- Email: [email protected]
The impact of delayed chemotherapy-induced nausea and vomiting on patients, health resource utilization and costs in German cancer centers
Original article
Annals of Oncology 15: 526–536, 2004 DOI: 10.1093/annonc/mdh110
The impact of delayed chemotherapy-induced nausea and vomiting on patients, health resource utilization and costs in German cancer centers A. Ihbe-Heffinger1*, B. Ehlken2, R. Bernard1, K. Berger2, C. Peschel3, H.-G. Eichler4,5, R. Deuson4, J. Thödtmann1 & F. Lordick3 1
Department of Pharmacy, Klinikum rechts der Isar, Technische Universität München; 2MERG, Medical Economics Research Group, Munich; Third Medical Department, Klinikum rechts der Isar, Technische Universität München, Germany; 4Merck & Co., Inc., Whitehouse Station, NJ, USA; 5 Department of Clinical Pharmacology, University of Vienna Medical School, Vienna, Austria 3
Background: Delayed chemotherapy-induced nausea and vomiting (CINV) remains a major adverse effect of cancer chemotherapy. We assessed, under current practice patterns, the occurrence and impact on healthcare resource utilization of CINV in patients receiving emetogenic chemotherapy. An additional aim of this study was to estimate costs imputable to CINV in the German healthcare environment. Materials and methods: This prospective, multi-center, cross-sectional cost-of-illness study was conducted in three hospitals and in three office-based facilities in Germany. Consecutive patients undergoing emetogenic chemotherapy (levels 4 or 5 according to Hesketh classification of emetogenicity) were enrolled. Data were obtained from preplanned chart reviews and from self-administered patient questionnaires. Analysis of direct costs was performed from the perspectives of third party payer (statutory sick fund), provider (hospital) and patients. Indirect costs were assessed on the basis of paid workdays lost. Results: During the 5-day observation period, 134 of 208 chemotherapy cycles observed (64.4%) were associated with at least one episode of nausea or vomiting. More patients experienced delayed than acute CINV (60.7% versus 32.8%), and more patients reported nausea than vomiting (62.5% versus 26.0%). A total of 68 patients (32.6%) utilized healthcare resources due to CINV. The most frequently used resources were rescue medications and outpatient hospital and office physician visits. Only one patient required hospitalization and only three patients lost workdays due to CINV. Average costs imputable to CINV per patient (with or without CINV) per treatment cycle incurred by third party payers and hospital providers were $49 and $48, respectively. Patient or treatment characteristics that were associated with high costs imputable to CINV were as follows: cisplatincontaining regimen; experience of emesis; and presence of delayed CINV. Conclusions: A substantial proportion of patients continue to experience CINV. This entails not only clinical but also economic consequences, and highlights a continuing need for improved utilization of existing antiemetic agents and for new, more efficacious treatments. The greatest improvements in patient care and potential for cost offset may be realized by preventing delayed CINV. Key words: chemotherapy-induced nausea and vomiting, cost-of-illness, healthcare utilization, health economics, healthcare provider, third party payer
Introduction Chemotherapy-induced nausea and vomiting (CINV) remains a major adverse effect of cancer chemotherapy, despite the availability of several antiemetic drug classes, including 5-HT3 receptor antagonists. Although not life-threatening, CINV has a major impact on a patient’s quality of life and ranks high on the list of factors most feared by patients receiving chemotherapy [1].
*Correspondence to: Dr A. Ihbe-Heffinger, Apotheke des Klinikums rechts der Isar, der Technischen Universität München, Ismaninger Strasse 22, 81675 München, Germany. Tel: +49-89-4140-2219; Fax: +49-89-4140-6365; E-mail: [email protected] © 2004 European Society for Medical Oncology
Aside from these clinical consequences, CINV has considerable economic impact. Direct costs include acquisition cost of antiemetic drugs and administration devices, additional patient care, such as nursing and physician time, and in some cases, extended hospitalization or readmission. CINV is often not confined to the acute period (developing within 24 h of receiving chemotherapy), but may manifest itself as delayed CINV (>24 h post-chemotherapy), which can persist for several days [2]. Therefore, indirect costs may result from reduced productivity at work or workdays lost. Economic analyses, including cost-of-illness studies, receive increasing attention from healthcare payers and providers to help inform choices in the delivery of care. Hence, quantifying the
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Received 27 August 2003; revised 28 November 2003; accepted 19 December 2003
527
Patients and methods Study design, setting and perspectives This prospective, multi-center, cross-sectional cost-of-illness study was conducted in three hospitals and in three office-based facilities in Germany. All centers were experienced in the administration of cancer chemotherapy. Centers were selected to allow for enrollment of a broad spectrum of patients requiring chemotherapy and to represent both hospital- and office-based cancer care settings which reflect current medical practice in Germany. Analysis of costs was performed from three perspectives: (a) the third party payer (‘Krankenkassen’ or statutory sick funds), (b) the provider (hospital) and (c) the patient.
Patient selection criteria Patients were eligible for inclusion if they were scheduled for treatment with emetogenic chemotherapy as defined by level 4 or 5 categorization on the emetogenicity scale proposed by Hesketh et al. [1], or with a level 4 or 5 combination chemotherapy according to the algorithm provided by Hesketh et al. [1]. All of these agents or combinations are associated with a ≥60% frequency of emesis [1]. The Hesketh scale and the corresponding algorithm provide practical means to predict the acute emetogenicity (≤24 h after chemotherapy) of individual chemotherapy agents or combination chemotherapy regimens in the absence of effective antiemetic prophylaxis [1]. According to the Hesketh scale, chemotherapy agents can be subdivided into five emetogenic levels: level 1 (<10% of patients experience acute emesis without antiemetic prophylaxis); level 2 (10–30%); level 3 (30–60%); level 4 (60–90%); and level 5 (>90%). For combination chemotherapy, the emetogenic level was determined by identifying the most emetogenic agent in the combination followed by an assessment of the relative contribution of the other agents to the overall level of emetogenicity. In doing so, the following rules apply: (a) level 1 agents do not contribute to the emetogenic level of a combination; (b) adding one or more level 2 agents increases the emetogenicity of the combination by one level greater than the most emetogenic agent in the combination; and (c) adding level 3 or 4 agents increases the emetogenicity of the combination by one level per agent. Patients were enrolled either at the start of cycles one or four of their treatment course. The study protocol provided for enrollment of consecutive patients. Exclusion criteria were as follows: scheduled chemotherapies that are known to cause delayed emesis beyond day 5 after the start of chemotherapy (e.g. 5-day cisplatin–etoposide–bleomycin for testicular cancer); life expectancy <12 weeks; Karnofsky score <60; presence of gastrointestinal disease
that had provoked emesis and/or nausea as defined by the National Cancer Institute–common toxicity criteria (NCI–CTC) [6] grade >0 for vomiting, and grade >2 for nausea; relevant previous CINV experience as defined by NCI–CTC [6] grade >0 for vomiting, and grade >2 for nausea; previously diagnosed CNS metastasis or primary CNS cancer; pregnancy. Patients receiving routine medication with emetic or antiemetic properties (e.g. opioids, butyrophenones) were not explicitly excluded. One to five per cent of patients received concomitant medications with antiemetogenic or emetogenic properties, depending on substance type. Because in most cases these medications were not prescribed routinely (with the exception of opioids containing pain medication), no further analysis was performed. Written informed consent was obtained from all patients; the study was approved by the ethics committee of the Medical Faculty of the Technische Universität München, Germany.
Data collection Data were obtained from preplanned, structured chart reviews and from selfadministered patient questionnaires. Chart data were abstracted by trained and experienced staff (e.g. a study nurse) of the Department of Pharmacy at the Klinikum rechts der Isar, Technische Universität München (MRI), or Medical Economics Research Group (MERG), Munich. Data extracted from medical charts included the following: patient demographics; clinical history and co-morbidity; clinical condition requiring chemotherapy; details of chemotherapy regimen and antiemetic prophylaxis and treatment; detailed vomiting and nausea experience during hospitalisation; and healthcare resource consumption due to vomiting and nausea. Patients were surveyed in the following areas: day-by-day experience of nausea and vomiting (frequency/level of severity; days 1–5); physician visits or hospitalisations required due to CINV; expenses incurred for over-thecounter (OTC) medication, transportation or home help required due to CINV; employment status; workdays lost due to CINV. For the purpose of this analysis, acute CINV was defined as at least one episode of nausea and/or vomiting during the first 24 h after the beginning of chemotherapy administration [2]. Delayed CINV was defined as at least one episode of nausea and/or vomiting that occurred >24 h after the beginning of chemotherapy administration, irrespective of the presence or absence of acute CINV in that patient. An emetic episode was defined as one episode of vomiting or a sequence of productive or unproductive retching episodes in very close succession. Severity of emesis was based on the NCI–CTC guidelines [6]. Patients were asked to rate their level of nausea on a three-point Likert scale (mild, moderate, severe).
Valuations and cost analysis Direct medical costs were defined as those of prophylactic or rescue antiemetic drugs, drug administration devices, personnel time associated with preparing and administering medication and tending to patients with CINV, extended hospitalisations or re-hospitalisations due to CINV, hospital outpatient or physician office visits due to CINV, as well as cost of OTC or complementary and/or alternative medications. Direct non-medical costs were those for transportation and need for assistance, such as additional childcare. Analysis of indirect costs was based exclusively on the number of workdays lost due to CINV. Costs not included were chemotherapy agents, preplanned office or hospital visits or hospitalisations for the purpose of chemotherapy administration, and other patient management costs not directly related to CINV. Diagnostic laboratory or radiography tests were also not included in the analysis as they do not generally play a relevant role in the management of CINV; in our patient sample, the cost for this line item was close to zero (data not shown). No attempt was made to quantify intangible costs such as the impact of CINV on a patient’s quality of life. All costs were calculated in 2002$ and not discounted.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
impact and costs of CINV in clinical practice, as opposed to clinical trial settings, is of importance for allocating resources to existing and novel antiemetic drug regimens. There are many reports on the incremental cost-effectiveness of antiemetic drugs, but information on the total cost per patient associated with CINV is limited [3]. Moreover, cost-of-illness data cannot necessarily be directly extrapolated from one country to another because of differences in treatment practice, drug regimens and drug dosages [4], as well as drug and labor costs [5]. Against this background, we set out to prospectively assess, under current practice patterns, the occurrence and impact on health resource utilization of CINV in patients receiving emetogenic chemotherapy. Additional aims of this study were to estimate the direct and indirect costs of CINV in the German healthcare environment, and to identify characteristics of patient subgroups that are associated with high costs imputable to CINV.
528 Unit costs were estimated from several different sources, depending on the payer perspective. Cost data sources were as follows:
Results
Provider (hospital) perspective: Bundesangestelltentarif 2003 (BAT; official tariffs for federal employees; personnel time); pharmacy acquisition cost for drugs administered in the hospital; DKG-NT, Bd.I: ‘Tarif der deutschen Krankenhausgesellschaft für die Abrechnung erbrachter Leistungen und für die Kostenerstattung vom Arzt an das Krankenhaus’, 2002 [infusion bags and tubing for intravenous (i.v.) drug administration in the hospital]; Statistisches Bundesamt: Fachserie 12, Reihe 6.3, 2000, Tab 7.2.1; Kosten je Pflegetag nach Krankenhaus Typen und Ländern (cost per day of hospitalisation).
