Cost of chemotherapy-induced ... - Wiley Online Library

1541

Cost of Chemotherapy-Induced Thrombocytopenia among Patients with Lymphoma or Solid Tumors Linda S. Elting, Dr.P.H.1 Scott B. Cantor, Ph.D.1 Charles G. Martin, Ph.D.2 Lois Hamblin, R.N.1 Danna Kurtin, Ph.D.1 Edgardo Rivera, M.D.3 Saroj Vadhan-Raj, M.D.4 Robert S. Benjamin, M.D.5 1

The Section of Health Services Research, Department of Biostatistics, The University of Texas M. D. Anderson Cancer Center, Houston, Texas.

2

Department of General Internal Medicine, The University of Texas M. D. Anderson Cancer Center, Houston, Texas.

3

Department of Breast Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas.

4

Department of Bioimmunotherapy, The University of Texas M. D. Anderson Cancer Center, Houston, Texas.

5

Sarcoma Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas.

BACKGROUND. The purpose of this study was to estimate the mean incremental cost of chemotherapy-induced thrombocytopenia and the drivers of cost. Another goal was to estimate the impact of depth and duration of thrombocytopenia on the cost of thrombocytopenia. METHODS. A retrospective cohort, consisting of a random sample of 75 solid tumor or lymphoma patients who developed chemotherapy-induced thrombocytopenia (ⱕ 50,000 platelets per ␮l), was chosen. The number of each type of resource used during 217 cycles with and 300 cycles without thrombocytopenia were multiplied by the cost of each resource and summed to yield the total cost of care (in 1999 dollars from the provider’s perspective). RESULTS. Compared with cycles without thrombocytopenia, the mean incremental cost of thrombocytopenia was $1037 per cycle. However, 60% of cycles were usual cost, with a mean cost of thrombocytopenia of $43 per cycle less than control cycles. Twelve percent of cycles were high cost (mean incremental cost ⫽ $612 per cycle); 28% were very high cost (mean incremental cost ⫽ $3519). The excess cost during high-cost cycles was due to high consumption of prophylactic platelet transfusions and during very high-cost cycles to both higher platelet transfusion consumption and to a high incidence of major bleeding episodes. CONCLUSIONS. Although thrombocytopenia is a common complication of chemotherapy, only 40% of cycles with thrombocytopenia would be considered high or very high cost. Interventions targeted at this subset of cycles could significantly reduce the cost of thrombocytopenia provided they are initiated early enough in the chemotherapy experience to be effective. Cancer 2003;97:1541–50. © 2003 American Cancer Society. DOI 10.1002/cncr.11195

KEYWORDS: thrombocytopenia, cost of care, cost of illness, incremental costs, complications of chemotherapy, platelet transfusions.

Supported in part by a grant from Pharmacia, Inc. Dr. Elting has received previous funding from Genetics Institute, the manufacturer of interleukin11. Address for reprints: Linda S. Elting, Dr.P.H., Chief, Section of Health Services Research, Department of Biostatistics, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd. (Box 196), Houston, TX 77030-4009; Fax: (713) 792-7990; E-mail: [email protected] Received May 22, 2002; revision received October 21, 2002; accepted October 24, 2002. © 2003 American Cancer Society

T

he relation between platelet counts and bleeding was first described in leukemia patients in 1962 and, later, in solid tumor patients in 1978 and 1984.1–3 Interest in chemotherapy-induced thrombocytopenia has increased recently, fueled, in part, by platelet transfusion shortages, controversies about appropriate platelet thresholds for prophylaxis,4 –17 the development of platelet growth factors,18 –22 and the need to control the cost of healthcare. We recently described the epidemiology and outcomes of chemotherapyinduced thrombocytopenia among patients with lymphoma and solid tumors.23 However, with a few exceptions, the literature is silent on the cost associated with this common complication of chemotherapy among patients with lymphoma or solid tumors.24,25 Accordingly, we examined the cost of care of patients with lymphoma or solid tumors

1542

CANCER March 15, 2003 / Volume 97 / Number 6

during chemotherapy cycles with and without thrombocytopenia. The four objectives of the study were to: 1) estimate the mean cost of chemotherapy cycles with and without thrombocytopenia, 2) identify the specific resources that “drive” cost, 3) estimate the incremental cost of thrombocytopenia compared with cycles without thrombocytopenia, and 4) estimate the financial impact of reductions in the duration of thrombocytopenia.

