used [Citation8]. The hourly labour cost was then calculated by adding side costs on top of pre-tax salary, resulting in an hourly labour cost of €18.89.
Data analysis
The main objective of the study was to estimate the mean cost of different states of PCa. The statistical significance of differences in costs and resource use between the disease states was tested by using confidence intervals. Fixed log linear multivariate models were built to analyse how background factors are associated with total costs. We used total costs as the dependent variable and, due to skewed distribution of cost variables, natural logarithm transformation was applied. Age, cohabiting, educational level, and symptoms and functioning scales (physical, role, social, emotional and cognitive functions) measured by a health-related quality of life instrument developed for cancer patients by the European Organisation for Research and Treatment of Cancer (EORTC QLQ-C30) were used as independent variables [Citation9,Citation10]. We built three different models: one for primary treatment, one combining remission and rehabilitation and one for metastatic disease. A risk level of 5% was used for type 1 error in all analyses. Analyses were performed with SPSS 22 (SPSS Inc., Chicago, IL, USA) [Citation11].
Results
Patients
The study population consisted of 611 PCa patients. Of these, 522 had local disease and 89 advanced disease. The patients’ ages ranged from 44 to 93 years (mean 69.2 years). Most of the patients were married or cohabiting (73%) and half had received higher education (49%). The mean age differed slightly between patients with localised (68.8 years) or advanced disease (72.1 years). Around one in five from the localised disease group was still working, compared to one in 10 in the metastatic disease group ().
Table I. Patient characteristics.
Direct health care costs
The mean total direct costs increased substantially after detection of metastases and differed markedly from the costs related to localised disease states. The distribution of specialist care costs between the disease states followed a similar pattern to the total direct health care costs. The costs were highest in the metastatic state of the disease ().
Table II. Direct healthcare costs (€).
The mean primary health care costs were relatively modest, being twice as high in newly diagnosed patients compared to the other states. Private health care costs were fairly similar to those observed in the control group. In localised disease, a small additional cost from the use of private health care was observed. However, in the metastatic state of the disease patients used less private health care than their control group ().
The mean costs of medicines, including inpatient and outpatient medicines, were lowest in recently diagnosed patients (primary) €141, while in the two other groups with localised disease medicine costs were slightly over €300. During the metastatic state, medicine costs increased sharply to a mean of €4354 ().
The mean travel costs were highest in the metastatic state (€263), otherwise they were even lower than the control group's costs ().
Productivity costs
Most of the patients were not working which was expected considering their advanced mean age. Of the patients, 102 (16.7%) were employed and of the 430 retired patients more than two thirds were receiving the state pension. Only 18 patients (2.9%) were receiving disability pension due to their cancer, and 28 patients (4.6%) were receiving disability pension due to reasons other than cancer. Seventeen patients (2.8%) reported being unemployed or not working ().
The number of days absent from work due to sick leave was relatively low, being on average less than one day during the six-month period. However, in the metastatic group it was almost four days. The mean number of days patients were absent from work due to early retirement was much higher, being highest in the metastatic and primary treatment groups (8.6 and 7.7 days, respectively) ().
Table III. Productivity losses.
The estimated productivity losses were highest in the primary and metastatic groups (). Most of the productivity losses came from early retirement and the impact of sick leave was limited, being meaningful only in the metastatic group.
Informal care
Only 34 patients (5.6%) received informal care due to PCa either from their family or others. The mean number of hours of informal care in localised disease was modest (around one hour per week or less) but increased with disease progression, being highest in the metastatic group (6.5 hours per week). The estimated value of informal care during the six-month period was highest in the patient group with advanced disease ().
Table IV. Informal care.
