Presented in part at the American Society of Clinical Oncology 2006 Annual Meeting, Atlanta, Georgia, June 2–6, 2006.
Despite recommendations to immunize all patients at an increased risk of influenza complications, the vaccine utilization among high-risk nonelderly adults remains low and its cost-effectiveness is unclear. In the current study, the authors analyzed the cost-effectiveness of influenza vaccination in working-age (ages 20–64 years) cancer patients.
The authors developed a decision-analytic model, from the societal perspective, using epidemiologic, vaccine effectiveness, resource utilization, cost, survival, and utility data from published sources, supplemented with data collected from the authors' own institutional accounting system. Two strategies were compared: influenza vaccination of working-age cancer patients and no vaccination. The base-case patient was assumed to be a 51-year-old cancer patient (the mean age for the National Cancer Institute's Surveillance, Epidemiology, and End Results [SEER] population of working-age patients within 5 years of cancer diagnosis).
The effectiveness of the influenza vaccine was 6.02 quality-adjusted life-years (QALYs) at a cost of $30.10. The effectiveness of the no vaccination strategy was 6.01 QALYs at a cost of $27.86. Compared with the no vaccination strategy, the incremental cost-effectiveness ratio of vaccinating working-age cancer patients would be $224.00 per QALY gained. Using the benchmark of $50,000 per QALY, the model was only sensitive to changes in cancer survival (threshold of 2.8 months).
Influenza infection is a major cause of morbidity and mortality around the world. In the U.S., epidemics of influenza annually affect 5% to 20% of the population, resulting in greater than 200,000 hospitalizations and 36,000 deaths.1, 2 The rates of serious morbidity and mortality are particularly high among individuals who have medical conditions that place them at increased risk of complications from influenza, such as underlying malignancies. Some studies have shown an increased incidence and duration of influenza infections among cancer patients.3, 4 In addition, influenza-related respiratory infections in patients with cancer have often been associated with costly hospitalizations, delays in potentially life-saving therapy, and death.5, 6 As a result, the Advisory Committee on Immunization Practice (ACIP) currently recommends annual vaccination against influenza for all individuals at increased risk of influenza-related complications, including those who are at risk from underlying immunosuppressive disease.7
However, despite the availability of a suitable vaccine and recommendations to immunize all patients at increased risk of complications from influenza, the rate of vaccination among high-risk adults age <65 years remains low. According to data from the 2004 National Health Interview Survey (NHIS), only 35% of all high-risk adults ages 18 to 64 years reported having received influenza vaccination compared with 65% of the elderly population.7 In a surveillance study focused exclusively on cancer patients, nonelderly adult cancer patients also were found to demonstrate a low rate of influenza immunization compared with their elderly counterparts (17% vs. 53%, respectively).8 Also of note is that the vaccine insurance coverage picture for young adults is far less positive than the coverage for the elderly. Recent estimates indicate that 59 million American adults ages 18 to 64 years have private insurance that does not include immunization benefits and another 30 million lack health insurance altogether.9 As a result, a total of 89 million nonelderly adults—half of this age group—lack influenza immunization benefits, whereas nearly all elderly Americans have coverage for influenza vaccine under Medicare part B.9
In addition, it is important to note that the low rates of influenza vaccination among cancer patients in particular also may be influenced by the controversy over the effectiveness of the vaccine in this high-risk subpopulation. In a survey of pediatric oncologists, a substantial number of the respondents reported not routinely recommending the vaccine for their patients.10 The main concern is that a potential suboptimal immune response resulting from the underlying immunosuppressive disease and/or the effects of cancer therapy could compromise the effectiveness of the vaccine. Notwithstanding that, evidence exists suggesting that even though cancer patients' immune response to influenza vaccination might be attenuated, it can still provide protection against influenza infections.11 Therefore, we analyzed the cost-effectiveness of influenza vaccination in working-age adults (ages 20–64 years) who are at an increased risk of influenza-related complications from underlying malignancies. We hypothesized that even though the immune response of adult cancer patients to the influenza vaccination might be attenuated, the vaccine may still be cost-effective in this high-risk subpopulation.