MATERIALS AND METHODS A random sample of 75 adult patients, stratified by underlying cancer, was selected from among all patients with solid tumors or non-Hodgkin lymphoma who developed chemotherapy-induced thrombocytopenia between January 1, 1995 and December 31, 1996. Among these, 41% were males, 56% had metastatic disease, 15% had bone marrow metastases, and 3% had a Zubrod performance status ⬎ 2. Their median age was 51 years. Lymphomas and breast cancers predominated (23% each), genitourinary cancers and sarcomas were present in 11% each, and lung, gastrointestinal, gynecological cancers, and melanomas were present in ⬍ 10% each. To ensure complete ascertainment of all healthcare resources consumed, only those cycles during which the patient resided in the Houston area and all care was provided at our institution were included. Patients with leukemia, bone marrow transplant and stem cell support recipients, and those receiving platelet growth factors were excluded. For each of the 75 eligible patients, all cycles complicated by thrombocytopenia between January 1994 and December 1996 were included. These 217 cycles are designated hereafter as “TCP cycles.” Previous studies of the cost of thrombocytopenia have shared a significant limitation: a lack of controls. In a study without a control population, the costs of bleeding and platelet transfusions can be readily attributed to thrombocytopenia, but the costs of monitoring, which could be attributed to follow-up for multiple toxicities, cannot be allocated to thrombocytopenia. Therefore, the total costs of thrombocytopenia cannot be estimated. The strength of our study is the inclusion of controls to facilitate estimation of the incremental costs of thrombocytopenia compared with controls without thrombocytopenia. To reduce variation in cost from differing patient or treatment characteristics, we chose, as control cycles, 300 cycles that occurred in the same patients who received the same chemotherapy regimens but were not complicated by thrombocytopenia. These are designated as “control cycles.” Fifty-seven percent of con-

trol cycles occurred before the patients’ first TCP cycles and the remainder after the first TCP cycle. Although the chemotherapy regimens were the same, the doses were occasionally different and, therefore, not necessarily equally myelosuppressive. However, we consider control cycles in the same patients on the same regimen to be more informative than cycles among different patients or those on other regimens. Because the TCP and control cycles occurred in the same patients, the two groups were very similar for most demographic and clinical status characteristics. However, more TCP cycles occurred during the metastatic phase of the patients’ clinical course, whereas control cycles occurred more frequently during the local phase (P ⫽ 0.04). Despite this difference, bone marrow metastases (13% and 15%, respectively) and poor performance status (8% and 6%, respectively) were similarly frequent in the two groups.

Data Collection The paper and electronic medical and billing records of eligible patients were examined for clinical information and for resource utilization. Data from paper sources were transcribed by physician and nurse abstractors, and data from databases were transferred electronically.

Definitions Thrombocytopenia was defined as a platelet count ⬍ 50,000 per ␮l. Bleeding episodes were dichotomized into minor episodes (World Health Organization [WHO] Grades 1 or 2, including petechiae, ecchymoses, superficial bleeding of gums, microscopic hematuria, blood-tinged sputum, mild epistaxis, and vaginal bleeding not requiring red cell transfusion) or major episodes (WHO Grades 3 or 4, including fatal hemorrhage at any site or epistaxis, vaginal bleeding, or major organ hemorrhage requiring red cell transfusion). Bleeding developed during 14 episodes; 13 episodes occurred during TCP cycles, one of which was fatal, and 1 episode developed during a control cycle. The duration of thrombocytopenia was computed using the last value carried forward method as is typical in observational studies and clinical trials. Performance status was measured on Day 1 of each cycle using the Zubrod score.26 Poor performance status was defined as Zubrod score ⬎ 2.

Measurement and Aggregation of Costs Cost was measured in 1999 dollars from the provider’s perspective. Costs were computed by multiplying the cost of individual resources (obtained from our hospital accounting system) by the number of resources

Cost of Thrombocytopenia/Elting et al.

1543

FIGURE 1. Components of cost studied, aggregation of components, and representative unit costs. consumed (obtained from patients’ medical and billing records). Incremental costs were computed by subtracting the cost of care during control cycles from the cost of care during TCP cycles. However, this does not yield the incremental cost attributable to thrombocytopenia. TCP cycles also were characterized by granulocytopenia, which occasionally led to growth factor administration, anemia that led to red cell transfusion and growth factor administration, and febrile neutropenia that often led to hospitalization. These costs were unrelated to TCP. Therefore, the “attributable” incremental cost of thrombocytopenia was calculated by adding the costs of monitoring, bleeding prophylaxis, and bleeding treatment for each cycle, then subtracting the cost during control cycles from the cost during TCP cycles. For bleeding prophylaxis and treatment, this yields the cost directly attributable to thrombocytopenia. For monitoring, this strategy yields only an estimate of the portion of the incremental cost for each cycle that may be attributable to thrombocytopenia because monitoring blood counts and clinic visits are done to detect multiple toxicities of chemotherapy. For most analyses, costs were aggregated into clinically meaningful categories, including the cost of chemotherapy, monitoring cost, bleeding prophylaxis cost, bleeding treatment cost, and cost of all other complications (Fig. 1). However, for the analysis of cost “drivers,” individual resources were disaggregated. For example, the cost of bleeding treatment was