Factors associated with total direct health care costs
Multivariate regression analysis revealed that patient characteristics and background factors explained 9–27% of the variance in total costs associated with PCa in various states of the disease (). From background factors, with newly diagnosed patients a higher education seemed to lead to more costs and in the metastatic state, cohabiting was associated with less cost burden. When analysing QLQ-C30 symptoms scales, emotional functioning in primary state, and physical functioning in remission/rehabilitation state were significantly associated with costs. From QLQ-C30 functionality scales, nausea/vomiting and financial difficulties only affected costs in the remission/rehabilitation state.
Table V. Multivariate analysis of cost drivers with the natural logarithm of total costs as the dependent variable.
Discussion
The total six-month costs varied quite substantially between the states of the disease. Our study assessed the real-life costs, which are incurred when patients are treated according to current standards. The analysis was focused on only those costs linked to PCa. This approach makes it possible to establish a more reliable estimate of the costs related to disease management beyond other health care utilisation within this aged population.
Our results support previous findings that PCa is an expensive cancer to treat, especially in the advanced state [Citation12]. Consequently, it will definitely have an impact on future health care costs, especially with an ageing population.
To make a robust evaluation of costs related to disease, it is important to include all the costs related to that condition regardless of who pays them. In our study, the analysis captured health care resource utilisation and direct costs both in the primary and specialist care settings, assessed productivity costs due to disease and also informal care given by family members or relatives and friends. Many studies in this area have focused on direct health care costs only [Citation13], although there have been some attempts to assess productivity costs also in the cancer setting [Citation5,Citation14–16].
In our study, the share of productivity costs out of total costs (combined direct health care costs and productivity costs) varied between 37% and 51% in different states of the disease. However, when informal care costs were also included, the share of indirect costs (productivity costs and informal care costs) increased to 47–64% out of total costs (combination of direct health care costs, productivity costs and informal costs). It has been estimated that indirect costs may result in up to 70% of the total burden of disease in breast cancer patients [Citation17]. A PCa cost study from the US evaluated indirect costs in PCa, including out-of-pocket costs, and came to the conclusion that up to 60% of the total cost burden in PCa was due to indirect costs [Citation18]. As breast cancer patients are on average younger than PCa patients and more often in their active work-life phase, it may be considered surprising that productivity costs and costs of informal care make up such a remarkable share of total costs in PCa.
Total direct health care costs for a six-month period varied in our analysis between €760 and €7423 depending on the state of the disease. These results are in line with other studies [Citation19], which have analysed costs of PCa patients in different settings, usually when introducing new treatment options. Nevertheless, we were not able to find studies which used a similar patient grouping, and therefore any attempts to compare results need to be made with caution.
Based on our results, direct costs for a six-month period in localised PCa, especially in the early treatment phase, may not be as high as in the case of some other cancers (i.e. colorectal, etc.) [Citation20]. Also, the estimates for a six-month period in the subsequent states (rehabilitation and remission) show that resource use is moderate and the costs of interventions or adjuvant medications used in PCa are relatively modest. The costs, however, start to accumulate when the disease progresses.
Loss of productivity in our analysis is based on the human capital approach. Productivity loss due to sick leave was modest, however, more losses being generated by early retirement. Overall, only 177 patients (29%) were younger than 65 years, and consequently still at working age, and only 17% of the total number of patients were working, curtailing the potential productivity losses.
Productivity losses varied between states of disease. In localised disease they were fairly low but naturally increased in advanced disease states. The share of productivity losses out of all costs was prominent (47–64%), and uncertainty remains over how to best evaluate them. Hanly et al. found in their study, using breast and prostate cancer as examples, that results may differ substantially between approaches. However, so far there is no consensus on this topic and the human capital approach remains the most commonly used [Citation5].
In the rehabilitation and remission groups there were less productivity losses illustrating that patients were doing fine and were able to continue their normal routines and working. Nevertheless, there were more productivity losses in the primary and metastatic states as more intensive care and interventions were required.