Description of patient sample
The provider perspective was not calculated for the office-based treatment settings because data regarding staff time and administration costs are lacking. Patient perspective: Costs incurred by patients were accounted for as reported by patients in the questionnaire (OTC drugs and complementary and/ or alternative treatments, need for extra childcare, or other assistance, transportation, etc.) Indirect costs (referring to the societal perspective): The human capital approach was applied to estimate indirect costs resulting from productivity losses due to CINV (lost workdays; [7]). On the basis of data for gross wage and number of persons in dependent employment retrieved from Statistisches Bundesamt, Germany, the monetary value of productivity loss for employed persons amounted to $93.74 per day. Lost personal time from daily activities due to CINV and lost unpaid work due to CINV were not considered. The number of units of resources consumed were directly obtained from the medical charts or patient questionnaires. Personnel time associated with the management of CINV was based on a publication by Scherbel et al. [8]. These authors quantified, in the German hospital setting, the average personnel time required for preparation and administration of injectable antiemetics, and average nursing time and materials (emesis basins, wash towels, etc.) associated with each episode of vomiting (for assisting patients with a mouthwash, cleaning, comforting patients and sometimes changing bed linen and nightwear, etc.). The time required for each task was multiplied by the average hourly salary and benefits for each group of professionals.
Statistical analysis The statistical analysis software package SAS, version 8.0 (Cary, NC, USA) was used for the analysis of data. Descriptive statistics (mean, median, minimum, maximum and standard deviation) were calculated for each cost item. Differences in frequencies or scores were tested by the chi-square test and Mann–Whitney U-test, respectively. Average (per patient) costs per chemotherapy cycle were calculated separately for prophylaxis, treatment and overall management (i.e. prophylaxis plus treatment) of CINV, for those cycles that were associated with an experience of CINV and for all cycles (with and without CINV). Differentiation between these samples is important because patients not experiencing CINV still incurred the costs of prophylaxis. Differences in the cost of CINV were explored across several subgroups of patients. These were based on the following criteria: cisplatin/non-cisplatincontaining regimen; chemotherapy treatment cycle; level of CINV; presence of delayed CINV; and adherence to the American Society of Clinical Oncologists (ASCO) guidelines [2] for prophylaxis of delayed CINV.
Frequency and severity of CINV The frequency and severity of acute and delayed CINV is summarized in Table 3. During the 5-day observation period, 134 of 208 cycles (64.4%) were associated with at least one reported episode of nausea or vomiting. Delayed CINV was reported more often than acute CINV (60.7% versus 32.8%), and nausea was reported more often than vomiting (62.5% versus 26.0%). One patient withdrew from chemotherapy because of CINV.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Third party payer (‘Krankenkassen’, statutory sick fund) perspective: Einheitlicher Bewertungsmaßstab (EBM; German tariff list for outpatient care); Lauer-Taxe: WINAPO® Lauer-Taxe, Lauer-Fischer, Germany (acquisition cost for drugs prescribed by or administered in outpatient facilities; Zahlenbericht 2001/2002 der PKV, Private Krankenversicherer (reimbursable cost per day of hospitalization).
A total of 244 patients were enrolled in six centers, 195 patients (80%) returned questionnaires. The distribution of demographic and clinical characteristics of responders was similar to that of non-responders. Questionnaires from seven patients had to be excluded from the analysis due to incomplete data or protocol violations. The disposition of the study participants are shown in Figure 1. Of the remaining 188 evaluable patients, 88 were male, 100 were female; mean age was 55.3 ± 11.2 years; 41.6% of patients were in the paid work force or self-employed, the remaining 58.4% were retired or unemployed. Among the employed patients, 53.9% were on sick leave during and after chemotherapy. Of the 188 patients, 122 and 66 were treated in hospital- and office-based settings, respectively. Twenty-five patients had undergone previous courses of chemotherapy treatments previous to cycle one, but had no relevant previous experience of CINV. Of 188 patients, 20 were enrolled twice, at the start of cycles one and four; the remaining patients were studied during one cycle only, 146 and 42 patients were enrolled at the start of cycles one and four, respectively. Hence, the total number of cycles studied was 208. There were no differences in results (adherence to ASCO guidelines, experience of CINV) from patients studied during their first treatment cycle, or when results from all cycles, including the repeat cycles of the 20 patients who were enrolled twice, were considered. Therefore, all cycles were combined for analysis and presentation of data. Diagnoses and chemotherapy regimens are listed by treatment setting in Table 1. The most frequent diagnoses were gastrointestinal, breast and lung cancers. Patient populations treated in the hospital and office-based settings were not comparable with regard to type of malignancy and chemotherapy regimen. Antiemetic regimens administered for prophylaxis of acute and delayed CINV, and the association between compliance with ASCO guidelines [2] and CINV experience, are shown in Table 2. All patients received prophylactic antiemetic treatment on day 1, and 89.2% of antiemetic regimens were in compliance with ASCO guidelines for prophylaxis of acute CINV [2]. 86.5% of patients received prophylaxis for delayed CINV (>24 h after chemotherapy), but only 49.2% of the antiemetic regimens were in compliance with the ASCO guidelines for prophylaxis of delayed CINV [2].
529
Consumption of healthcare resources and workdays lost Utilization of health resources and workdays lost due to CINV are shown in Table 4. A total of 68 patients (33%) utilized healthcare resources due to CINV (Figure 1). The most frequently used resources were rescue medications and outpatient hospital and office physician visits. In contrast, only one patient required hospitalization, and only three patients lost workdays due to CINV.
Cost of CINV The average cost per treatment cycle resulting in an experience of CINV, incurred by third party payers and by patients, as well as the indirect costs, are detailed in Table 5. The corresponding cost figures from the perspective of the provider (hospital) are presented in Table 6. In Tables 5 and 6, costs are given as the mean cost per treatment cycle for two different groups, those patients who experienced CINV, and all patients. The acquisition and administration costs for antiemetic drugs are separated for prophylactic and treatment (rescue) use, to account for the fact that even patients not experiencing CINV still incurred the costs of prophylaxis [3].
We conducted several subgroup analyses from the hospital provider perspective to identify patient or treatment characteristics that were associated with high costs imputable to CINV (Table 7). These characteristics were as follows: cisplatin containing regimen; experience of emesis; and presence of delayed CINV (P <0.05 versus reference group). Non-adherence to ASCO guidelines for delayed CINV was associated with slightly higher cost, but the difference did not reach the level of significance. The cost of CINV for patients studied during cycle four of their chemotherapy was not statistically significantly different from cycle-one patients.
Discussion The goal of antiemetic treatment is to prevent nausea and vomiting completely [2]. However, despite improvements and considerable research efforts, this goal is still elusive for a significant number of patients. Results from our prospective study support the concept that CINV remains a relevant adverse effect of chemotherapy in daily clinical practice. A majority of patients (64.4%) experienced at least one episode of nausea or vomiting, even though all of them
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Figure 1. Disposition of study patients.
530 Table 1. Type of malignancy and chemotherapy (n = 188 patients)a Treatment setting, n (%) Hospital-based setting
Office-based setting
All patients
(n = 122)
(n = 66)
(n = 188)
Type of malignancy 64 (52.5)
9 (13.6)
73 (38.8)
Breast
17 (13.9)
36 (54.6)
53 (28.2)
Lung
21 (17.2)
6 (9.1)
27 (14.4)
3 (2.5)
6 (9.1)
9 (4.8)
10 (8.2)
6 (9.1)
16 (8.5)
7 (5.7)
3 (4.6)
10 (5.3)
Cisplatin
69 (56.7)
2 (3.0)
71 (37.8)
Carboplatin
11 (9.0)
15 (22.7)
26 (13.8)
Oxaliplatin
9 (7.4)
5 (7.6)
14 (7.4)
Epirubicine
Ovarial Lymphoma Other Chemotherapy
a
a
17 (13.9)
24 (36.3)
41 (21.8)
Doxirubicine
8 (6.6)
10 (15.2)
18 (9.6)
Cyclophosphamide
5 (4.1)
7 (10.6)
12 (6.4)
Other
3 (2.5)
3 (4.5)
6 (3.2)
For chemotherapy combinations, the agent with the greatest emetogenic risk is shown.
received prophylactic medication (Tables 2 and 3). This finding is in line with previous observational data and results from controlled clinical trials (for a review, see [9]). The unsatisfactory therapeutic outcome is driven by a lack of control of delayed CINV. Twice as many patients experienced delayed as opposed to acute CINV (Table 3). In part, this is likely to be the result of suboptimal therapeutic management. More than 50% of patients received a prophylactic regimen for delayed CINV that was not in agreement with ASCO guidelines, and a significantly higher proportion of these undertreated patients experienced delayed CINV than appropriately treated patients (71.6% versus 49.5%; Table 2). In contrast, 89.2% of patients received prophylaxis for acute CINV that was in compliance with ASCO guidelines. Our findings highlight that, unlike recommendations for the prevention of acute CINV, guidelines for the prevention of delayed CINV are still not widely implemented [10–13]. The prevention of delayed CINV is recognized to be of equal importance to the need to prevent acute CINV [2, 14] and, in particular, improved management of delayed CINV remains a priority. However, our results also show that even optimal management of CINV is not expected to fully achieve the therapeutic goal: nearly 50% of our patients who were managed according to guidelines experienced delayed CINV (Table 2). Our results also show that chemotherapy-induced nausea affects far more patients than vomiting (62.5% versus 26.0%; Table 3), which is in line with the notion that the introduction of 5-HT3 receptor antagonists has effected a reduction in the frequency of vomiting, but not in the frequency of nausea episodes [15]. This is noteworthy since patients, in contrast to physicians
and nurses, typically view control of nausea as more important than control of vomiting [16, 17]. In aggregate, these findings highlight the limited effectiveness of currently available antiemetic prophylaxis, particularly for delayed CINV and nausea [18], and a need for new pharmacological agents with improved efficacy. A further aim of this study was to estimate costs imputable to CINV so as to provide information for future decisions on resource allocation. We separately present data on consumption of healthcare resources (Table 4) and cost data (Tables 5 and 6), to facilitate the application of our findings to cost analyses in other healthcare settings where unit costs may differ. Most line items shown in Table 4 and their relative contributions to overall cost probably lend themselves to extrapolation to other healthcare environments. Corresponding data from a similar analysis conducted in Canada were broadly comparable to our findings [3], including the small number of patients requiring hospitalization due to CINV. However, the small number of workdays lost, and the low indirect cost resulting from this observation, may not be readily applicable to other healthcare settings. In our sample, only three of those patients who experienced CINV and were expected to work lost a total of 4.4 paid workdays (~35 working hours). O’Brien et al. [3] found that, among 72 Canadian patients experiencing CINV, there were total losses of 198 h of paid employment and 409 h of unpaid employment, which explained the fact that, in their study, indirect cost accounted for about two-thirds of the total cost of CINV. Similarly, Lindley et al. [19] reported that 23% of their patients surveyed in a US cancer care center, “were unable to go to work due to emesis”.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Gastrointestinal
531 Table 2. Prophylactic antiemetic treatments Antiemetic agents used for prophylaxis of acute emesis (n = 204 cycles)a 5-HT3/corticosteroid
Cycles, %
Percentage of cycles with acute CINV
Yes (89.2)
33.0 (N and/or V)
75.0 9.8
5-HT3/corticosteroid/D2
4.4
5-HT3
6.8
5-HT3/H1
0.5
5-HT3/D2
0.5
D2
0.5
Corticosteroid
2.0
Corticosteroid/D2
0.5
9.3 (V)
No (10.8)
31.8 (N and/or V) 9.1 (V)
0.0
Total No. of patients, %
100
(100)
Antiemetic agents used for prophylaxis of delayed emesis (n = 201 cycles)a
Cycles, %
Antiemetic regimen in agreement with ASCO guidelines for delayed emesis? (%)
Percentage of cycles with delayed CINV
Yes (49.2)
49.5* (N and/or V)
Corticosteroid/5-HT3
6.9
Corticosteroid/5-HT3/H1
0.5
Corticosteroid/5-HT3/D2
30.8
Corticosteroid
8.0
Corticosteroid/H1
0.5
Corticosteroid/D2 5-HT3
2.5 21.9
5-HT3/D2
2.5
5-HT3/H1
0.5
D2
10.9
D2/H1
No (50.8)
0.5 b
1.0
No prophylaxis
13.5
Corticosteroid
11.1* (V)
Total No. of patients, %
100
71.6 (N and/or V) 33.3 (V)
(100)
a
Four and seven cycles, respectively, with missing values regarding onset of acute or delayed symptoms. Patients in this group received a cisplatin-containing regimen. *P >0.005 versus patients receiving antiemetic regimen not in agreement with ASCO guidelines. ASCO, American Society of Clinical Oncology; CINV, chemotherapy-induced nausea and vomiting; D2, dopamine2-receptor antagonists; H1, histamine1-receptor antagonists; 5HT3, 5-hydroxytryptamine3-receptor antagonists; N, nausea; V, vomiting.