MATERIALS AND METHODS
We developed a decision-analytic model to estimate the costs and health outcomes associated with influenza infections in working-age cancer patients under 2 different scenarios: inactivated influenza vaccination versus no influenza vaccination (Fig. 1). We derived the parameters for the decision tree by reviewing published articles on the effectiveness of the influenza vaccine, as well as on the epidemiology, resource utilization, cost, and quality of life associated with influenza infection among working-age cancer patients, supplemented with data collected from our institutional accounting system. The time horizon for the risk of influenza infection, protection from the vaccine, and cost of vaccination was 1 year because vaccination must be repeated annually due to changes in the influenza viral structure from year to year. The base-case patient for the model was a 51-year-old cancer patient within 5 years of diagnosis (the mean age for the 5-year prevalent cancer cases between 20–64 years of age from the National Cancer Institute's Surveillance, Epidemiology, and End Results [SEER] population).12 We developed the decision-analytic model and performed all the analyses using DATA 4.0 software (Decision Analysis by TreeAge Software Inc., Williamstown, MA). We adopted the benchmark of U.S. $50,000 per quality-adjusted life-year (QALY) as the threshold for acceptable cost-effectiveness and examined the robustness of the results by performing 1-way sensitivity analyses of plausible ranges for the following key parameters: the rate of influenza infection, the effectiveness and cost of the influenza vaccination, and the patient's life expectancy. In addition, we also performed 2-way sensitivity analysis on the 2 parameters most subject to uncertainty and variability (ie, the rate of influenza infection and vaccine effectiveness) to account for the seasons in which the circulating strains and the vaccine strains are a poor match, which could potentially render the vaccine negligibly effective. The study was submitted to our Institutional Review Board and was determined to be exempt from review.
We estimated the probability of influenza infection based on published data from the Tecumseh community study, in which approximately 10% of the community in Tecumseh, Michigan and its surrounding rural area was followed for acute respiratory illness between 1965 and 1971 and 1976 through 1981.1 We calculated the annualized incidence of influenza infection based on the reported age-adjusted frequency of influenza infection observed from 1977 through 1981 in the Tecumseh study participants age ≥20 years (Table 1).1
Table 1. Base-case Probabilities and Underlying Assumptions
ED indicates Emergency Department; OTC, over-the-counter.
Risk of receiving a viral respiratory antigen test in the ED
The incident cancer cases (22% of the 5-year cancer prevalence) would receive a viral respiratory antigen test in the ED because they are likely to have active disease and therefore be more thoroughly investigated when presenting to the ED with influenza-like symptoms, particularly fever.
Risk of influenza-related office visit utilization (among patients not requiring hospitalization or ED visits)
The incident cancer cases (22% of the 5-year prevalent cancer cases) would require an office visit because they are likely to be immunosuppressed (due to the underlying disease and/or treatment) and therefore are oriented to seek medical care if experiencing influenza symptoms.
Risk of OTC medication utilization (among influenza-infected patients not requiring hospitalization)
All influenza-infected cancer patients not requiring hospitalization would utilize OTC medication for the relief of flu symptoms during the initial 5 days of influenza infection.
Risk of antiviral medication utilization (among influenza-infected patients not requiring hospitalization)
The incident cancer cases (22% of the 5-year prevalent cancer cases) are likely to be immunosuppressed (due to the underlying disease and/or treatment) and therefore to receive antiviral medication in case of a diagnosed influenza infection.