The costs of informal care have been less studied than direct costs or productivity losses but they may also constitute a meaningful impact on the total burden of disease. On average, the total number of hours of informal care received was very modest, which, however, was not true for those 34 (6%) patients who actually had access to care and used it intensively. Most of these patients had more advanced disease. Our results did not reveal any reasons why these patients needed this amount of informal care or whether other patients received some other type of support that our study was not able to capture. Nevertheless, informal care plays an important part in the support network of PCa patients and more research is needed to reveal its true role.
The most important limitation of our study is the cross-sectional study design. It might have been more effective to follow the same patients throughout their disease progression. However, this would presumably have caused a much longer follow-up time and a higher number of enrolled patients. Our response rate was 61.5%, which can be seen as a limitation. However, we did not have access to non-respondents’ medical records and therefore were not able to verify whether they had more severe disease. We did not try to categorise patients inside our groups according to treatments they had received but included all the patients in the analyses as they represent the real-life situation in PCa care. This is also one of the strengths of our study. Most other analyses study certain interventions and patients are included or excluded according to the intervention that is under evaluation. Although this gives a robust estimate of the efficacy of the intervention, it does not reflect the everyday situation in health care units where different types of PCa patients are treated with multiple interventions.
We may have missed some out-of-pocket costs mainly related to travelling. We included only costs which had been reimbursed to patients but may have missed some travel costs related to visits using the patient's own car or transport from friends. Overall, as there was very little difference between our patient population and the control group, the potentially missed costs probably play only a minor role.
One of the strengths in our analysis is the fact that the costs of primary health care were included. This gives much needed information on costs related to PCa outside specialist units. We have also included all the medicines related to PCa regardless of whether they were dispensed in the hospital or delivered by a pharmacy.
Our analysis regarding the costs related to different states of PCa has one distinguishing feature compared to many other similar types of cost analyses. We used a control group to estimate the incremental resource use and costs related to PCa compared to usual health care usage of persons with a fairly advanced mean age. We did not try to estimate the overall cost of PCa in a cost-of- disease type analysis, but focused on the incremental cost linked to management of this disease in different states. Cost studies in the past have focused on assessing direct costs only, and therefore much important information has remained unknown. In fact, ignoring productivity costs may have a large effect on the outcome of evaluation, and may even be decisive in considering whether an intervention can be considered cost-effective or not [Citation21]. This may also be true in the case of informal care [Citation22]. Our results support this, and highlight the importance of taking all cost elements into consideration, especially if an evaluation is meant to be performed from a societal perspective.
Conclusion
Direct costs related to different states of PCa are significant. However, productivity losses and costs of informal care also play a major role when estimating the total burden of PCa. Excluding such a large share of costs from cost effectiveness considerations might have a significant impact on the decision making process of health economic evaluations or health technology assessments (HTA).
Additional studies are still needed to better appreciate the impact of the increasing economic burden of PCa management. The ageing of the population will substantially increase the demand for all health care resources, and PCa is definitely one of the areas where a significant number of patients will be diagnosed and treated. From a health policy perspective, more research is needed to make appropriate choices on resource allocation decisions to guarantee the rational use of health care resources.
Acknowledgments
This work was supported by the Cancer Society of Finland and GlaxoSmithKline Oy, Finland. The study sponsors were not involved in the study design, or the collection, analysis and interpretation of data nor were they involved in the writing of the manuscript; or in the decision to submit the manuscript for publication. All authors participated in the design of the study, data collection and drafting of the manuscript.
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References
- Bray F, Ren JS, Masuyer E, Ferlay J. Estimates of global cancer prevalence for 27 sites in the adult population in 2008. Int J Cancer 2013;132:1133–45.
- Piper NY, Kusada L, Lance R, Foley J, Moul J, Seay T. Adenocarcinoma of the prostate: An expensive way to die. Prostate Cancer Prostatic Dis 2002;5:164–6.
- Torvinen S, Färkkilä N, Sintonen H, Saarto T, Roine RP, Taari K. Health-related quality of life in prostate cancer. Acta Oncol 2013;52:1094–101.