disaggregated into the number of hospital days, ICU days, platelet units, RBC units, procedures, visits to the Emergency Department, laboratory tests, and pharmaceuticals. Costs of key resources are included in Figure 1. The cost of platelet units varied depending on the type of product used. The costs of random and single donor platelets were $119 and $181 per unit, respectively. Professional fees were not available in the hospital accounting system and were, therefore, not included in the analysis. A goal of this study was to identify “high-cost” cycles. Therefore, the TCP cycles were divided into three groups based on their attributable incremental costs (the incremental costs of monitoring, bleeding prophylaxis, and bleeding treatment) compared with control cycles. Cycles with an attributable cost ⬍ 1 standard deviation higher than the mean attributable cost among Control cycles were considered “usual cost.” Those with an attributable cost ⬎ 1 and ⬍ 3 standard deviations higher than control cycles were considered “high cost.” The remaining cycles, with attributable costs ⱖ 3 standard deviations higher than control cycles were considered “very high cost.” Because of the importance of depth of thrombocytopenia in determining clinical outcomes, costs were categorized by several clinically important thresholds. Mean and incremental costs per day between 20,000 and 50,000 platelets were compared with those incurred per day between 10,000 and 20,000

1544


FIGURE 2. Comparison of the distribution of cost components between cycles with (TCP cycles) and without (control cycles) thrombocytopenia. platelets (at which point transfusions may be initiated) and those ⬍ 10,000 platelets (at which point the risk of spontaneous bleeding increases). Mean and incremental costs per day also were computed based on the ultimate nadir of platelet count for each cycle.

Statistical Considerations Differences between proportions were examined using Pearson chi square test (or Fisher exact test, when appropriate). Differences between means were examined for statistical significance using Student t test. Stepwise multiple linear regression was used to identify specific cost components that “drive” total cost, using total cost as the dependent variable and the number of individual resources consumed as independent variables. All statistical analyses were done using BMDP Dynamic (BMDP Statistical Software, Inc., Los Angeles, CA).

RESULTS The estimated mean total cost of care during TCP cycles ($6866) exceeded that observed during control cycles ($4875) by 40% (Fig. 2). The absolute costs of chemotherapy agents and administration and monitoring after chemotherapy were virtually identical for TCP and control cycles. Management of bleeding, infections, and other complications accounted for more than 50% of the total cost of care during TCP cycles.

Drivers of Cost Considering only those costs attributable to thrombocytopenia (monitoring, bleeding prophylaxis, and bleeding therapy), platelet products accounted for 51%, hospitalization or Emergency Department visits 14%, and routine monitoring visits and blood counts accounted for 31% of attributable costs. These data

provide insight into potential targets for across-theboard reductions in the cost of thrombocytopenia. Stepwise multiple linear regression was used to assess the impact of individual cost components on the variance in cost. This analysis provides insight into the resources that contribute to high-cost episodes and that could be opportunities for targeted efforts to reduce the cost of very high-cost patients. Hospital days and Emergency Department visits for complications other than thrombocytopenia accounted for 59% and 7% of the variance in total cost, respectively, but are not the topic of this report (Table 1). Platelet units transfused prophylactically accounted for 15% of the variance in total cost, whereas hospital days and Emergency Department visits for bleeding accounted for 8% and 6%, respectively. Although routine monitoring visits and blood counts accounted for 31% of the total cost attributable to thrombocytopenia, their costs were very stable across cycles and contributed very little to the variance in cost from one cycle to another. Other individual resources, such as units of platelets or red blood cells transfused for bleeding and ICU days, also contributed very little to the variance in total cost.

Incremental Cost Attributable to Thrombocytopenia The estimated mean costs of chemotherapy ($2365 and $2206, respectively) and monitoring following chemotherapy ($458 and $436, respectively) were similar for TCP cycles and control cycles (Table 2). However, the costs of managing bleeding ($237 and $14, respectively) and other complications ($3014 and $2219, respectively) were significantly higher during TCP cycles compared with control cycles (P ⫽ 0.03 and 0.04, respectively). Bleeding prophylaxis added an average of $792 to the cost of TCP cycles. The mean incremental cost of TCP cycles was $1991 compared with control cycles (Table 2). However, only $1037 (the incremental costs of monitoring, bleeding prophylaxis, and bleeding treatment) is attributable to thrombocytopenia. The additional costs of chemotherapy ($159) were primarily due to the addition of granulocyte or erythrocyte growth factors when pancytopenia was expected. Granulocytopenia was largely responsible for the incremental cost ($795) in other complications (primarily febrile neutropenia). Although the overall mean attributable incremental cost of TCP was $1037 per cycle, 60% of the TCP cycles were categorized in the usual cost group (as defined in Materials and Methods), with a mean attributable cost of $43 (95% CI; ⫺$63, ⫺$24) per cycle less than the control cycles (Fig. 3). The 26 (12%) high-cost cycles had a mean attributable incremental