b
The discrepancy may be attributable to differences in data accrual (e.g. “unable to go to work” need not be the same as counting actual workdays lost) or valuation of non-marketed labor time lost [3]. However, it might also be ascribed to true differences in local practice regarding sick leave. Most of the patients in our sample were on sick leave for the period during and after chemotherapy, irrespective of the occurrence of nausea or vomiting. This, and the fact that a majority of our patients were retired or unemployed, necessarily limited the impact of CINV on lost workdays in our study. Hence, this finding may be limited in scope to German settings. Quantification of costs imputable to CINV is not straightforward in a setting where some patients receive chemotherapy in hospitals while others are seen in office practices. Unit prices are not always
consistent across settings because hospitals often receive special discounts from manufacturers. Even more complicating, though, is the reimbursement structure currently in force in Germany. This implies, for example, that hospitals receive a lump sum per day of hospitalization from the third party payer (sick funds). The sum is not related to the actual amount of, for example, antiemetic medication or staff time required. Hence, when hospital patients require additional CINV treatment, the hospital provider, but not the third party payer, incurs extra drug and personnel costs. In contrast, when patients receive rescue medication in an office setting, the third party payer will reimburse drugs, while the provider incurs no drug costs. This necessitates separate calculation of costs for different perspectives and patient groups, and, for quantification of average cost per patient incurred by the third
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
5-HT3/corticosteroid/H1
No prophylaxis
Antiemetic regimen in agreement with ASCO guidelines for acute emesis? (%)
532 Table 3. Frequency and severity of acute and delayed chemotherapy-induced nausea and vomiting (n = 208 cycles) Grade of nausea and vomiting
Acute (≤24 h post chemotherapy), n (%)a,c
Delayed (>24 h post chemotherapy), n (%)b,c
Total No. of cycles, n (%)
(n = 204)
(n = 201)
(n = 208)
Vomiting by NCI grade 1–2
22 (10.8)
3–4
42 (20.9)
45 (21.6)
6 (3.0)
6 (3.0)
7 (3.4)
28 (13.7)
48 (23.9)
52 (25.0)d
Mild
17 (8.3)
41 (20.3)
48 (23.1)
Moderate
22 (10.7)
44 (21.9)
44 (21.2)
Severe
27 (13.2)
33 (16.4)
37 (17.8)
All grades
66 (32.3)
118 (58.7)
129 (62.0)e
67 (32.8)
122 (60.7)
134 (64.4)
All grades Nausea by severity
a
Including patients with or without delayed symptoms. Including patients with or without acute symptoms. c Four and seven cycles, respectively, with missing values regarding onset of acute or delayed symptoms. d Two cycles with missing values regarding severity of vomiting; in total, 54 cycles (26.0%) with vomiting. e One cycle with missing values regarding severity of nausea; in total, 130 cycles (62.5%) with nausea. NCI, National Cancer Institute [6]. b
Table 4. Utilization of healthcare resources and workdays lost due to CINV (total number of evaluable cycles, n = 208) Type of resource consumption (units)
No. of cycles patients utilizing resources/ reporting lost workdays, n (%)
Total No. of units consumed/full workdays lost
Rescue medication i.v. (dosages)
13 (6.3)
29
Rescue medication oral & rectal (dosages)
20 (9.6)
270
Office physician visits (individual visit)
23 (11.1)
28
Hospital outpatient visits (individual visit)
6 (2.9)
6
(Extended) hospitalization (days)
1 (0.5)
3
Ambulance transport (rides)
1 (0.5)
1
OTC/alternative medication (purchases)
11 (5.3)
15
Taxi/public transportationa (rides)
15 (7.2)
15
Extra home help/childcare (episodes)
3 (1.4)
NE
Workdays lost (full-day equivalents)b
3 (10.0)
a
a
4.4
a
Patient out-of-pocket payment. Based on 30 cycles where patients are expected to work (patients are not retired and had no sick leave due to chemotherapy). CINV, chemotherapy-induced nausea and vomiting; i.v., intravenous; NE, not evaluated; OTC, over-the-counter medication.
b
party payer, requires data from different settings. We therefore attempted to recruit a case mix of patients from different types of treatment centers. The major cost driver from a sick funds’ perspective was the acquisition cost of antiemetic drugs for prophylaxis and treatment of CINV ($40.27 of a total cost of $48.66 per patient; Table 5); as described earlier, this was incurred exclusively in the office-based setting. Given the German reimbursement structure, it is not surprising that the main cost driver for hospital providers was staff time and, to a small extent, drug administration devices and other materials ($29.94 of a total cost of $48.27 per patient; Table 6).
More importantly, nearly one-third of the total cost per patient imputable to CINV was due to the treatment of CINV. This was true for both the third party payer and the hospital-provider perspectives, and represents the potential for cost offset that might be saved by improved management of CINV. Out-of-pocket payments by patients and indirect costs represent only a comparatively small fraction of the total cost of CINV in our sample. We have not estimated CINV cost from the perspective of officebased physicians due to a lack of data on staff time requirements and valuation of staff time. Hourly wage rates in office-based
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Any nausea and/or vomiting
533 Table 5. Cost of CINV incurred by third party payers (sick funds) and patients, and indirect costsa Type of cost
Mean cost per treatment cycle per patient in 2002,$ Cycles with CINV (n = 134)
All cycles (n = 208)
Direct cost incurred by third party payer 7.22 ± 30.07 [0.00]
4.65 ± 11.25 [0.00]
2.99 ± 9.29 [0.00]
CINV-attributable (extended) hospitalizations
5.34 ± 61.80 [0.00]
3.44 ± 49.60 [0.00]
Ambulance transport
3.04 ± 35.22 [0.00]
1.96 ± 28.27 [0.00]
Subtotal: costs of treatment of CINV
24.23 ± 119.59 [0.00]
15.61 ± 96.56 [0.00]
Medication to prevent CINV (prophylaxis)c
45.37 ± 65.39 [0.00]
33.05 ± 57.07 [0.00]
Total: costs of prophylaxis plus treatment
69.60 ± 135.18 [33.21]
48.66 ± 113.01 [0.00]
Direct out-of-pocket costs incurred by patients OTC/alternative medication
1.46 ± 7.06 [0.00]
0.94 ± 5.71 [0.00]
Taxi/public transportation
0.79 ± 2.36 [0.00]
0.51 ± 1.93 [0.00]
Extra home help/childcare
2.39 ± 13.27 [0.00]
1.54 ± 10.65 [0.00]
Total
4.64 ± 27.90 [0.00]
2.99 ± 22.47 [0.00]
3.05 ± 22.74 [0.00]
1.97 ± 18.28 [0.00]
Indirect costs Workdays lost—total cost per patient Direct cost—third party payer (cost of prophylaxis plus treatment)
69.60 ± 135.18 [33.21]
48.66 ± 113.01 [0.00]
Direct cost—patient
4.64 ± 27.90 [0.00]
2.99 ± 22.47 [0.00]
Indirect cost
3.05 ± 22.74 [0.00]
1.97 ± 18.28 [0.00]
Total
77.30 ± 146.59 [36.14]
53.62 ± 122.72 [0.00]
a
Costs are given as the mean per treatment cycle for two different denominators: cycles associated with CINV experience, and all cycles. Mean ± SD [median] costs are presented because the data were not normally distributed. b Includes only rescue medication administered outside hospital, because drugs administered in-hospital are included in hospital day rates. c Cost for prophylaxis in patients receiving chemotherapy at the hospital is calculated at $0, because cost for prophylaxis is included in the flat rate third party payers reimburse for chemotherapy treatment. CINV, chemotherapy-induced nausea and vomiting; OTC, over the counter medication.
settings are dependent on the income of individual office practices. Hence, they are not publicly available and may vary considerably across centers. Even though we may have captured an almost complete spectrum of costs directly attributable to CINV, we are also not presenting cost data from a societal perspective. We recognise that this is a shortcoming of our analysis [4], since resource allocation decisions should ideally be made from a societal perspective [20]. However, the complexity of the cost structure in Germany, which may result in double counting of individual line items and a lack of data, virtually preclude calculation of this perspective. For example, data on the true cost per day in hospital or per outpatient visit have not been published or otherwise available to us. Given these limitations, the best ‘proxy’ estimate of societal cost of CINV is likely obtained by adding the costs incurred by third party payers and patients, and indirect costs. This yields a total cost estimate of $53.62 per patient per treatment cycle (Table 5). To provide further insight into the relative influence of different patient or treatment characteristics on the overall cost of CINV, we constructed several subgroups of patients. This analysis was
limited to patients treated in hospital settings from the provider’s perspective. The rationale for excluding patients treated in physician offices was the difference in cost structure between the settings, and the fact that allocation of patients to the different settings was non-random, and the lack of comparability of the groups with regard to type of malignancy and chemotherapy (Table 1) would have confounded subgroup analyses from the combined populations. Not surprisingly, cisplatin-containing chemotherapies were associated with higher average costs of prophylaxis and treatment of CINV than non-cisplatin regimens (Table 7). Subgroup analyses further show that overall cost of CINV was associated with presence of vomiting, presence of delayed CINV, and inadequate prophylaxis for delayed CINV (as indicated by a lack of adherence with ASCO guidelines for delayed CINV, although this did not reach the level of significance; Table 7). These findings are directly relevant to patient-management decisions, as they identify subgroups of patients for whom improved CINV prophylaxis is likely to result in the highest levels of cost savings or cost offset.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
11.20 ± 36.90 [0.00]
CINV-attributable outpatient physician visits
Medication to treat CINVb
534 Table 6. Cost of CINV incurred by provider (hospital)a Mean cost per treatment cycle per patient in 2002, $ Cycles with CINV (n = 73) Medication to treat CINV (rescue medication)b
1.32 ± 6.17 [0.00]
All cycles (n = 137) 0.71 ± 4.54 [0.00] 0.37 ± 1.44 [0.00]
25.30 ± 109.13 [0.00]
13.48 ± 80.41 [0.00]
Subtotal
27.31 ± 109.20 [0.00]
14.55 ± 80.62 [0.00]
Medication to prevent CINV (prophylaxis)
17.38 ± 9.73 [15.98]
17.25 ± 9.52 [15.98]
Prophylactic drug administration devices and staff time
16.51 ± 5.07 [20.60]
16.46 ± 5.07 [20.60]
Subtotal
33.89 ± 11.42 [36.13]
33.72 ± 11.09 [36.58]
Total
61.20 ± 109.68 [38.08]
48.27 ± 81.33 [36.58]
a
The provider perspective is not calculated for the office-based treatment settings, because data regarding staff time and administration costs are lacking. Costs are given as means per treatment cycle for two different denominators (cycles associated with CINV experience, and all cycles), because patients not experiencing CINV during a particular cycle still incurred the costs of prophylaxis. Mean ± SD [median] costs are presented because the data were not normally distributed. b Includes only rescue medication administered inside hospital.