- Färkkilä N, Torvinen S, Roine RP, Sintonen H, Hänninen J, Taari K, et al. Health-related quality of life among breast, prostate, and colorectal cancer patients with end-stage disease. Qual Life Res 2014;23:1387–94.
- Hanly P, Timmons A, Walsh PM, Sharp L. Breast and prostate cancer productivity costs: A comparison of the human capital approach and friction cost approach. Value in Health 2012;15:429–36.
- Hujanen T, Kapiainen S, Tuominen U, Pekurinen M. Terveydenhuollon yksikkökustannukset Suomessa vuonna 2006. Helsinki: Stakesin työpapereita 3/2008. [cited 10.8.2014]. Available from: https://www.julkari.fi/bitstream/handle/10024/76754/T3–2008-VERKKO.pdf?sequence = 1.
- Hoefman RJ, van Exel J, Brouwer W. How to include informal care in economic evaluations. Pharmacoeconomics 2013;31:1105–19.
- Statistics Finland. Official statistics of Finland. Structure of earnings 2010. 2011. Statistics of Finland. [cited 10.8.2014]. Available from: http://www.stat.fi.
- Aaronson NK, Ahmedzai S, Bergman B, Bullinger M, Cull A, Duez NJ, et al. The European Organization for Research and Treatment of Cancer QLQ-C30: A quality- of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst 1993;85:365–76.
- Fayers P, Aaronson N, Bjordal K, Groenvold M, Curran D, Bottomley A, et al. The EORTC QLQ-C30 scoring manual, 3rd ed. Brussels: European Organisation for Research and Treatment of Cancer; 2001.
- Statistical Package for the Social Sciences [computer program]. Version 21, 2012.
- Kalseth J, Halsteinli V, Halvorsen T, Kalseth B, Anthun K, Peltola M, et al. Cost of cancer in Nordic countries: A comparative study of health care costs and public income loss compensation payments related to cancer in the Nordic countries in 2007; 2011. [cited 10.10.2014]. Available from: http://www.ncu.nu.
- Sanyal C, Aprikian AG, Chevalier S, Cury FL, Dragomir A. Direct cost for initial management of prostate cancer: A systematic review. Curr Oncol 2013;20:522–31.
- Brown MI, Lipscomb J, Snyder C. The burden of illness of cancer: Economic cost and quality of life. Annu Rev Public Health 2001;22:91–113.
- Yabroff KR, Davis WW, Lamont EB, Fahey A, Topor M, Brown ML, et al. Patient time costs associated with cancer care. J Natl Cancer Inst 2007;99:14–23.
- Bradley CJ, Yabroff KR, Dahman B, Feuer EJ, Mariotto A, Brown ML. Productivity costs of cancer mortality in the United States: 2000–2020. J Natl Cancer Inst 2008; 100:1763–70.
- Lidgren M, Wilking N, Jönsson B. Cost of breast cancer in Sweden in 2002. Eur J Health Econ 2007;8:5–15.
- Jayadevappa R, Schwartz JS, Chhatre S, Gallo JJ, Wein AJ, Malkowicz SB. The burden of out-of-pocket and indirect costs of prostate cancer. Prostate 2010;70: 1255–64.
- Mehra M, Wu Y, Dhawan R. Healthcare resource use in advanced prostate cancer patients treated with docetaxel. J Med Econ 2012;15:836–43.
- Färkkilä N, Torvinen S, Sintonen H, Saarto T, Järvinen H, Hänninen J, et al. Costs of colorectal cancer in different states of disease. Acta Oncol Epub 2014 Dec 18:1–9.
- Krol M, Papenburg J, Koopmanschap M, Brouwer W. Do productivity costs matter? The impact of including productivity costs on the incremental costs of interventions targeted at depressive disorders. Pharmacoeconomics 2011; 29:601–19.
- Hoefman RJ, van Excel J, Brouwer W. How to include informal care in economic evaluations. Pharmacoeconomics 2013;31:1105–19.