1545

TABLE 1 Impact of Individual Components, “Cost Drivers,” on Total Cost of Care

Component of cost

% of total cost per cycle

Step entered in regression model

Multiple R2 (% of variance explained by the cumulative model)

Hospital days — all other complications Prophylactic platelet units Hospital Days — bleeding ED visits — all other complications ED visits — bleeding Chemotherapy All other resources

39% 11% 2% 5% 1% 34% 8%

1 2 3 4 5 6 7–12

0.59 0.74 0.82 0.89 0.95 0.99 0.9999

Change in R2 (% of variance explained by the component) 0.59 0.15 0.08 0.07 0.06 0.04 0.01

ED: Emergency Department; of total cost: cost of component divided by total cost (univariate); multiple R2: % of variance in cost explained by the total model at that step (multivariate); change in R2: % of variance in cost explained by each cost component, having accounted for the impact of components entered at earlier steps (multivariate); all other resources: platelet counts, CBC counts, clinic visits, platelet units for bleeding, RBC units, ICU days.

TABLE 2 Estimated Incremental Cost of Thrombocytopenia

Chemotherapy Monitoring Bleeding prophylaxis Bleeding treatment All other complications Total Total Attributable

TCP cycles Mean (95% CI)

Control cycles Mean (95% CI)

Incremental cost of TCP (95% CI)

P value*

$2365 (1943–2787) $ 458 (436–480) $ 792 (561–1023) $ 237 (8–466) $3014 (2370–3659) $6866 (5830–7901) $1486 (1149–1824)

$2206 (1934–2479) $ 436 (409–463) $ 0 $ 14 (0–42) $2219 (1843–2595) $4875 (4391–5359) $ 450 (410–490)

$ 159 (⫺312–638) $ 22 (⫺12–59) $ 792 (596–988) $ 223 (26–420) $ 795 (93–1498) $1991 (945–3037) $1037 (888–1186)

0.53 0.21 0.001 0.03 0.04 0.0007 ⬍ 0.0001

* P value compares the mean cost of TCP and control cycles; CI: confidence interval; TCP: thrombocytopenia; Total attributable: monitoring cost ⫹ bleeding prophylaxis ⫹ bleeding treatment; Incremental cost: mean cost (TCP cycles) ⫺ mean cost (control cycles).

cost of $612 (95% CI; $546 –$678) per cycle, due largely to the cost of prophylactic platelet transfusions. The 61 (28%) very high-cost cycles had even higher costs for prophylactic platelet transfusions but in addition had significantly higher incremental costs of bleeding treatment.

Potential Savings from Reduction in the Depth or Duration of Thrombocytopenia The previous data suggest that a reduction in the duration of thrombocytopenia likely would have little effect on the overall cost of care of solid tumor and lymphoma patients during approximately 60% of cycles with TCP (those in the usual cost group). Although only 40% of TCP cycles had high or very high attributable incremental costs, reduction of costs in this portion of the population could lead to considerable savings if interventions could be targeted to this population. For that reason, we examined differences in the clinical characteristics of the patients during these three groups of cycles.

FIGURE 3. Distribution of the mean cost components among control, high, and usual cost cycles. High and usual cost cycles differed in both the depth and duration of thrombocytopenia. Only 12% of usual cost cycles had a nadir ⬍ 20,000 platelets compared with 92% of high-cost cycles (P ⬍ 0.0001) (Table 3). The mean duration of thrombocytopenia

1546

CANCER March 15, 2003 / Volume 97 / Number 6 TABLE 3 Characteristics of Patients during Control, High, and Usual Cost Cycles TCP cycles

Platelet nadir (%) ⬎ 20,000 10,000–20,000 ⬍ 10,000 TCP duration (mean days) Platelet transfusions (%) Transfusion episodes (mean per cycle) Minor bleeding (%) Major bleeding (%) Cancer (%) Lymphoma GU Sarcoma Breast Lung Others

Control cycles (n ⴝ 300)

Usual cost (n ⴝ 130)

High cost (n ⴝ 26)

Very high cost (n ⴝ 61)

0 ⬍1 0 0 ⬍1

88 10 2 5.6 2 0.2 1 0

8 77 15 7.5 96 1.4 4 0

11 41 48 11.0 95 2.9 7 10

19 15 10 29 5 22

17 12 9 37 7 18

31 23 12 8 4 22

28 25 20 12 2 13

TCP: thrombocytopenia; GU: genitourinary.