How do our findings compare with results from similar studies? We found two reports concerned with CINV-attributable costs in Germany. Ihbe-Heffinger et al. [21] quantified the costs of drug acquisition and administration associated with the control of CINV from a hospital perspective. In their cohort of breast cancer patients receiving moderately emetogenic chemotherapy, total costs (comprising prophylaxis and treatment of CINV) ranged from $35 to $73, depending on the type of antiemetic drugs used. This is in good agreement with our data from the hospital treatment setting. Herold and Hieke [5] compared the cost of toxicity of standard chemotherapy given to patients with non-Hodgkin’s lymphoma in several countries. In their retrospective analysis, the average cost to third party payers for managing CINV in Germany was $358 per patient for a complete course of treatment, which typically comprised six cycles of chemotherapy. Hence, average cost per patient per cycle amounts to approximately $60, which is also in line with our findings of $48.66 in a more heterogeneous patient population. It is difficult to compare our findings on costs of CINV with other reports because published information is limited. To the best of our knowledge, there is only one study that is similar in design to ours [3], but it predates the availability of 5-HT3 receptor antagonists, and relates to the Canadian healthcare environment. A number of methodological aspects of this study deserve discussion. Our study was prospective and was based on specifically designed questionnaires; all data acquisition was performed or monitored by experienced and trained personnel. These study features allowed us to capture cost items with a degree of accuracy and completeness that may not usually be available from retrospective chart review or medical claims data. This study was not restricted to a particular cancer type and we have deliberately enrolled a heterogeneous group of cancer
patients with a variety of malignancies. Patients received a broad range of chemotherapies, all of which are considered highly emetogenic [1]. This is expected to make our results relevant for extrapolation to most cancer patient populations on highly emetogenic treatments. Our study may have systematically underestimated the true cost of CINV in at least two ways. Firstly, and as discussed above, the results indicate that management of CINV may have been suboptimal, and improved management might result in cost savings that are not apparent from our study, since we did not consider the cost of chemotherapy administration. Stewart et al. [22] observed retrospectively that the improved control of CINV by the introduction of 5-HT3 antagonists enabled the administration of larger doses of chemotherapy in shorter periods of time. As a result, some multi-day regimens were successfully converted to singleday regimens, and prolonged infusions given on an inpatient basis were converted to relatively short outpatient infusions. Hence, our study may have underestimated the resources wasted (i.e. not saved) due to the continuing high rate of CINV. Conceivably, more savings or cost offsets could be achieved by reducing the frequency and severity of CINV, either by optimizing the management with existing medication or by use of novel antiemetic agents with improved efficacy. Secondly, we have not factored in less tangible and hidden costs, such as unpaid family labor and time in caring for patients, or patients’ inability to perform household work, which are often ignored but may not be insignificant [23, 24]. Nevertheless, the present results document that a substantial proportion of patients continue to experience CINV. This entails not only clinical but also economic consequences, and highlights a continuing need for improved utilization of existing antiemetic agents and for new, more efficacious treatments. Our data indicate
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
0.69 ± 1.93 [0.00]
Staff time and material consumption associated with managing episode of CINV
Rescue-drug administration devices
535 Table 7. CINV-attributable cost in 2002$ by subgroup (hospital-provider perspective)a n
Prophylaxis
137
33.72 ± 11.09
14.55 ± 80.62
48.27 ± 81.33
Non-cisplatin
55
26.67 ± 12.70
7.11 ± 18.57
33.79 ± 22.11
1.00 (ref.)
Cisplatin
82
38.44 ± 6.54
19.54 ± 103.05
57.98 ± 102.68
1.72*
No CINV
64
33.52 ± 10.80
0.00
33.52 ± 10.80
1.00 (ref.)
Nausea only
44
34.28 ± 10.30
5.45 ± 10.17
39.73 ± 19.93
1.19
Emesis (any level)
29
33.30 ± 13.11
60.48 ± 169.15
93.77 ± 169.28
2.80*
Emesis level 1–2
25
33.04 ± 11.88
25.21 ± 23.08
58.26 ± 25.93
1.74
Emesis level 3–4
3
37.67 ± 25.67
374.50 ± 478.43
412.17 ± 472.22
12.29
Delayed, no
67
33.36 ± 10.69
33.36 ± 10.69
1.00 (ref.)
Delayed, yes
66
33.81 ± 11.69
29.25 ± 114.61
63.06 ± 115.03
1.89*
Delayed, yes
100
37.04 ± 8.21
6.45 ± 16.31
43.49 ± 21.08
1.00 (ref.)
Delayed, no
37
24.73 ± 12.87
36.46 ± 152.14
61.19 ± 153.40
1.41
All cycles
Treatment
Total (prophylaxis + treatment)
Normalized total cost NA
Cisplatin-containing chemotherapies
Level of acute or delayed CINVb
0.00
Adherence to ASCO guidelines for prophylaxis of delayed CINV
Mean ± SD costs are indicated as per patient per cycle. To facilitate interpretation of cost differences between different patient subgroups, costs are also expressed as multiples of a designated reference subgroup. Note that these analyses include only results from patients treated in a hospital setting; patients treated in an office-based care setting were excluded. b One cycle with missing values regarding severity of emesis. c Four cycles with missing values regarding time of onset of symptoms. *P <0.05 versus reference group. CINV, chemotherapy-induced nausea and vomiting; NA, not applicable; ref., reference. a
that the greatest improvement in patient care and potential for cost offset may be realized by avoidance of delayed CINV.
Acknowledgements This study was supported by Merck & Co., Inc., Whitehouse Station, NJ, USA.
References 1. Hesketh PJ, Kris MG, Grunberg SM et al. Proposal for classifying the acute emetogenicity of cancer chemotherapy. J Clin Oncol 1997; 15: 103–109. 2. Gralla RJ, Osoba D, Kris MG et al. Recommendations for the use of antiemetics: evidence-based, clinical practice guidelines. J Clin Oncol 1999; 17: 2971–2994. 3. O’Brien BJ, Rusthoven J, Rocci A et al. Impact of chemotherapy-associated nausea and vomiting on patients’ functional status and on costs: survey of five Canadian centres. Can Med Assoc J 1993; 149: 296–302. 4. Uyl-de Groot CA, Wait S, Buijt I. Economics and health-related quality of life in antiemetic therapy: recommendations for trial design. Eur J Cancer 2000; 36: 1522–1535. 5. Herold M, Hieke K. Costs of toxicity during chemotherapy with CHOP, COP/CVP, and fludarabine. Eur J Health Econom 2002; 3: 166–172.
6. Cancer Therapy Evaluation Program. Common toxicity criteria, version 2.0. DCTD, NCI, NIH and DHHS, March 1998; [online] www.ctep.cancer.gov/ forms/ctcv2nom-4–30–99-final3.pdf (date last accessed: July 2003). 7. Hannoveraner Konsens Gruppe. Dt. Empfehlungen zur gesundheitsökonomischen Evaluation—Revidierte Fassung des Hannoveraner Konsens. Gesundheitsökonomie Qualitätsmanagement 1999; 4: A62–A65. 8. Scherbel G, Bodenmüller-Kroll R, Matthias M et al. Antiemetische Therapieformen Kosten und Effektivität im Vergleich. Arzneimitteltherapie 1994; 6 (Suppl 17): 1–8. 9. Beckwith MC, Mullin S. Prevention and management of chemotherapyinduced nausea and vomiting, part 1. Hospital Pharmacy 2001; 36: 67–82. 10. Mertens WC, Higby DJ, Brown D et al. Improving the care of patients with regard to chemotherapy-induced nausea and emesis: the effect of feedback to clinicians on adherence to antiemetic prescribing guidelines. J Clin Oncol 2003; 21: 1373–1378. 11. Roila F, Donati D, Tamberi S et al. Delayed emesis: patterns, prognostic factors and optimal treatment. Support Care Cancer 2002; 10: 88–95. 12. Warr D. Standard treatment of chemotherapy-induced emesis. Support Care Cancer 1997; 5: 12–16. 13. Rubenstein EB, Litke SE, Jose MA. Chemotherapy-induced nausea and emesis in the 5HT3 era: a national survey of hematologists/oncologists. Proc Am Soc Clin Oncol 2002; 21: 260 (Abstr 2860). 14. Roila F. Dexamethasone alone or in combination with ondansetron for the prevention of delayed nausea and vomiting induced by chemotherapy. Italian Group for Antiemetic Research. N Engl J Med 2000; 342: 1554– 1559.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Presence of delayed CINVc
536 20. Russell LB, Gold MR, Siegel JE et al. The role of cost-effectiveness analysis in health and medicine. JAMA 1996; 276: 1172–1177. 21. Ihbe-Heffinger A, Kuhn W, Sattler D et al. Cost-minimization analysis for the use of 5-HT3 receptor antagonists in moderately emetogenic chemotherapy. A prospective pharmacoeconomic evaluation in German hospitals. Eur J Cancer 2000; 36 (Suppl 3): 18. 22. Stewart DJ, Dahrouge S, Coyle D, Evans WK. Cost of treating and preventing nausea and vomiting in patients receiving chemotherapy. J Clin Oncol 1999; 17: 344–351. 23. Hayman JA, Langa KM, Kabeto MU et al. Estimating the cost of informal caregiving for elderly patients with cancer. J Clin Oncol 2001; 19: 3219– 3225. 24. Lindley CM, Hirsch JD. Nausea and vomiting and cancer patients’ quality of life: a discussion of professor Selby’s paper. Br J Cancer 1992; 66 (Suppl XIX): S26–S29.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
15. Roscoe JA, Morrow GR, Hickok JT, Stern RM. Nausea and vomiting remain a significant clinical problem: trends over time in controlling chemotherapy-induced nausea and vomiting in 1413 patients treated in community clinical practices. J Pain Symptom Manage 2000; 20: 113– 121. 16. Bonneterre J, Hecquet B, Atenis A et al. How do patients and physicians decide which antiemetic is best in a cross over study? Proc Am Soc Clin Oncol 1991: 10: 323 (Abstr 1136). 17. Lecomte S, Lemenager M, Duquenne I et al. Quality of life in breast cancer patients receiving adjuvant chemotherapy. Proc Am Soc Clin Oncol 1995: 14: 308 (Abstr 900). 18. Morrow GR, Hickok JT, Rosenthal SN. Progress in reducing nausea and emesis. Comparisons of ondansetron (Zofran), granisetron (Kytril), and tropisetron (Navoban). Cancer 1995; 76: 343–357. 19. Lindley CM, Hirsch JD, O’Neill CV et al. Quality of life consequences of chemotherapy-induced emesis. Qual Life Res 1992; 1: 331–340.