We obtained an estimate of the effectiveness of the influenza vaccine in working-age cancer patients by performing a meta-analysis of published studies of seroconversion in adult cancer patients after immunization with split virion influenza vaccine.13–18 Based on the threshold hemagglutination inhibition (HAI) antibody titer of 32 to 40, which represents the level at which approximately half of the individuals would be protected from influenza infection and disease, we assumed that seroconversion would equate to 50% of clinical effectiveness (Table 1).19, 20 To obtain the overall average protection rate of influenza vaccine in adult cancer patients, we determined that both the sample size and the quality of the studies of vaccine seroconversion in cancer patients should be taken into account. We graded the quality of the studies as: 1) lower quality if the antibody level threshold used to measure seroconversion was a serum HAI titer of <40 or a 4-fold or greater increase in the HAI antibody titer; and 2) higher quality if the threshold level used was an HAI titer of ≥40. Based on these criteria, we defined the quality-adjusted protection rate, Poverall as follows:
in which qsj is the quality score for the jth study, nj is the sample size for the jth study, and pj is the protection rate observed in the jth study, which was obtained by multiplying the seroconversion rate observed in the jth study by 50%. Because some of the study sample sizes were small, we do not think that the Gaussian approximation to binomial distribution would be applicable. As such, we used a modification of the Bernard and Bravo method to obtain an estimate of the 95% confidence interval for the overall quality-adjusted protection rate, Poverall, by using the bootstrap Efron.21, 22
The probability of hospitalization was based on reported estimates of annual influenza-related hospitalizations of cancer patients.6 We divided the reported annual number of influenza-related admissions among cancer patients ages 20 to 64 years by the annual number of cancer patients in this age group estimated to become infected with the influenza virus, which we obtained by applying the annualized incidence of influenza observed in the Tecumseh study to the 5-year cancer prevalence for the SEER population ages 20 to 64 years (Table 1).1, 6, 12
We derived the probability of emergency department (ED) utilization from the rate of influenza-related ED visits that required hospitalization in the general population reported in the study conducted by Cox et al.23 We obtained the total number of influenza-infected cancer patients who required an ED visit by dividing the estimated number of cancer patients who were hospitalized throughout EDs nationwide in the study by Cooksley et al.6 by the reported proportion of ED visits that required hospitalization in the study by Cox et al.23 We subsequently excluded the patients who were hospitalized through the ED in the study by Cooksley et al. and divided the resulting number by the annual number of working-age cancer patients estimated to become infected with the influenza virus (Table 1).1, 6, 12
Survival and Quality of Life Data
We estimated survival using the declining exponential approximation of life expectancy (DEALE) method.24 To obtain the estimated life expectancy of the base-case cancer patient, we applied the cancer-specific mortality rate of those ages 20 to 64 years (derived from SEER data) to the life expectancy of a 51 year-old person (obtained from U.S. vital statistics life tables).12, 25 For the influenza-infected patients requiring hospitalization, we in addition applied the reported influenza-related inhospital mortality rate for cancer patients ages 18 to 64 years reported by Cooksley et al.6 We then discounted life expectancy at an annual rate of 3%.
We obtained the utility weight for influenza infection from a published study in which the utility weight for a day with influenza symptoms was derived from the Quality of Well Being Scale.26 We applied the reported utility weight for a day with influenza symptoms to the reported mean number of days associated with influenza infection among adult participants of the Tecumseh study.1, 26 Finally, we obtained the utility measure for malignancies from a study in which a large, randomly selected sample of cancer patients (all causes) was asked to evaluate their current state of health by the time-tradeoff method.27
Resource Utilization and Cost Data
We obtained cost estimates of influenza-related hospitalizations from the reported mean cost per influenza-related stay for cancer patients ages 18 to 65 years and inflated it to 2005 U.S. dollars using the Consumer Price Index (CPI) for medical care (Table 2).6, 28 With regard to the cost of ED and office visits, we conducted a microcosting analysis to identify and quantify the resources used by a cancer patient infected with the influenza virus who required emergency care and office visits by surveying oncologists at our comprehensive cancer center and within the community. We then estimated costs by multiplying the number of each of the resources identified as being required during influenza-related ED and office visits by their unit costs (in 2005 U.S.dollars) obtained from our institutional accounting system, which uses cost-to-charge-ratios for determining costs (Table 2). The cost of the influenza vaccine administration also was obtained by microcosting analysis through interviewing the providers directly involved with patient care and with the employee vaccination program at our institution. With regard to the costs of pharmaceuticals and the inactivated influenza vaccine, they were based on the pharmaceutical industry's average wholesale price as listed in the 2005 Red Book (Table 2).29
Table 2. Influenza-related Costs (Per Patient) Among Working-age Cancer Patients
Cost per 2005 U.S.$
ED indicates Emergency Department; OTC, over-the-counter.