TABLE 4 Mean Cost and Incremental Cost Per Day of Thrombocytopenia TCP cycles

Mean cost per day ⬍ 50,000 Incremental costa per day ⬍ 50,000 a

Nadir 20,000–50,000 (n ⴝ 123)

Nadir 10,000–20,000 (n ⴝ 59)

Nadir < 10,000 (n ⴝ 35)

$93 $76

$214 $197

$369 $352

Mean cost per day of control cycles, without thrombocytopenia ⫽ $17; Incremental cost ⫽ mean cost per day of TCP cycle ⫺ mean cost per day of control cycles.

was ⬍ 1 week (5.6 days) for usual cost cycles and exceeded 1 week (7.5 days) for high-cost cycles (P ⫽ 0.004). Very high-cost cycles differed from highcost cycles in that the duration of thrombocytopenia was significantly longer (11.0 days and 7.5 days, respectively, P ⫽ 0.02). A nadir ⬍ 10,000 was observed in 48% of very high-cost cycles compared with 15% of high-cost cycles (P ⫽ 0.01). These statistically significant differences in the depth and duration of thrombocytopenia were reflected in the clinically significant difference in the incidence of major bleeding episodes during very high-cost cycles compared with high-cost cycles (10% vs. 0%, P ⫽ 0.02). As might be expected, high- and very highcost cycles were more common among patients with lymphoma and sarcoma who receive high doses of myelosuppressive therapy. They also were more

common among patients with genitourinary malignancies who are high risk for bleeding episodes. Because of the significant difference in platelet nadirs experienced by patients during usual and high-cost cycles, we examined the mean attributable costs and incremental costs per day for several clinically important platelet thresholds. The mean attributable cost per day ⬍ 50,000 platelets varied, depending on the ultimate platelet nadir, from a low of $93 per day during cycles with a nadir between 20,000 and 50,000 platelets to a high of $369 per day during cycles with a nadir ⬍ 10,000 platelets (Table 4). As would be expected, days ⬍ 10,000, at which point spontaneous bleeding would be expected, were significantly more costly on average ($875) than those at higher levels. Compared with a mean attributable cost of $17 per day for control cycles,


the incremental cost during TCP cycles varied from $76 to $352 per day.

TABLE 5 Comparison of Published Estimates of the Cost of Platelet Transfusions

DISCUSSION Published information about the costs associated with chemotherapy-induced thrombocytopenia is limited. A few reports have described costs among patients with leukemia or bone marrow transplant recipients.27,29 These populations are significant because of their high risk of bleeding and their high platelet transfusion requirements per cycle. Other reports have described the platelet transfusion requirements for solid tumor patients undergoing surgery but do not address platelet transfusions during thrombocytopenia in this population.30 –32 However, despite the typically low number of platelet transfusions per cycle among patients with solid tumors or lymphoma, the large size of the total population makes management of their episodes of thrombocytopenia an economically significant issue from a public policy perspective and a subject worthy of study.33 The mean incremental cost of thrombocytopenia among our patients with solid tumors or lymphoma ($1037) was comparable to that reported by Malone in 1995 ($1035),24 whereas Calhoun et al. reported an estimate of direct costs almost triple those observed in our study ($3268).25 However, it is not clear that Calhoun’s estimate applies to a cycle of chemotherapy because follow-up appears to have continued for up to 3 months. Although the populations and methodologies differ somewhat, when these estimates are standardized to 1999 dollars, our costs remain very similar to Malone’s ($1037 vs. $1288) but lower than Calhoun’s ($3268). However, the similarity of our results to those reported by Malone (from a very large study) underscores the generality of our results and argues against the primary limitation of our study, which is the inclusion of patients from only a single center. The retrospective nature of the data collection also could have resulted in incomplete ascertainment of resource utilization and bleeding episodes and a corresponding underestimate of the incremental cost attributable to thrombocytopenia. The similarity of our cost estimates to previous estimates from large studies also relieves these concerns. Although the limitations of our study may have caused an underestimate of the true cost of thrombocytopenia, the unique chemotherapy regimens used in some patients at our institution may have caused overestimates in some subsets because of the intensity of the regimens. The similarity of our estimates of cost to Malone’s suggests that the magnitude of any overestimation is small.

1547

Author

Year

Reported cost per platelet transfusion

Cost in 1999 dollarsa

McCullough32 Balducci35 Lill27 Ackerman36 Snider34 Lopez-Plaza37 Kaushansky21 This report

1988 1992 1995 1996 1996 1996 1998 1999

$138c $304c $712b $536 $551 $313c $300c $590

$250c $401c $808b $589 $605 $344c $311c $590

a

Costs standardized to 1999 dollars using the consumer price index for medical care from the U.S. Bureau of Labor Statistics (available at www.bls.gov). b Result of time in motion study. All other estimates were obtained from hospital accounting systems. c Cost of platelet products only. No supplies, premedication, administration, or other costs included. All other estimates include supplies and administration costs.