Annals of Oncology 15: 526–536, 2004 DOI: 10.1093/annonc/mdh110
The impact of delayed chemotherapy-induced nausea and vomiting on patients, health resource utilization and costs in German cancer centers A. Ihbe-Heffinger1*, B. Ehlken2, R. Bernard1, K. Berger2, C. Peschel3, H.-G. Eichler4,5, R. Deuson4, J. Thödtmann1 & F. Lordick3 1
Department of Pharmacy, Klinikum rechts der Isar, Technische Universität München; 2MERG, Medical Economics Research Group, Munich; Third Medical Department, Klinikum rechts der Isar, Technische Universität München, Germany; 4Merck & Co., Inc., Whitehouse Station, NJ, USA; 5 Department of Clinical Pharmacology, University of Vienna Medical School, Vienna, Austria 3
Background: Delayed chemotherapy-induced nausea and vomiting (CINV) remains a major adverse effect of cancer chemotherapy. We assessed, under current practice patterns, the occurrence and impact on healthcare resource utilization of CINV in patients receiving emetogenic chemotherapy. An additional aim of this study was to estimate costs imputable to CINV in the German healthcare environment. Materials and methods: This prospective, multi-center, cross-sectional cost-of-illness study was conducted in three hospitals and in three office-based facilities in Germany. Consecutive patients undergoing emetogenic chemotherapy (levels 4 or 5 according to Hesketh classification of emetogenicity) were enrolled. Data were obtained from preplanned chart reviews and from self-administered patient questionnaires. Analysis of direct costs was performed from the perspectives of third party payer (statutory sick fund), provider (hospital) and patients. Indirect costs were assessed on the basis of paid workdays lost. Results: During the 5-day observation period, 134 of 208 chemotherapy cycles observed (64.4%) were associated with at least one episode of nausea or vomiting. More patients experienced delayed than acute CINV (60.7% versus 32.8%), and more patients reported nausea than vomiting (62.5% versus 26.0%). A total of 68 patients (32.6%) utilized healthcare resources due to CINV. The most frequently used resources were rescue medications and outpatient hospital and office physician visits. Only one patient required hospitalization and only three patients lost workdays due to CINV. Average costs imputable to CINV per patient (with or without CINV) per treatment cycle incurred by third party payers and hospital providers were $49 and $48, respectively. Patient or treatment characteristics that were associated with high costs imputable to CINV were as follows: cisplatincontaining regimen; experience of emesis; and presence of delayed CINV. Conclusions: A substantial proportion of patients continue to experience CINV. This entails not only clinical but also economic consequences, and highlights a continuing need for improved utilization of existing antiemetic agents and for new, more efficacious treatments. The greatest improvements in patient care and potential for cost offset may be realized by preventing delayed CINV. Key words: chemotherapy-induced nausea and vomiting, cost-of-illness, healthcare utilization, health economics, healthcare provider, third party payer
Introduction Chemotherapy-induced nausea and vomiting (CINV) remains a major adverse effect of cancer chemotherapy, despite the availability of several antiemetic drug classes, including 5-HT3 receptor antagonists. Although not life-threatening, CINV has a major impact on a patient’s quality of life and ranks high on the list of factors most feared by patients receiving chemotherapy [1].
*Correspondence to: Dr A. Ihbe-Heffinger, Apotheke des Klinikums rechts der Isar, der Technischen Universität München, Ismaninger Strasse 22, 81675 München, Germany. Tel: +49-89-4140-2219; Fax: +49-89-4140-6365; E-mail: [email protected] © 2004 European Society for Medical Oncology
Aside from these clinical consequences, CINV has considerable economic impact. Direct costs include acquisition cost of antiemetic drugs and administration devices, additional patient care, such as nursing and physician time, and in some cases, extended hospitalization or readmission. CINV is often not confined to the acute period (developing within 24 h of receiving chemotherapy), but may manifest itself as delayed CINV (>24 h post-chemotherapy), which can persist for several days [2]. Therefore, indirect costs may result from reduced productivity at work or workdays lost. Economic analyses, including cost-of-illness studies, receive increasing attention from healthcare payers and providers to help inform choices in the delivery of care. Hence, quantifying the
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Received 27 August 2003; revised 28 November 2003; accepted 19 December 2003
527
Patients and methods Study design, setting and perspectives This prospective, multi-center, cross-sectional cost-of-illness study was conducted in three hospitals and in three office-based facilities in Germany. All centers were experienced in the administration of cancer chemotherapy. Centers were selected to allow for enrollment of a broad spectrum of patients requiring chemotherapy and to represent both hospital- and office-based cancer care settings which reflect current medical practice in Germany. Analysis of costs was performed from three perspectives: (a) the third party payer (‘Krankenkassen’ or statutory sick funds), (b) the provider (hospital) and (c) the patient.
Patient selection criteria Patients were eligible for inclusion if they were scheduled for treatment with emetogenic chemotherapy as defined by level 4 or 5 categorization on the emetogenicity scale proposed by Hesketh et al. [1], or with a level 4 or 5 combination chemotherapy according to the algorithm provided by Hesketh et al. [1]. All of these agents or combinations are associated with a ≥60% frequency of emesis [1]. The Hesketh scale and the corresponding algorithm provide practical means to predict the acute emetogenicity (≤24 h after chemotherapy) of individual chemotherapy agents or combination chemotherapy regimens in the absence of effective antiemetic prophylaxis [1]. According to the Hesketh scale, chemotherapy agents can be subdivided into five emetogenic levels: level 1 (<10% of patients experience acute emesis without antiemetic prophylaxis); level 2 (10–30%); level 3 (30–60%); level 4 (60–90%); and level 5 (>90%). For combination chemotherapy, the emetogenic level was determined by identifying the most emetogenic agent in the combination followed by an assessment of the relative contribution of the other agents to the overall level of emetogenicity. In doing so, the following rules apply: (a) level 1 agents do not contribute to the emetogenic level of a combination; (b) adding one or more level 2 agents increases the emetogenicity of the combination by one level greater than the most emetogenic agent in the combination; and (c) adding level 3 or 4 agents increases the emetogenicity of the combination by one level per agent. Patients were enrolled either at the start of cycles one or four of their treatment course. The study protocol provided for enrollment of consecutive patients. Exclusion criteria were as follows: scheduled chemotherapies that are known to cause delayed emesis beyond day 5 after the start of chemotherapy (e.g. 5-day cisplatin–etoposide–bleomycin for testicular cancer); life expectancy <12 weeks; Karnofsky score <60; presence of gastrointestinal disease
that had provoked emesis and/or nausea as defined by the National Cancer Institute–common toxicity criteria (NCI–CTC) [6] grade >0 for vomiting, and grade >2 for nausea; relevant previous CINV experience as defined by NCI–CTC [6] grade >0 for vomiting, and grade >2 for nausea; previously diagnosed CNS metastasis or primary CNS cancer; pregnancy. Patients receiving routine medication with emetic or antiemetic properties (e.g. opioids, butyrophenones) were not explicitly excluded. One to five per cent of patients received concomitant medications with antiemetogenic or emetogenic properties, depending on substance type. Because in most cases these medications were not prescribed routinely (with the exception of opioids containing pain medication), no further analysis was performed. Written informed consent was obtained from all patients; the study was approved by the ethics committee of the Medical Faculty of the Technische Universität München, Germany.
Data collection Data were obtained from preplanned, structured chart reviews and from selfadministered patient questionnaires. Chart data were abstracted by trained and experienced staff (e.g. a study nurse) of the Department of Pharmacy at the Klinikum rechts der Isar, Technische Universität München (MRI), or Medical Economics Research Group (MERG), Munich. Data extracted from medical charts included the following: patient demographics; clinical history and co-morbidity; clinical condition requiring chemotherapy; details of chemotherapy regimen and antiemetic prophylaxis and treatment; detailed vomiting and nausea experience during hospitalisation; and healthcare resource consumption due to vomiting and nausea. Patients were surveyed in the following areas: day-by-day experience of nausea and vomiting (frequency/level of severity; days 1–5); physician visits or hospitalisations required due to CINV; expenses incurred for over-thecounter (OTC) medication, transportation or home help required due to CINV; employment status; workdays lost due to CINV. For the purpose of this analysis, acute CINV was defined as at least one episode of nausea and/or vomiting during the first 24 h after the beginning of chemotherapy administration [2]. Delayed CINV was defined as at least one episode of nausea and/or vomiting that occurred >24 h after the beginning of chemotherapy administration, irrespective of the presence or absence of acute CINV in that patient. An emetic episode was defined as one episode of vomiting or a sequence of productive or unproductive retching episodes in very close succession. Severity of emesis was based on the NCI–CTC guidelines [6]. Patients were asked to rate their level of nausea on a three-point Likert scale (mild, moderate, severe).
Valuations and cost analysis Direct medical costs were defined as those of prophylactic or rescue antiemetic drugs, drug administration devices, personnel time associated with preparing and administering medication and tending to patients with CINV, extended hospitalisations or re-hospitalisations due to CINV, hospital outpatient or physician office visits due to CINV, as well as cost of OTC or complementary and/or alternative medications. Direct non-medical costs were those for transportation and need for assistance, such as additional childcare. Analysis of indirect costs was based exclusively on the number of workdays lost due to CINV. Costs not included were chemotherapy agents, preplanned office or hospital visits or hospitalisations for the purpose of chemotherapy administration, and other patient management costs not directly related to CINV. Diagnostic laboratory or radiography tests were also not included in the analysis as they do not generally play a relevant role in the management of CINV; in our patient sample, the cost for this line item was close to zero (data not shown). No attempt was made to quantify intangible costs such as the impact of CINV on a patient’s quality of life. All costs were calculated in 2002$ and not discounted.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
impact and costs of CINV in clinical practice, as opposed to clinical trial settings, is of importance for allocating resources to existing and novel antiemetic drug regimens. There are many reports on the incremental cost-effectiveness of antiemetic drugs, but information on the total cost per patient associated with CINV is limited [3]. Moreover, cost-of-illness data cannot necessarily be directly extrapolated from one country to another because of differences in treatment practice, drug regimens and drug dosages [4], as well as drug and labor costs [5]. Against this background, we set out to prospectively assess, under current practice patterns, the occurrence and impact on health resource utilization of CINV in patients receiving emetogenic chemotherapy. Additional aims of this study were to estimate the direct and indirect costs of CINV in the German healthcare environment, and to identify characteristics of patient subgroups that are associated with high costs imputable to CINV.
528 Unit costs were estimated from several different sources, depending on the payer perspective. Cost data sources were as follows:
Results
Provider (hospital) perspective: Bundesangestelltentarif 2003 (BAT; official tariffs for federal employees; personnel time); pharmacy acquisition cost for drugs administered in the hospital; DKG-NT, Bd.I: ‘Tarif der deutschen Krankenhausgesellschaft für die Abrechnung erbrachter Leistungen und für die Kostenerstattung vom Arzt an das Krankenhaus’, 2002 [infusion bags and tubing for intravenous (i.v.) drug administration in the hospital]; Statistisches Bundesamt: Fachserie 12, Reihe 6.3, 2000, Tab 7.2.1; Kosten je Pflegetag nach Krankenhaus Typen und Ländern (cost per day of hospitalisation).