Indirect costs associated with work absenteeism were based on the mean per capita income by age obtained from the U.S. Census 2000, inflated to 2005 U.S. dollars using the general CPI and age-adjusted to the 5-year cancer prevalence for the SEER population ages 20–64 years (Table 2).12, 28, 30 We assumed that influenza-infected patients who required hospitalization would be absent from work for the length of the hospitalization, which was based on the reported mean length of stay for influenza-related admissions of working-age cancer patients.6 We also assumed that influenza-infected cancer patients who required an ED visit or an office visit would be absent from work for the day.
The meta-analysis of the 6 published studies of seroconversion in adult cancer patients after immunization with a split virion influenza vaccine produced a summary protection rate, Poverall, of .33 (95% Bootstrap confidence interval [95% CI], 0.27–0.38) (Table 1). Based on this estimate, the effectiveness of the inactivated influenza vaccination during an average influenza year would be 6.02 QALYs at a cost of U.S. $30.10 (Table 3), whereas the effectiveness of the no vaccination strategy would be 6.01 QALYs at a cost of U.S. $27.86. Therefore, compared with the no vaccination strategy, the incremental cost-effectiveness ratio of vaccinating working-age cancer patients would be $224.00 per QALY.
Table 3. Results of Sensitivity Analysis
Incremental cost/effectiveness (U.S. $/QALY)
QALY indicates quality-adjusted life-years; Dominant, when the vaccination strategy is both cheaper and more effective than the no vaccination strategy; 95% CI, 95% confidence interval; Poverall, overall quality-adjusted influenza vaccine protection rate in cancer patients.
Incidence of influenza
The lower age-adjusted annual rates observed in the Tecumseh study.1
No influenza vaccination
The upper age-adjusted annual rates observed in the Tecumseh study.1
No influenza vaccination
The lower limit of the 95% CI for Poverall
No influenza vaccination
The upper limit of the 95% CI for Poverall
No influenza vaccination
Average wholesale price of the inactivated influenza vaccine.29
No influenza vaccination
Cost of vaccination plus office visit.
No influenza vaccination
Cost-effectiveness threshold for duration of cancer survival.
No influenza vaccination
Life expectancy of a 51-year-old person (obtained from U.S. vital statistics life tables).30
No influenza vaccination
In 1-way sensitivity analyses using the benchmark of U.S. $50,000 per QALY, the decision model was robust to plausible changes in the values of influenza incidence, vaccine effectiveness, and vaccination price (Table 3). If values for influenza incidence vary through the reported age-adjusted lower and upper rates of influenza infection observed in adult participants in the Tecumseh study from 1977 through 1981,1 the vaccination strategy will be cost-effective, with a maximum cost per QALY of $12,117 (Table 3). It is interesting to note that the vaccination strategy becomes the dominant strategy (ie, both cheaper and more effective than the no vaccination strategy) when the incidence of influenza is ≥12.9%. In turn, if values for vaccine effectiveness vary through the 95% Bootstrap CI for the calculated rate of protection in cancer patients, the vaccination strategy will still be cost-effective with a maximum cost per QALY of $2305 (Table 3). It is also interesting to note that the incremental cost-effectiveness of the vaccination strategy would only reach the threshold of U.S. $50,000 per QALY when the vaccine effectiveness rate is equal to 0.04. In addition, even if all cancer patients seek their oncologists to get the vaccine (with the cost of an office visit being added to the vaccination cost), the influenza vaccination strategy would still be cost-effective at a cost of $16,330 per QALY (Table 3). In contrast, the model was sensitive to changes in cancer survival, with a survival cost-effectiveness threshold of 2.8 months (Table 3).
In a 2-way sensitivity analysis of vaccine effectiveness and the rate of influenza infection, vaccine effectiveness was allowed to vary from 0 to the upper limit of the 95% CI for the estimated rate of protection in cancer patients, while at the same time the incidence of influenza infection varied from 0 to the upper age-adjusted annual rate of infection observed in the Tecumseh study.1 As shown in the lower left corner of Figure 2, the only circumstances in which the incremental cost-effectiveness of the vaccination strategy would go beyond the U.S. $50,000 per QALY threshold are when both the effectiveness and the probability of influenza infection are concomitantly below the respective ranges used in the 1-way sensitivity analysis, or when either one is at its extreme low-end values (Figure 2).