Although very high-cost episodes included hospital days for management of bleeding, the single most significant contributor to incremental costs was platelet transfusions. For that reason, estimates of incremental costs would be expected to be sensitive to changes in the cost of a transfusion. For this study we used a conservative estimate of the cost of a platelet transfusion, $590. This estimate compares favorably to previously published estimates of the total cost of transfusion when these are standardized to 1999 dollars (Table 5). If only those estimates that include all transfusion costs are considered, our estimate is among the lowest ($589 –$808). This suggests that our data provide conservative estimates of the incremental cost of thrombocytopenia. If the range in previously published transfusion cost estimates is the result of variation in transfusion costs across institutions, the incremental cost of thrombocytopenia for high- and very high-cost cycles could be higher at other centers. For example, substituting a cost of $808 for platelet transfusions in our study would increase the mean incremental cost of high- and very high-cost cycles an average of $3006 and $10,264 per cycle, respectively. However, it should be noted that the mean attributable cost of usual cost cycles would still be $2 less than control cycles, owing to the small percentage of patients that required transfusion during these cycles. Our costs may be underestimated for other reasons, as well. Professional fees were not included, nor were the personnel costs of telephone follow-up conversations with outpatients. We addressed only the direct medical costs of thrombocytopenia. Calhoun et

1548


al. observed that the indirect costs of thrombocytopenia actually exceeded the direct medical costs.25 Finally, the additional costs of refractoriness to platelet transfusions (which may have been incurred during cycles before or after the 2-year study period) are not reflected in our cost estimates. These costs can be substantial, and up to 20% of patients may become refractory to platelet transfusion during the course of therapy.31,38,39 Meehan reported a two-fold increase in the cost of hospitalization of platelet refractory patients compared with those who received platelet transfusion but were not refractory. In that prospective study of costs and platelet transfusion utilization, patients with refractory status accounted for 66% of all platelet transfusions.31 For this reason, the incremental cost of thrombocytopenia may be underestimated in our study.

Implications for Cost Containment There are two options for cost-containment interventions: 1) across-the board interventions applied to every cycle, aimed at reducing the cost of every episode of thrombocytopenia, 2) interventions targeted to the highest cost episodes. Our data suggest that in 60% of chemotherapy cycles complicated by thrombocytopenia, efforts at cost reduction would have little impact. Because of the transient and limited nature of the thrombocytopenia, prophylactic transfusions are uncommon and the risk of bleeding is no greater than among cycles without thrombocytopenia in the same patients. The costs of monitoring are remarkably similar for usual cost cycles and for those without thrombocytopenia, and it is unlikely that monitoring would be reduced even if thrombocytopenia did not occur. These findings argue against across-the-board interventions. Conversely, the remaining cycles are associated with prolonged and profound thrombocytopenia with correspondingly high risks of bleeding and high requirements for prophylactic platelet transfusion. Interventions targeted at these cycles could have a significant impact on the overall cost of thrombocytopenia. However, the success of a targeted strategy would depend on overcoming two potential barriers: 1) the safety of confining prophylaxis to high-cost cycles, 2) a sensitive and specific method for identifying high-cost cycles. Because the cost of platelet transfusions is the most significant distinguishing characteristic of highand very high-cost cycles, interventions to reduce the need for transfusions are the obvious choices for targeted interventions. Such interventions could include dose reductions or platelet growth factors. However, the safety and efficacy of such strategies are not clin-

ically obvious. Dose reductions may reduce antitumor effect in some cancers.40 In the case of growth factors, because thrombocytopenia tends to be cumulative and growth factors are most effective when used in patients with adequate bone marrow reserve, it is probably most efficacious to identify patients at risk for high-cost cycles early in their chemotherapy experience during low-cost cycles. Limiting prophylaxis to high-risk cycles, when bone marrow reserves are inadequate, could make management more difficult. The safety of such an approach must be evaluated in a prospective trial. Assuming that a targeted strategy was proven to be safe, it would still be necessary to devise a method for identifying high- and very high-cost cycles in advance. Although not reported here, we have developed a clinical prediction rule, the Bleeding Risk Index, which identifies patients at high risk of bleeding on Day 1 of a chemotherapy cycle. Factors associated with bleeding during the subsequent cycle include a prior history of bleeding, a Day 1 platelet count ⬍ 75,000, the presence of bone marrow metastases, poor performance status, genitourinary or gynecological malignancy, therapy during the cycle with agents that affect platelet function (heparin, nonsteroidal anti-inflammatory agents, penicillins, cephalosporins, thiazide diuretics, and histamine type 2 antagonists), and chemotherapy during the cycle with agents highly toxic to the bone marrow (cisplatin, mitomycin C, dacarbazine, and lomustine).41 It is possible that this index or an adaptation of it could be helpful in targeting patients and chemotherapy cycles for cost-reduction interventions.