Description of patient sample
The provider perspective was not calculated for the office-based treatment settings because data regarding staff time and administration costs are lacking. Patient perspective: Costs incurred by patients were accounted for as reported by patients in the questionnaire (OTC drugs and complementary and/ or alternative treatments, need for extra childcare, or other assistance, transportation, etc.) Indirect costs (referring to the societal perspective): The human capital approach was applied to estimate indirect costs resulting from productivity losses due to CINV (lost workdays; [7]). On the basis of data for gross wage and number of persons in dependent employment retrieved from Statistisches Bundesamt, Germany, the monetary value of productivity loss for employed persons amounted to $93.74 per day. Lost personal time from daily activities due to CINV and lost unpaid work due to CINV were not considered. The number of units of resources consumed were directly obtained from the medical charts or patient questionnaires. Personnel time associated with the management of CINV was based on a publication by Scherbel et al. [8]. These authors quantified, in the German hospital setting, the average personnel time required for preparation and administration of injectable antiemetics, and average nursing time and materials (emesis basins, wash towels, etc.) associated with each episode of vomiting (for assisting patients with a mouthwash, cleaning, comforting patients and sometimes changing bed linen and nightwear, etc.). The time required for each task was multiplied by the average hourly salary and benefits for each group of professionals.
Statistical analysis The statistical analysis software package SAS, version 8.0 (Cary, NC, USA) was used for the analysis of data. Descriptive statistics (mean, median, minimum, maximum and standard deviation) were calculated for each cost item. Differences in frequencies or scores were tested by the chi-square test and Mann–Whitney U-test, respectively. Average (per patient) costs per chemotherapy cycle were calculated separately for prophylaxis, treatment and overall management (i.e. prophylaxis plus treatment) of CINV, for those cycles that were associated with an experience of CINV and for all cycles (with and without CINV). Differentiation between these samples is important because patients not experiencing CINV still incurred the costs of prophylaxis. Differences in the cost of CINV were explored across several subgroups of patients. These were based on the following criteria: cisplatin/non-cisplatincontaining regimen; chemotherapy treatment cycle; level of CINV; presence of delayed CINV; and adherence to the American Society of Clinical Oncologists (ASCO) guidelines [2] for prophylaxis of delayed CINV.
Frequency and severity of CINV The frequency and severity of acute and delayed CINV is summarized in Table 3. During the 5-day observation period, 134 of 208 cycles (64.4%) were associated with at least one reported episode of nausea or vomiting. Delayed CINV was reported more often than acute CINV (60.7% versus 32.8%), and nausea was reported more often than vomiting (62.5% versus 26.0%). One patient withdrew from chemotherapy because of CINV.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Third party payer (‘Krankenkassen’, statutory sick fund) perspective: Einheitlicher Bewertungsmaßstab (EBM; German tariff list for outpatient care); Lauer-Taxe: WINAPO® Lauer-Taxe, Lauer-Fischer, Germany (acquisition cost for drugs prescribed by or administered in outpatient facilities; Zahlenbericht 2001/2002 der PKV, Private Krankenversicherer (reimbursable cost per day of hospitalization).
A total of 244 patients were enrolled in six centers, 195 patients (80%) returned questionnaires. The distribution of demographic and clinical characteristics of responders was similar to that of non-responders. Questionnaires from seven patients had to be excluded from the analysis due to incomplete data or protocol violations. The disposition of the study participants are shown in Figure 1. Of the remaining 188 evaluable patients, 88 were male, 100 were female; mean age was 55.3 ± 11.2 years; 41.6% of patients were in the paid work force or self-employed, the remaining 58.4% were retired or unemployed. Among the employed patients, 53.9% were on sick leave during and after chemotherapy. Of the 188 patients, 122 and 66 were treated in hospital- and office-based settings, respectively. Twenty-five patients had undergone previous courses of chemotherapy treatments previous to cycle one, but had no relevant previous experience of CINV. Of 188 patients, 20 were enrolled twice, at the start of cycles one and four; the remaining patients were studied during one cycle only, 146 and 42 patients were enrolled at the start of cycles one and four, respectively. Hence, the total number of cycles studied was 208. There were no differences in results (adherence to ASCO guidelines, experience of CINV) from patients studied during their first treatment cycle, or when results from all cycles, including the repeat cycles of the 20 patients who were enrolled twice, were considered. Therefore, all cycles were combined for analysis and presentation of data. Diagnoses and chemotherapy regimens are listed by treatment setting in Table 1. The most frequent diagnoses were gastrointestinal, breast and lung cancers. Patient populations treated in the hospital and office-based settings were not comparable with regard to type of malignancy and chemotherapy regimen. Antiemetic regimens administered for prophylaxis of acute and delayed CINV, and the association between compliance with ASCO guidelines [2] and CINV experience, are shown in Table 2. All patients received prophylactic antiemetic treatment on day 1, and 89.2% of antiemetic regimens were in compliance with ASCO guidelines for prophylaxis of acute CINV [2]. 86.5% of patients received prophylaxis for delayed CINV (>24 h after chemotherapy), but only 49.2% of the antiemetic regimens were in compliance with the ASCO guidelines for prophylaxis of delayed CINV [2].
529
Consumption of healthcare resources and workdays lost Utilization of health resources and workdays lost due to CINV are shown in Table 4. A total of 68 patients (33%) utilized healthcare resources due to CINV (Figure 1). The most frequently used resources were rescue medications and outpatient hospital and office physician visits. In contrast, only one patient required hospitalization, and only three patients lost workdays due to CINV.
Cost of CINV The average cost per treatment cycle resulting in an experience of CINV, incurred by third party payers and by patients, as well as the indirect costs, are detailed in Table 5. The corresponding cost figures from the perspective of the provider (hospital) are presented in Table 6. In Tables 5 and 6, costs are given as the mean cost per treatment cycle for two different groups, those patients who experienced CINV, and all patients. The acquisition and administration costs for antiemetic drugs are separated for prophylactic and treatment (rescue) use, to account for the fact that even patients not experiencing CINV still incurred the costs of prophylaxis [3].
We conducted several subgroup analyses from the hospital provider perspective to identify patient or treatment characteristics that were associated with high costs imputable to CINV (Table 7). These characteristics were as follows: cisplatin containing regimen; experience of emesis; and presence of delayed CINV (P <0.05 versus reference group). Non-adherence to ASCO guidelines for delayed CINV was associated with slightly higher cost, but the difference did not reach the level of significance. The cost of CINV for patients studied during cycle four of their chemotherapy was not statistically significantly different from cycle-one patients.
Discussion The goal of antiemetic treatment is to prevent nausea and vomiting completely [2]. However, despite improvements and considerable research efforts, this goal is still elusive for a significant number of patients. Results from our prospective study support the concept that CINV remains a relevant adverse effect of chemotherapy in daily clinical practice. A majority of patients (64.4%) experienced at least one episode of nausea or vomiting, even though all of them
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Figure 1. Disposition of study patients.
530 Table 1. Type of malignancy and chemotherapy (n = 188 patients)a Treatment setting, n (%) Hospital-based setting
Office-based setting
All patients
(n = 122)
(n = 66)
(n = 188)
Type of malignancy 64 (52.5)
9 (13.6)
73 (38.8)
Breast
17 (13.9)
36 (54.6)
53 (28.2)
Lung
21 (17.2)
6 (9.1)
27 (14.4)
3 (2.5)
6 (9.1)
9 (4.8)
10 (8.2)
6 (9.1)
16 (8.5)
7 (5.7)
3 (4.6)
10 (5.3)
Cisplatin
69 (56.7)
2 (3.0)
71 (37.8)
Carboplatin
11 (9.0)
15 (22.7)
26 (13.8)
Oxaliplatin
9 (7.4)
5 (7.6)
14 (7.4)
Epirubicine
Ovarial Lymphoma Other Chemotherapy
a
a
17 (13.9)
24 (36.3)
41 (21.8)
Doxirubicine
8 (6.6)
10 (15.2)
18 (9.6)
Cyclophosphamide
5 (4.1)
7 (10.6)
12 (6.4)
Other
3 (2.5)
3 (4.5)
6 (3.2)
For chemotherapy combinations, the agent with the greatest emetogenic risk is shown.
received prophylactic medication (Tables 2 and 3). This finding is in line with previous observational data and results from controlled clinical trials (for a review, see [9]). The unsatisfactory therapeutic outcome is driven by a lack of control of delayed CINV. Twice as many patients experienced delayed as opposed to acute CINV (Table 3). In part, this is likely to be the result of suboptimal therapeutic management. More than 50% of patients received a prophylactic regimen for delayed CINV that was not in agreement with ASCO guidelines, and a significantly higher proportion of these undertreated patients experienced delayed CINV than appropriately treated patients (71.6% versus 49.5%; Table 2). In contrast, 89.2% of patients received prophylaxis for acute CINV that was in compliance with ASCO guidelines. Our findings highlight that, unlike recommendations for the prevention of acute CINV, guidelines for the prevention of delayed CINV are still not widely implemented [10–13]. The prevention of delayed CINV is recognized to be of equal importance to the need to prevent acute CINV [2, 14] and, in particular, improved management of delayed CINV remains a priority. However, our results also show that even optimal management of CINV is not expected to fully achieve the therapeutic goal: nearly 50% of our patients who were managed according to guidelines experienced delayed CINV (Table 2). Our results also show that chemotherapy-induced nausea affects far more patients than vomiting (62.5% versus 26.0%; Table 3), which is in line with the notion that the introduction of 5-HT3 receptor antagonists has effected a reduction in the frequency of vomiting, but not in the frequency of nausea episodes [15]. This is noteworthy since patients, in contrast to physicians
and nurses, typically view control of nausea as more important than control of vomiting [16, 17]. In aggregate, these findings highlight the limited effectiveness of currently available antiemetic prophylaxis, particularly for delayed CINV and nausea [18], and a need for new pharmacological agents with improved efficacy. A further aim of this study was to estimate costs imputable to CINV so as to provide information for future decisions on resource allocation. We separately present data on consumption of healthcare resources (Table 4) and cost data (Tables 5 and 6), to facilitate the application of our findings to cost analyses in other healthcare settings where unit costs may differ. Most line items shown in Table 4 and their relative contributions to overall cost probably lend themselves to extrapolation to other healthcare environments. Corresponding data from a similar analysis conducted in Canada were broadly comparable to our findings [3], including the small number of patients requiring hospitalization due to CINV. However, the small number of workdays lost, and the low indirect cost resulting from this observation, may not be readily applicable to other healthcare settings. In our sample, only three of those patients who experienced CINV and were expected to work lost a total of 4.4 paid workdays (~35 working hours). O’Brien et al. [3] found that, among 72 Canadian patients experiencing CINV, there were total losses of 198 h of paid employment and 409 h of unpaid employment, which explained the fact that, in their study, indirect cost accounted for about two-thirds of the total cost of CINV. Similarly, Lindley et al. [19] reported that 23% of their patients surveyed in a US cancer care center, “were unable to go to work due to emesis”.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Gastrointestinal
531 Table 2. Prophylactic antiemetic treatments Antiemetic agents used for prophylaxis of acute emesis (n = 204 cycles)a 5-HT3/corticosteroid
Cycles, %
Percentage of cycles with acute CINV
Yes (89.2)
33.0 (N and/or V)
75.0 9.8
5-HT3/corticosteroid/D2
4.4
5-HT3
6.8
5-HT3/H1
0.5
5-HT3/D2
0.5
D2
0.5
Corticosteroid
2.0
Corticosteroid/D2
0.5
9.3 (V)
No (10.8)
31.8 (N and/or V) 9.1 (V)
0.0
Total No. of patients, %
100
(100)
Antiemetic agents used for prophylaxis of delayed emesis (n = 201 cycles)a
Cycles, %
Antiemetic regimen in agreement with ASCO guidelines for delayed emesis? (%)
Percentage of cycles with delayed CINV
Yes (49.2)
49.5* (N and/or V)
Corticosteroid/5-HT3
6.9
Corticosteroid/5-HT3/H1
0.5
Corticosteroid/5-HT3/D2
30.8
Corticosteroid
8.0
Corticosteroid/H1
0.5
Corticosteroid/D2 5-HT3
2.5 21.9
5-HT3/D2
2.5
5-HT3/H1
0.5
D2
10.9
D2/H1
No (50.8)
0.5 b
1.0
No prophylaxis
13.5
Corticosteroid
11.1* (V)
Total No. of patients, %
100
71.6 (N and/or V) 33.3 (V)
(100)
a
Four and seven cycles, respectively, with missing values regarding onset of acute or delayed symptoms. Patients in this group received a cisplatin-containing regimen. *P >0.005 versus patients receiving antiemetic regimen not in agreement with ASCO guidelines. ASCO, American Society of Clinical Oncology; CINV, chemotherapy-induced nausea and vomiting; D2, dopamine2-receptor antagonists; H1, histamine1-receptor antagonists; 5HT3, 5-hydroxytryptamine3-receptor antagonists; N, nausea; V, vomiting.