Our estimates of the costs and health outcomes associated with influenza infections in vaccinated and unvaccinated cancer patients ages 20 to 64 years indicate that, based on the generally accepted cost-effectiveness benchmark of $50,000 per QALY, it would be cost-effective to vaccinate all patients in this age group who have been diagnosed with cancer within the previous 5 years and have a life expectancy of at least 3 months. Our base-case findings were relatively comparable to those reported on another high-risk subpopulation, the elderly population. According to the analysis conducted by the U.S. Office of Technology Assessment, inactivated influenza vaccination of the population age ≥65 years would be highly cost-effective, by producing net savings per year of healthy life gained.31 However, in contrast to the elderly, the rate of vaccination among high-risk working-age adults in general, and cancer patients in particular, remains very low.7, 8 In addition, a substantial number of oncologists report not routinely recommending the vaccine for their patients.10 In light of the results of the current study showing that immunization of working-age cancer patients diagnosed within the previous 5 years and who have a life expectancy of at least 3 months would cost $224.00 per QALY gained (which is well within the lower end of the range generally accepted as cost effective) it is crucial to identify and implement effective strategies to increase the vaccine utilization rate in this subpopulation. Based on previous research cited above, new research should focus on strategies to increase the recommendations of oncologists and primary care providers, as well as insurance reimbursement for influenza vaccination in this particular subpopulation.
However, the current study has several limitations. As with all studies that utilize meta-analysis, the validity of the summary estimate produced is limited by the heterogeneity of the populations studied. Our meta-analysis included a mixed population of adult patients with hematologic and solid neoplasms; 24.2% were adults with hematologic malignancies, whereas 75.8% were adults with solid tumors or lymphomas. As a result, the hematologic malignancies may have been overrepresented in our meta-analysis sample, which could have led to an underestimatation of the vaccine effectiveness in adult cancer patients in general because there is evidence that the vaccine response rates in patients with hematologic malignancies may be lower than that in patients with solid tumors.32 In addition, we have also assumed that all cancer patients within 5 years of a cancer diagnosis would have an attenuated immune response to the influenza vaccination, which in fact may only be the case for those patients who have active disease and/or are under (or have recently received) immunosuppressive cancer treatment. As a result of these limitations, we may have underestimated the cost-effectiveness of the influenza vaccine in working-age cancer patients diagnosed within 5 years.
Second, we did not include the outcomes associated with the adverse effects of the vaccine in our analysis, and by so doing, we may have underestimated the costs associated with the vaccination, which would have led to a corresponding overestimate of the vaccine cost-effectiveness. However, it is important to note that experience with influenza vaccinations in children with cancer (mostly leukemia patients) who are receiving highly immunosuppressive therapy, and therefore are at highest risk for vaccine adverse reactions, has shown a pattern of minimal transient local reactions and negligible systemic adverse effects after influenza vaccination.33–38
Third, we assumed that the rates of influenza infection and influenza-related emergency care utilization to hospitalization among cancer patients were the same as those of the general population, which may have led to an underestimation of the cost-effectiveness of the vaccine because there is evidence suggesting an increase in the incidence and severity of influenza infection among patients with cancer when compared with healthy controls (Table 1).3, 4
Lastly, we assumed that the rates of hospitalization and ED utilization among vaccinated patients were the same as those of unvaccinated patients (Table 1). In doing so, we may have underestimated the cost-effectiveness of the vaccine because there is evidence suggesting that influenza vaccination may reduce the severity of illness and the consequent utilization of healthcare resources when it does not prevent the infection from occurring.39–41 As a result, if future clinical trials of influenza vaccine in working-age cancer patients confirm a reduction in the severity of illness and the utilization of healthcare resources among those receiving vaccines who become infected with the influenza virus, the cost-effectiveness profile of the influenza vaccine in this subpopulation is likely to be even more favorable than the profile presented herein.
We conclude that, despite the attenuated immune response of cancer patients to influenza vaccination, the inactivated influenza vaccine is cost-effective for working-age cancer patients diagnosed within the previous 5 years and who have a life expectancy of at least 3 months. The results of the current study suggest that, from a societal perspective, all working-age cancer patients who are within 5 years of diagnosis and have a life expectancy of ≥3 months should be vaccinated against influenza.