Implications for the Cost-Effectiveness of Platelet Growth Factors During the last decade, platelet growth factors have been developed with the goal of mitigating the effects of thrombocytopenia.42 Two such agents, oprelvekin and thrombopoietin, have been extensively studied, and the former is commercially available in the United States. In the studies reported to date, platelet growth factors have reduced rather than eliminated the need for prophylactic platelet transfusions.18 –21 Although many of these studies involved mandatory transfusion at a threshold of 20,000 platelets, a conservative trigger for platelet transfusion by today’s standards of practice, it is unlikely that a lower threshold would have completely eliminated the need for transfusions. However, even a reduction in the amount of prophylactic platelet transfusions would have a significant impact on the financial bottom line for the high- and very high-cost cycles reported in this study. If a tar-


geted strategy is proven to be safe and the Bleeding Risk Index or some other clinical algorithm can be used to identify these high-cost cycles, it is possible that platelet growth factors that are clinically effective will also be cost-effective at prices typical of growth factors.

REFERENCES 1.

2.

3.

4.

5.

6. 7.

8. 9. 10.

11.

12.

13.

14.

15.

16. 17.

18.

Gaydos LA, Freireich EJ, Mantel N. The quantitative relation between platelet count and hemorrhage in patients with acute leukemia. N Engl J Med. 1962;266:905–909. Belt RJ, Leite C, Haas CD, et al. Incidence of hemorrhagic complications in patients with cancer. JAMA. 1978;239:2571– 2574. Dutcher JP, Schiffer CA, Aisner M, et al. Incidence of thrombocytopenia and serious hemorrhage among patients with solid tumors. Cancer. 1984;53:557–562. Goldberg GL, Gibbon DG, Smith HO, et al. Clinical impact of chemotherapy-induced thrombocytopenia in patients with gynecologic cancer. J Clin Oncol. 1994;12:2317–2320. Fanning J, Hilgers RD, Murray KP, et al. Conservative management of chemotherapy-induced thrombocytopenia in women with gynecologic cancers. Gynecol Oncol. 1995;59: 191–193. Beutler E. An iconoclastic view of conventional wisdom in hematology. Arch Intern Med. 1979;139:221–223. Baer MR, Bloomfield CD. Controversies in transfusion medicine—Prophylactic platelet transfusion therapy: pro. Transfusion. 1992;32:377–380. Schiffer CA. Prophylactic platelet transfusion [editorial]. Transfusion. 1992;32:295–298. Beutler E. Platelet transfusions: the 20,000/␮L trigger. Blood. 1993;81:1411–1413. Pisciotto PT, Benson K, Hume AB, et al. Prophylactic versus therapeutic platelet transfusion practices in hematology and/or oncology patients. Transfusion. 1995;35:498 –502. Norfolk DR, Ancliffe PJ, Contreras M, et al. Concensus conference on platelet transfusion, Royal College of Physicians of Edinburgh. B J Haemat. 1998;101:609 – 617. Heckman K, Weiner G, Davis C, et al. Randomized study of prophylactic platelet transfusion threshold during induction therapy for adult acute leukemia: 10,000/ml versus 20,000/ ml. J Clin Oncol. 1997;15:1143–1149. Rebulla P, Finazzi G, Marangoni F, et al. The threshold for prophylactic platelet transfusions in adults with acute myeloid leukemia. N Engl J Med. 1997;337:1870 –1875. Wandt H, Frank M, Ehninger G, et al. Safety and cost effectiveness of a 10 x 109/L trigger for prophylactic platelet transfusions compared with 20 x 109/L trigger: a prospective comparative trial in 105 patients with acute myeloid leukemia. Blood. 1998;91:3601–3606. Schiffer CA, Anderson KC, Bennett CL, et al. Platelet transfusion for patients with cancer: clinical practice guidelines of the American Society of Clinical Oncology. J Clin Oncol. 2001;19:1519 –1538. McCullough J. Current issues with platelet transfusion in patients with cancer. Semin Hematol. 2000;37(Suppl 4):3–10. Rinder H, Arbini A, Snyder E. Optimal dosing and triggers for prophylactic use of platelet transfusions. Curr Opin Hematol. 1999;6:437– 441. Tepler I, Elias L, Smith II JW, et al. A randomized placebocontrolled trial of recombinant human interleukin-11 in

19.

20.