b
The discrepancy may be attributable to differences in data accrual (e.g. “unable to go to work” need not be the same as counting actual workdays lost) or valuation of non-marketed labor time lost [3]. However, it might also be ascribed to true differences in local practice regarding sick leave. Most of the patients in our sample were on sick leave for the period during and after chemotherapy, irrespective of the occurrence of nausea or vomiting. This, and the fact that a majority of our patients were retired or unemployed, necessarily limited the impact of CINV on lost workdays in our study. Hence, this finding may be limited in scope to German settings. Quantification of costs imputable to CINV is not straightforward in a setting where some patients receive chemotherapy in hospitals while others are seen in office practices. Unit prices are not always
consistent across settings because hospitals often receive special discounts from manufacturers. Even more complicating, though, is the reimbursement structure currently in force in Germany. This implies, for example, that hospitals receive a lump sum per day of hospitalization from the third party payer (sick funds). The sum is not related to the actual amount of, for example, antiemetic medication or staff time required. Hence, when hospital patients require additional CINV treatment, the hospital provider, but not the third party payer, incurs extra drug and personnel costs. In contrast, when patients receive rescue medication in an office setting, the third party payer will reimburse drugs, while the provider incurs no drug costs. This necessitates separate calculation of costs for different perspectives and patient groups, and, for quantification of average cost per patient incurred by the third
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
5-HT3/corticosteroid/H1
No prophylaxis
Antiemetic regimen in agreement with ASCO guidelines for acute emesis? (%)
532 Table 3. Frequency and severity of acute and delayed chemotherapy-induced nausea and vomiting (n = 208 cycles) Grade of nausea and vomiting
Acute (≤24 h post chemotherapy), n (%)a,c
Delayed (>24 h post chemotherapy), n (%)b,c
Total No. of cycles, n (%)
(n = 204)
(n = 201)
(n = 208)
Vomiting by NCI grade 1–2
22 (10.8)
3–4
42 (20.9)
45 (21.6)
6 (3.0)
6 (3.0)
7 (3.4)
28 (13.7)
48 (23.9)
52 (25.0)d
Mild
17 (8.3)
41 (20.3)
48 (23.1)
Moderate
22 (10.7)
44 (21.9)
44 (21.2)
Severe
27 (13.2)
33 (16.4)
37 (17.8)
All grades
66 (32.3)
118 (58.7)
129 (62.0)e
67 (32.8)
122 (60.7)
134 (64.4)
All grades Nausea by severity
a
Including patients with or without delayed symptoms. Including patients with or without acute symptoms. c Four and seven cycles, respectively, with missing values regarding onset of acute or delayed symptoms. d Two cycles with missing values regarding severity of vomiting; in total, 54 cycles (26.0%) with vomiting. e One cycle with missing values regarding severity of nausea; in total, 130 cycles (62.5%) with nausea. NCI, National Cancer Institute [6]. b
Table 4. Utilization of healthcare resources and workdays lost due to CINV (total number of evaluable cycles, n = 208) Type of resource consumption (units)
No. of cycles patients utilizing resources/ reporting lost workdays, n (%)
Total No. of units consumed/full workdays lost
Rescue medication i.v. (dosages)
13 (6.3)
29
Rescue medication oral & rectal (dosages)
20 (9.6)
270
Office physician visits (individual visit)
23 (11.1)
28
Hospital outpatient visits (individual visit)
6 (2.9)
6
(Extended) hospitalization (days)
1 (0.5)
3
Ambulance transport (rides)
1 (0.5)
1
OTC/alternative medication (purchases)
11 (5.3)
15
Taxi/public transportationa (rides)
15 (7.2)
15
Extra home help/childcare (episodes)
3 (1.4)
NE
Workdays lost (full-day equivalents)b
3 (10.0)
a
a
4.4
a
Patient out-of-pocket payment. Based on 30 cycles where patients are expected to work (patients are not retired and had no sick leave due to chemotherapy). CINV, chemotherapy-induced nausea and vomiting; i.v., intravenous; NE, not evaluated; OTC, over-the-counter medication.
b
party payer, requires data from different settings. We therefore attempted to recruit a case mix of patients from different types of treatment centers. The major cost driver from a sick funds’ perspective was the acquisition cost of antiemetic drugs for prophylaxis and treatment of CINV ($40.27 of a total cost of $48.66 per patient; Table 5); as described earlier, this was incurred exclusively in the office-based setting. Given the German reimbursement structure, it is not surprising that the main cost driver for hospital providers was staff time and, to a small extent, drug administration devices and other materials ($29.94 of a total cost of $48.27 per patient; Table 6).
More importantly, nearly one-third of the total cost per patient imputable to CINV was due to the treatment of CINV. This was true for both the third party payer and the hospital-provider perspectives, and represents the potential for cost offset that might be saved by improved management of CINV. Out-of-pocket payments by patients and indirect costs represent only a comparatively small fraction of the total cost of CINV in our sample. We have not estimated CINV cost from the perspective of officebased physicians due to a lack of data on staff time requirements and valuation of staff time. Hourly wage rates in office-based
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Any nausea and/or vomiting
533 Table 5. Cost of CINV incurred by third party payers (sick funds) and patients, and indirect costsa Type of cost
Mean cost per treatment cycle per patient in 2002,$ Cycles with CINV (n = 134)
All cycles (n = 208)
Direct cost incurred by third party payer 7.22 ± 30.07 [0.00]
4.65 ± 11.25 [0.00]
2.99 ± 9.29 [0.00]
CINV-attributable (extended) hospitalizations
5.34 ± 61.80 [0.00]
3.44 ± 49.60 [0.00]
Ambulance transport
3.04 ± 35.22 [0.00]
1.96 ± 28.27 [0.00]
Subtotal: costs of treatment of CINV
24.23 ± 119.59 [0.00]
15.61 ± 96.56 [0.00]
Medication to prevent CINV (prophylaxis)c
45.37 ± 65.39 [0.00]
33.05 ± 57.07 [0.00]
Total: costs of prophylaxis plus treatment
69.60 ± 135.18 [33.21]
48.66 ± 113.01 [0.00]
Direct out-of-pocket costs incurred by patients OTC/alternative medication
1.46 ± 7.06 [0.00]
0.94 ± 5.71 [0.00]
Taxi/public transportation
0.79 ± 2.36 [0.00]
0.51 ± 1.93 [0.00]
Extra home help/childcare
2.39 ± 13.27 [0.00]
1.54 ± 10.65 [0.00]
Total
4.64 ± 27.90 [0.00]
2.99 ± 22.47 [0.00]
3.05 ± 22.74 [0.00]
1.97 ± 18.28 [0.00]
Indirect costs Workdays lost—total cost per patient Direct cost—third party payer (cost of prophylaxis plus treatment)
69.60 ± 135.18 [33.21]
48.66 ± 113.01 [0.00]
Direct cost—patient
4.64 ± 27.90 [0.00]
2.99 ± 22.47 [0.00]
Indirect cost
3.05 ± 22.74 [0.00]
1.97 ± 18.28 [0.00]
Total
77.30 ± 146.59 [36.14]
53.62 ± 122.72 [0.00]
a
Costs are given as the mean per treatment cycle for two different denominators: cycles associated with CINV experience, and all cycles. Mean ± SD [median] costs are presented because the data were not normally distributed. b Includes only rescue medication administered outside hospital, because drugs administered in-hospital are included in hospital day rates. c Cost for prophylaxis in patients receiving chemotherapy at the hospital is calculated at $0, because cost for prophylaxis is included in the flat rate third party payers reimburse for chemotherapy treatment. CINV, chemotherapy-induced nausea and vomiting; OTC, over the counter medication.
settings are dependent on the income of individual office practices. Hence, they are not publicly available and may vary considerably across centers. Even though we may have captured an almost complete spectrum of costs directly attributable to CINV, we are also not presenting cost data from a societal perspective. We recognise that this is a shortcoming of our analysis [4], since resource allocation decisions should ideally be made from a societal perspective [20]. However, the complexity of the cost structure in Germany, which may result in double counting of individual line items and a lack of data, virtually preclude calculation of this perspective. For example, data on the true cost per day in hospital or per outpatient visit have not been published or otherwise available to us. Given these limitations, the best ‘proxy’ estimate of societal cost of CINV is likely obtained by adding the costs incurred by third party payers and patients, and indirect costs. This yields a total cost estimate of $53.62 per patient per treatment cycle (Table 5). To provide further insight into the relative influence of different patient or treatment characteristics on the overall cost of CINV, we constructed several subgroups of patients. This analysis was
limited to patients treated in hospital settings from the provider’s perspective. The rationale for excluding patients treated in physician offices was the difference in cost structure between the settings, and the fact that allocation of patients to the different settings was non-random, and the lack of comparability of the groups with regard to type of malignancy and chemotherapy (Table 1) would have confounded subgroup analyses from the combined populations. Not surprisingly, cisplatin-containing chemotherapies were associated with higher average costs of prophylaxis and treatment of CINV than non-cisplatin regimens (Table 7). Subgroup analyses further show that overall cost of CINV was associated with presence of vomiting, presence of delayed CINV, and inadequate prophylaxis for delayed CINV (as indicated by a lack of adherence with ASCO guidelines for delayed CINV, although this did not reach the level of significance; Table 7). These findings are directly relevant to patient-management decisions, as they identify subgroups of patients for whom improved CINV prophylaxis is likely to result in the highest levels of cost savings or cost offset.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
11.20 ± 36.90 [0.00]
CINV-attributable outpatient physician visits
Medication to treat CINVb
534 Table 6. Cost of CINV incurred by provider (hospital)a Mean cost per treatment cycle per patient in 2002, $ Cycles with CINV (n = 73) Medication to treat CINV (rescue medication)b
1.32 ± 6.17 [0.00]
All cycles (n = 137) 0.71 ± 4.54 [0.00] 0.37 ± 1.44 [0.00]
25.30 ± 109.13 [0.00]
13.48 ± 80.41 [0.00]
Subtotal
27.31 ± 109.20 [0.00]
14.55 ± 80.62 [0.00]
Medication to prevent CINV (prophylaxis)
17.38 ± 9.73 [15.98]
17.25 ± 9.52 [15.98]
Prophylactic drug administration devices and staff time
16.51 ± 5.07 [20.60]
16.46 ± 5.07 [20.60]
Subtotal
33.89 ± 11.42 [36.13]
33.72 ± 11.09 [36.58]
Total
61.20 ± 109.68 [38.08]
48.27 ± 81.33 [36.58]
a
The provider perspective is not calculated for the office-based treatment settings, because data regarding staff time and administration costs are lacking. Costs are given as means per treatment cycle for two different denominators (cycles associated with CINV experience, and all cycles), because patients not experiencing CINV during a particular cycle still incurred the costs of prophylaxis. Mean ± SD [median] costs are presented because the data were not normally distributed. b Includes only rescue medication administered inside hospital.
