Economic evaluation of hematopoietic stem cell transplantation

Abstract

Hematopoietic stem cell transplantation is a highly expensive clinical intervention with considerable therapeutic benefit but serious adverse effects on health status in some circumstances. Consequently, it is an important target for economic evaluation in which the monetary costs and clinical consequences of optional treatment strategies are compared. The need for such formal assessment is further demanded by the expanded use of hematopoietic stem cell transplantation globally, including its utilization in developing countries. With respect to costs, those incurred by patients and families are often substantial, while those incurred by hospitals may be inadequately reimbursed. Determination of consequences should not be limited to measurements of clinical effectiveness but rather include adjustments for quality of life. Rigorous economic evaluation can provide hard evidence in deliberations of value for money, especially in the context of limited resources for health care.

Keywords:

• Economic
• Evaluation
• Stem cell
• Transplantation

Introduction

Economic evaluation of health care programs is the comparative analysis of clinically reasonable alternatives with respect to both costs and consequences.1 The former focuses usually on the direct monetary costs borne by the health care sector, commonly excluding the costs to other stakeholders including patients and families. The latter includes various outcome measures of the clinical interventions: the consequences.

Economic evaluations are classified according to the way in which the consequences are measured.1 In cost-benefit analysis, this is in terms of monetary units, e.g. dollars, in cost-effectiveness analysis, this is in terms of natural units such as life years (survival), and in cost-utility analysis, consequences are measured using preference scores of health-related quality of life for outcome health states from which quality-adjusted life years are then estimated. A model to generate such estimates for allogeneic bone marrow transplantation was published more than 15 years ago.2

The purpose of the current contribution is not to provide an exhaustive review of published experience in this area, but rather to highlight the important considerations with selected references to the literature and, when possible, to exemplify the challenges faced in conducting hematopoietic stem cell transplantation (HSCT) in low and middle income countries. The emphasis will be on allogeneic procedures for, in general, these are associated with higher costs and more adverse consequences than the autologous counterparts.

In a review of the most rapidly escalating hospital costs in the USA, for the period of 2004–2007, the Healthcare Costs and Utilization Project identified bone marrow transplantation as the procedure at the top of the list.3 The aggregate costs grew by almost 85% in this interval, to $1·3 billion in 2007. The incremental growth was more related to an increased number of procedures (60%) than to an increase in the mean cost of hospital stay (25%).

The challenges involved in economic evaluation of stem cell transplantation (SCT) are addressed in a comprehensive systematic review undertaken from the perspective of the National Health Service (NHS) in the UK.4 The authors concluded that sibling donor allogeneic SCT is more effective than chemotherapy alone in adults and children with acute myeloid leukemia in first complete remission, except for those with good risk disease, and for adults with acute lymphoblastic leukemia in first complete remission, while autologous SCT is equal to or less effective than chemotherapy alone. However, ‘No firm conclusions could be drawn, regarding the cost-effectiveness of SCT in the UK NHS’. This conclusion was based on the limitations of available evidence.

Consideration of Variables

There are numerous important variables pertaining to SCT in the current context. These are illustrated in Table 1. In regard to stem cell source, published information is heterogeneous with respect to equality. Reports are often retrospective regarding costs, with a mixture of disease and risk categories, different donor types and sources of stem cells, and non-uniform conditioning regimens. These challenges are exemplified in a publication from Houston, Texas.5 Nonetheless, in addition to more rapid engraftment, peripheral blood has been reported to offer a less expensive alternative than bone marrow in both the allogeneic (children with acute lymphoblastic leukemia in first complete remission)6 and autologous settings (adults with relapsed lymphoma).7 The former showed equivalence in disease-free survival, and the latter in less supportive care and better quality of life.

Table 1. Important variables in stem cell transplantation

Item	Variables
Stem cell source	Bone marrow, peripheral blood, cord blood
Donor type	Sibling, matched unrelated donor, haplo-identical donor
Conditioning intensity	Myeloablative, reduced intensity, non-myeloablative techniques
Other considerations	Use of granulocyte colony-stimulating factor, T-cell depletion, etc.

With respect to umbilical cord blood, there is a highly informative report from the Institute of Medicine that was commissioned by the US Congress in 2003 to determine the optimum number of units that should be stored nationally.8 It was noted that 22 non-profit organizations had a collective inventory of 80 000 units. The study revealed that the cost per life year gained as a function of inventory level was increasing and concluded that ‘an ethnically diverse inventory of 150,000 cord blood stem cell units would help provide appropriate matches for 80 to 90 percent of patients seeking matched cord blood stem cell transplants’.