21. 22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

1549

cancer patients with severe thrombocytopenia due to chemotherapy. Blood. 1996;87:3607–3614. Isaacs C, Robert NJ, Bailey FA, et al. Randomized placebocontrolled study of recombinant human interleukin-11 to prevent chemotherapy-induced thrombocytopenia in patients with breast cancer receiving dose-intensive cyclophosphamide and doxorubicin. J Clin Oncol. 1997;15:3368 – 3377. Basser R. Clinical use of thrombopoietic growth factors. In: American Society of Clinical Oncology Educational Book. Alexandria, VA: American Society of Clinical Oncology, 1999: 266 –274. Kaushansky K. Drug therapy: thrombopoietin. N Engl J Med. 1998;339:746 –754. Elting LS. Integrating prophylactic measures into practice guidelines for platelet growth factors. Oncology. 1999;13: 587–589. Elting LS, Rubenstein EB, Martin CG, et al. Incidence, cost and outcomes of bleeding and chemotherapy dose modification among solid tumor patients with chemotherapy-induced thrombocytopenia. J Clin Oncol. 2001;19:1137–1146. Malone D, Sullivan S, Black D, et al. The cost of treating chemotherapy-induced thrombocytopenia. Proceedings of ASCO [abstract]. J Clin Oncol. 1995;14:305. Calhoun E, Chang C-H, Welshman E, et al. Evaluating the total costs of chemotherapy-induced toxicity: results from a pilot study with ovarian cancer patients. Oncologist. 2001;6: 441– 445. Zubrod CG, Schneiderman M, Frei E, et al. Appraisal of methods for the study of chemotherapy of cancer in man: comparative trial of nitrogen mustard and trimethylene thiophosphoramide. J Chron Dis. 1960;11:7–33. Lill M, Snider C, Calhoun L, et al. Epidemiology and cost analysis of platelet transfusions at a tertiary care medical center. Blood Cell and Bone Marrow Transplantation Symposium, Keystone CO, January 1996 [abstract]. Hartmann O, Corroller A, Blaise D, et al. Peripheral blood stem cell and bone marrow transplantation for solid tumors and lymphomas: hematologic recovery and costs: a randomized, controlled trial. Ann Intern Med. 1997;126:600 – 607. Bernstein SH, Nademanee AP, Vose JM, et al. A multicenter study of platelet recovery and utilization in patients after myeloablative therapy and hematopoietic stem cell transplantation. Blood. 1998;91:3509 –3517. Jefferies L, Sachais B, Young D. Blood transfusion costs by diagnosis-related groups in 60 university hospitals in 1995. Transfusion. 2001;41:522–529. Meehan K, Matias C, Rathore S, et al. Platelet transfusions: utilization and associated costs in a tertiary care hospital. Am J Hemat. 2000;64:251–256. McCullough J, Steeper T, Connelly D, et al. Platelet utilization in a university hospital. J Am Med Assoc. 1988;259:2414 – 2418. Zimmerman R, Buscher M, Linhardt C, et al. A survey of blood component use in a German university hospital. Transfusion. 1997;37:1075–1083. Snider C, Erder H, LaBrecque J, et al. What are the true costs of platelet transfusions? A prospective time motion study of resource utilization associated with platelet transfusions at UCLA Medical Center [abstract]. Blood. 1996;88:333a.

1550


35. Balducci L, Benson K, Lyman G, et al. Cost-effectiveness of white cell-reduction filters in treatment of adult acute myelogenous leukemia. Transfusion. 1993;33:665– 670. 36. Ackernan S, Klumpp T, Guzman G, et al. Economic consequences of alterations in platelet transfusion dose: analysis of a prospective, randomized, double-blind trial. Transfusion. 2000;40:1457–1462. 37. Lopez-Plaza I, Weissfeld J, Triulzi D. The cost-effectiveness of reducing donor exposures with single-donor versus pooled random-donor platelets. Transfusion. 1999;39:925– 932. 38. Freedman J, Gafni A, Garvey M, et al. A cost-effectiveness evaluation of platelet crossmatching and HLA matching in

39. 40.

41.

42.

the management of alloimmunized thrombocytopenic patients. Transfusion. 1989;29:201–207. Webb I, Anderson K. Risks, costs, and alternatives to platelet transfusions. Leukemia Lymphoma. 1999;34:71– 84. Savarese DM, Hsieh C, Stewart FM. Clinical impact of chemotherapy dose escalation in patients with hematologic malignancies and solid tumors. J Clin Oncol. 1997;15:2981–2995. Elting L, Martin C, Cantor S, et al. A clinical prediction rule to guide the use of prophylactic platelet growth factors and platelet transfusions. Cancer. 2002;94:3252–3262. Kuter D, Cebon J, Harker L, et al. Platelet growth factors: potential impact on transfusion medicine. Transfusion. 1999;39:321–332.