How do our findings compare with results from similar studies? We found two reports concerned with CINV-attributable costs in Germany. Ihbe-Heffinger et al. [21] quantified the costs of drug acquisition and administration associated with the control of CINV from a hospital perspective. In their cohort of breast cancer patients receiving moderately emetogenic chemotherapy, total costs (comprising prophylaxis and treatment of CINV) ranged from $35 to $73, depending on the type of antiemetic drugs used. This is in good agreement with our data from the hospital treatment setting. Herold and Hieke [5] compared the cost of toxicity of standard chemotherapy given to patients with non-Hodgkin’s lymphoma in several countries. In their retrospective analysis, the average cost to third party payers for managing CINV in Germany was $358 per patient for a complete course of treatment, which typically comprised six cycles of chemotherapy. Hence, average cost per patient per cycle amounts to approximately $60, which is also in line with our findings of $48.66 in a more heterogeneous patient population. It is difficult to compare our findings on costs of CINV with other reports because published information is limited. To the best of our knowledge, there is only one study that is similar in design to ours [3], but it predates the availability of 5-HT3 receptor antagonists, and relates to the Canadian healthcare environment. A number of methodological aspects of this study deserve discussion. Our study was prospective and was based on specifically designed questionnaires; all data acquisition was performed or monitored by experienced and trained personnel. These study features allowed us to capture cost items with a degree of accuracy and completeness that may not usually be available from retrospective chart review or medical claims data. This study was not restricted to a particular cancer type and we have deliberately enrolled a heterogeneous group of cancer
patients with a variety of malignancies. Patients received a broad range of chemotherapies, all of which are considered highly emetogenic [1]. This is expected to make our results relevant for extrapolation to most cancer patient populations on highly emetogenic treatments. Our study may have systematically underestimated the true cost of CINV in at least two ways. Firstly, and as discussed above, the results indicate that management of CINV may have been suboptimal, and improved management might result in cost savings that are not apparent from our study, since we did not consider the cost of chemotherapy administration. Stewart et al. [22] observed retrospectively that the improved control of CINV by the introduction of 5-HT3 antagonists enabled the administration of larger doses of chemotherapy in shorter periods of time. As a result, some multi-day regimens were successfully converted to singleday regimens, and prolonged infusions given on an inpatient basis were converted to relatively short outpatient infusions. Hence, our study may have underestimated the resources wasted (i.e. not saved) due to the continuing high rate of CINV. Conceivably, more savings or cost offsets could be achieved by reducing the frequency and severity of CINV, either by optimizing the management with existing medication or by use of novel antiemetic agents with improved efficacy. Secondly, we have not factored in less tangible and hidden costs, such as unpaid family labor and time in caring for patients, or patients’ inability to perform household work, which are often ignored but may not be insignificant [23, 24]. Nevertheless, the present results document that a substantial proportion of patients continue to experience CINV. This entails not only clinical but also economic consequences, and highlights a continuing need for improved utilization of existing antiemetic agents and for new, more efficacious treatments. Our data indicate
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
0.69 ± 1.93 [0.00]
Staff time and material consumption associated with managing episode of CINV
Rescue-drug administration devices
535 Table 7. CINV-attributable cost in 2002$ by subgroup (hospital-provider perspective)a n
Prophylaxis
137
33.72 ± 11.09
14.55 ± 80.62
48.27 ± 81.33
Non-cisplatin
55
26.67 ± 12.70
7.11 ± 18.57
33.79 ± 22.11
1.00 (ref.)
Cisplatin
82
38.44 ± 6.54
19.54 ± 103.05
57.98 ± 102.68
1.72*
No CINV
64
33.52 ± 10.80
0.00
33.52 ± 10.80
1.00 (ref.)
Nausea only
44
34.28 ± 10.30
5.45 ± 10.17
39.73 ± 19.93
1.19
Emesis (any level)
29
33.30 ± 13.11
60.48 ± 169.15
93.77 ± 169.28
2.80*
Emesis level 1–2
25
33.04 ± 11.88
25.21 ± 23.08
58.26 ± 25.93
1.74
Emesis level 3–4
3
37.67 ± 25.67
374.50 ± 478.43
412.17 ± 472.22
12.29
Delayed, no
67
33.36 ± 10.69
33.36 ± 10.69
1.00 (ref.)
Delayed, yes
66
33.81 ± 11.69
29.25 ± 114.61
63.06 ± 115.03
1.89*
Delayed, yes
100
37.04 ± 8.21
6.45 ± 16.31
43.49 ± 21.08
1.00 (ref.)
Delayed, no
37
24.73 ± 12.87
36.46 ± 152.14
61.19 ± 153.40
1.41
All cycles
Treatment
Total (prophylaxis + treatment)
Normalized total cost NA
Cisplatin-containing chemotherapies
Level of acute or delayed CINVb
0.00
Adherence to ASCO guidelines for prophylaxis of delayed CINV
Mean ± SD costs are indicated as per patient per cycle. To facilitate interpretation of cost differences between different patient subgroups, costs are also expressed as multiples of a designated reference subgroup. Note that these analyses include only results from patients treated in a hospital setting; patients treated in an office-based care setting were excluded. b One cycle with missing values regarding severity of emesis. c Four cycles with missing values regarding time of onset of symptoms. *P <0.05 versus reference group. CINV, chemotherapy-induced nausea and vomiting; NA, not applicable; ref., reference. a
that the greatest improvement in patient care and potential for cost offset may be realized by avoidance of delayed CINV.
Acknowledgements This study was supported by Merck & Co., Inc., Whitehouse Station, NJ, USA.
References 1. Hesketh PJ, Kris MG, Grunberg SM et al. Proposal for classifying the acute emetogenicity of cancer chemotherapy. J Clin Oncol 1997; 15: 103–109. 2. Gralla RJ, Osoba D, Kris MG et al. Recommendations for the use of antiemetics: evidence-based, clinical practice guidelines. J Clin Oncol 1999; 17: 2971–2994. 3. O’Brien BJ, Rusthoven J, Rocci A et al. Impact of chemotherapy-associated nausea and vomiting on patients’ functional status and on costs: survey of five Canadian centres. Can Med Assoc J 1993; 149: 296–302. 4. Uyl-de Groot CA, Wait S, Buijt I. Economics and health-related quality of life in antiemetic therapy: recommendations for trial design. Eur J Cancer 2000; 36: 1522–1535. 5. Herold M, Hieke K. Costs of toxicity during chemotherapy with CHOP, COP/CVP, and fludarabine. Eur J Health Econom 2002; 3: 166–172.
6. Cancer Therapy Evaluation Program. Common toxicity criteria, version 2.0. DCTD, NCI, NIH and DHHS, March 1998; [online] www.ctep.cancer.gov/ forms/ctcv2nom-4–30–99-final3.pdf (date last accessed: July 2003). 7. Hannoveraner Konsens Gruppe. Dt. Empfehlungen zur gesundheitsökonomischen Evaluation—Revidierte Fassung des Hannoveraner Konsens. Gesundheitsökonomie Qualitätsmanagement 1999; 4: A62–A65. 8. Scherbel G, Bodenmüller-Kroll R, Matthias M et al. Antiemetische Therapieformen Kosten und Effektivität im Vergleich. Arzneimitteltherapie 1994; 6 (Suppl 17): 1–8. 9. Beckwith MC, Mullin S. Prevention and management of chemotherapyinduced nausea and vomiting, part 1. Hospital Pharmacy 2001; 36: 67–82. 10. Mertens WC, Higby DJ, Brown D et al. Improving the care of patients with regard to chemotherapy-induced nausea and emesis: the effect of feedback to clinicians on adherence to antiemetic prescribing guidelines. J Clin Oncol 2003; 21: 1373–1378. 11. Roila F, Donati D, Tamberi S et al. Delayed emesis: patterns, prognostic factors and optimal treatment. Support Care Cancer 2002; 10: 88–95. 12. Warr D. Standard treatment of chemotherapy-induced emesis. Support Care Cancer 1997; 5: 12–16. 13. Rubenstein EB, Litke SE, Jose MA. Chemotherapy-induced nausea and emesis in the 5HT3 era: a national survey of hematologists/oncologists. Proc Am Soc Clin Oncol 2002; 21: 260 (Abstr 2860). 14. Roila F. Dexamethasone alone or in combination with ondansetron for the prevention of delayed nausea and vomiting induced by chemotherapy. Italian Group for Antiemetic Research. N Engl J Med 2000; 342: 1554– 1559.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
Presence of delayed CINVc
536 20. Russell LB, Gold MR, Siegel JE et al. The role of cost-effectiveness analysis in health and medicine. JAMA 1996; 276: 1172–1177. 21. Ihbe-Heffinger A, Kuhn W, Sattler D et al. Cost-minimization analysis for the use of 5-HT3 receptor antagonists in moderately emetogenic chemotherapy. A prospective pharmacoeconomic evaluation in German hospitals. Eur J Cancer 2000; 36 (Suppl 3): 18. 22. Stewart DJ, Dahrouge S, Coyle D, Evans WK. Cost of treating and preventing nausea and vomiting in patients receiving chemotherapy. J Clin Oncol 1999; 17: 344–351. 23. Hayman JA, Langa KM, Kabeto MU et al. Estimating the cost of informal caregiving for elderly patients with cancer. J Clin Oncol 2001; 19: 3219– 3225. 24. Lindley CM, Hirsch JD. Nausea and vomiting and cancer patients’ quality of life: a discussion of professor Selby’s paper. Br J Cancer 1992; 66 (Suppl XIX): S26–S29.
Downloaded from http://annonc.oxfordjournals.org/ at University of Texas at San Antonio on August 22, 2014
15. Roscoe JA, Morrow GR, Hickok JT, Stern RM. Nausea and vomiting remain a significant clinical problem: trends over time in controlling chemotherapy-induced nausea and vomiting in 1413 patients treated in community clinical practices. J Pain Symptom Manage 2000; 20: 113– 121. 16. Bonneterre J, Hecquet B, Atenis A et al. How do patients and physicians decide which antiemetic is best in a cross over study? Proc Am Soc Clin Oncol 1991: 10: 323 (Abstr 1136). 17. Lecomte S, Lemenager M, Duquenne I et al. Quality of life in breast cancer patients receiving adjuvant chemotherapy. Proc Am Soc Clin Oncol 1995: 14: 308 (Abstr 900). 18. Morrow GR, Hickok JT, Rosenthal SN. Progress in reducing nausea and emesis. Comparisons of ondansetron (Zofran), granisetron (Kytril), and tropisetron (Navoban). Cancer 1995; 76: 343–357. 19. Lindley CM, Hirsch JD, O’Neill CV et al. Quality of life consequences of chemotherapy-induced emesis. Qual Life Res 1992; 1: 331–340.