Turning to the issue of donor type, there is evidence that the use of matched unrelated donors is more expensive than sibling donors, and one study from Holland reported that one-third of the costs associated with the former are incurred in the search process.9 In what seems to be a counter-intuitive outcome, the experience of a single center in China was that the cost of stem cell collections was lower in human leukocyte antigen haplo-identical/mismatched related donors with combined factor-mobilized peripheral blood and bone marrow harvests than in human leukocyte antigen-identical matched donors with factor-mobilized peripheral blood harvests alone.10 In a clinical trial comparing conventional chemotherapy with high-dose chemotherapy and autologous stem cell ‘rescue’, the latter resulted in substantially more morbidity and costs with no improvement in survival for women with metastatic breast cancer.11

On the matter of different conditioning regimens, there is no clear economic consensus. Again this often reflects the lack of rigor in study design, limiting interpretation even when results have face validity.12 Better studies, such as that from Saito et al.,13 generally support the finding of lower costs associated with reduced intensity regimens. However, initial cost savings from non-myeloablative conditioning may be offset by higher expenditures related to late complications and re-admissions.14

The challenge of differing methodologies is well illustrated in the comparison of transplantation versus imatinib in patients with chronic myeloid leukemia. Using a Monte Carlo microsimulation, Skrepnek and Ballard demonstrated that, over a 2-year period, imatinib was both less costly and more efficacious than matched unrelated donor transplants;15 but in a report from the Czech Republic, in which the costs of imatinib were estimated, these exceeded the costs associated with SCT using a reduced-intensity conditioning regimen, again over a 2-year period.16 Another example of ‘apples and oranges’.

Among the group of ‘other considerations’ is that of manipulating the graft. In the T-cell Depletion Study, the 3-year disease-free survival was compared in patients who received unrelated donor transplants with or without T-cell deletion.17 More than 400 patients participated in this multi-center study and there was no significant difference in disease-free survival between the two arms. Moreover, the treatment costs were similar in the two groups, with less acute graft-versus-host disease being offset by more serious infections among the patients who received T-cell deleted grafts.

Issues for Developing Countries

The European Group for Blood and Marrow Transplantation (EBMT) reported a decade ago that one in 2000 persons per 10 years was likely to be treated by SCT in western Europe.18 The total number of reported transplants had risen from 18 in 1973, when EBMT was founded, to more than 20 000 in the year 2000 with a corresponding increase in transplant teams from eight to more than 600 in 35 countries. Team density was defined as the number of teams per 10 million inhabitants. Countries were grouped into three categories of health care systems: social insurance-based or Bismarckian type, tax funded or Beveridge type, and centralized or Semashko type. Transplant efficiency was calculated by dividing the transplant rate in each country by its health care expenditures (HCE) as a fraction of gross national product (GNP) per capita, and grouped by health care system. Team density varied from 0 to 19·6 and was closely correlated with transplant rates that, in turn, were clearly correlated with both HCE and GNP per capita. Transplant efficiency ranged from 11·56 to 44·8 transplants per 10 million inhabitants for each US$10 000 spent on HCE per capita, with a similar distribution between health care systems. Saturation appeared in countries with a team density greater than 10 and the correlation of transplant rates with HCE and GNP per capita was ‘limited to countries with low economic strength’.

Against this backdrop, it is hardly surprising that SCT is not a prevalent health care intervention in low income countries. Consequently, reports of associated costs of SCT in these settings are few and the numbers of patients are small. Ruiz-Arguelles19 and Gomez-Almaguer20 have led the development of non-myeloablative conditioning regimens in Mexico. While such transplants are apparently relatively inexpensive, as a result of being undertaken in the outpatient setting,19 the determination of costs may not be fully comprehensive and important sequelae, such as life-threatening acute and chronic graft-versus-host disease are common and have not been subject to economic evaluation.19 Experience in Thailand, on 67 adult patients with acute myeloid leukemia who received allogeneic or autologous SCT with peripheral blood or bone marrow, identified costs in the first year of US$20 000–US$25 000,21 not very different from those recorded in Mexico. In Egypt, the average cost of SCT, in 16 patients with a variety of disorders, was 75 000 Egyptian pounds,22 equivalent to US$12 500. The challenges related to funding for SCT in developing countries are compounded by inadequate reimbursement (see accompanying article by Seber). Even in one of the world’s wealthiest countries, Norway, the gap between the mean cost of SCT (US$106 825) and the funds provided on average by the government (US$36 404) exceeded US$70 000 per transplant; and that more than a decade ago.23

Value for Money

In a meta-analysis of the cost-effectiveness of treatment for adults with acute lymphoblastic leukemia in first complete remission, a statistically significant survival advantage for allogeneic transplantation over autologous transplantation and chemotherapy alone was demonstrated.24 Moreover, the cost per life year gained was less than the conventional European threshold of €50 000. Even this excellent study did not attempt a cost-utility analysis; an important goal in view of the considerable burden of morbidity in survivors of allogeneic SCT, both in childhood25 and adult life.26 So what constitutes value for money in SCT?27 A perspective from the USA28 had the following conclusion: ‘Estimates of the economic value of a life year vary widely from $24,777 to $428,286. However, for purposes of gauging the economic benefits of health interventions, estimates in the range of $100,000 to $160,000 are appropriate’.28 Whether such estimates from 10 years ago remain appropriate in the current global financial crisis, and to what extent these are relevant elsewhere in the world, especially in countries with limit resources, is a matter for proper debate.