Printer Friendly

A cost-effectiveness analysis of hepatitis B vaccine in predialysis patients.

Hemodialysis patients and staff remain at risk for infection from hepatitis B virus (HBV). The problem of HBV infection in hemodialysis is compounded by the large proportion of patients who become chronic carriers of the virus, increasing the possibility of spread to their contacts. (Szmuness, Prince, Grady, et al. 1974). In the 1970s, the Centers for Disease Control (CDC) (1977) issued guidelines to reduce the spread of HBV among dialysis patients and their contacts. These guidelines have helped to effectively diminish the incidence of HBV in dialysis patients by approximately 90 percent (Alter, Favero, Miller, et al. 1989). Despite adoption of these guidelines, as well as the availability of the first HBV vaccine, a large pool of susceptible patients remains and recurrent outbreaks of HBV infection have been reported (Alter, Favero, Miller, et al. 1989; Alter, Favero, and Maynard 1986).

Plasma-derived HBV vaccine (Heptavax B--Merck, Sharp & Dohme) was intended to reduce the prevalence of HBV infection. Response rates in normal vaccinees were reported to be greater than 95 percent with virtually complete protection for responders (Szmuness, Stevens, Harley, et al. 1980; Crosnier, Jungers, Courouche, et al. 1981a). Unfortunately, randomized trials of vaccine efficacy in dialysis patients reported conflicting results. Two European studies showed a reduction in the incidence of HBV infection in immunized dialysis patients (Crosnier, Jungers, Courouche, et al. 1981b; Desmyter, deGroote, Colaert, et al. 1983), but a large collaborative study in U.S. dialysis units showed no diminution in the incidence of HBV infection in vaccinated dialysis patients (Stevens, Alter, Taylor, et al. 1984). Lack of vaccine efficacy was attributed to its low immunogenicity among dialysis patients.

Poor antibody response to HBV vaccine in dialysis patients has been well documented (range 50-60 percent at one year) (Stevens, Alter, Taylor, et al. 1984; Bommer, Ritz, Andrassy, et al. 1983; Benhamou, Courouche, Jungers, et al. 1984; de Graeff, Dankert, de Zeewur, et al. 1985). An alternative immunization strategy would vaccinate patients with renal insufficiency prior to dialysis when such patients may be less immunologically compromised (Bommer, Gussendorf, Jilg, et al. 1984; Seaworth, Drucker, Starling, et al. 1988). Seaworth and colleagues (1988) demonstrated a response rate of 81 percent with standard plasma-derived vaccine among a group of predominantly hypertensive and diabetic patients with a mean serum creatinine of 290 mol/L. Whether this predialysis vaccination strategy represented a better alternative for immunizing dialysis patients remains to be established. In order to assess the cost effectiveness of a predialysis vaccine strategy we compared the cost and effectiveness of that strategy with the available alternative strategies of (1) vaccinating patients at the time of dialysis and (2) not vaccinating patients at all.

METHODS

DECISION ANALYSIS MODEL

We constructed a decision analysis model to compare three possible strategies for preventing HBV infection in the dialysis population. The model focuses initially on patients with renal insufficiency who are not yet dialysis dependent (predialysis group). The initial decision is made to vaccinate (strategy 1) or not to vaccinate predialysis patients. For patients not initially vaccinated, two options remain: vaccinate when the patients require dialysis (strategy 2), or never vaccinate (strategy 3). For each strategy, health outcomes occurring within three years were included in the analysis. We chose a three-year time frame for the analysis because a majority of the surviving predialysis and dialysis patients would require additional vaccination at that time to maintain antibody titers above CDC guidelines (U.S.

Department of Health and Human Services 1990).

Several assumptions were made to facilitate the analysis: (1) all dialysis staff receive the HBV vaccine (vaccination is highly efficacious for this population), (2) the duration of protection for vaccine responders is three years, (3) there are no adverse reactions to the vaccine, (4) patients are not screened for immunity before vaccination (screening has already been shown not to be cost effective in populations with this prevalence of HBV) (Mulley, Silverstein, and Dienstag 1982), (5) patients who develop an HBV infection do not acquire it again during the three-year time frame of the analysis, (6) surface antigen carriers pose a risk only to their immediate household contacts, and (7) predialysis patients who do not require dialysis within the time frame of the analysis do not acquire HBV infection. Within the three-year time frame, events were modeled in the context of three sequential one-year periods. The annual rates of progression to dialysis and of contracting HBV infection were assumed to be constant.

Estimates of vaccine efficacy and costs were based on a protocol for TABULAR DATA OMITTED vaccination that includes three standard doses of the plasma-derived vaccine given at zero, one, and six months beginning with dialysis for strategy 2. The predialysis protocol for patients with moderate renal insufficiency (i.e., serum creatinine of 290 mol/L) includes the standard three-dose schedule plus an additional dose given at dialysis.

VARIABLES USED IN THE ANALYSIS FOR VACCINE EFFICACY AND HEPATITIS OUTCOME

The probabilities for the important variables used in the analysis are listed in Table 1. We estimated the response to vaccination in dialysis patients to be 55 percent based on data from Stevens and colleagues' large randomized trial (Stevens, Alter, Taylor, et al. 1984). Based on the Seaworth, Drucker, Starling, et al. (1988) data, we used 80 percent as the vaccine response rate in predialysis patients. The incidence rate for HBV infection among unvaccinated dialysis patients used in the baseline analysis was based on the annual survey of all U.S. dialysis centers performed by Alter, Gavero, Miller, et al. (1989). The probability of contracting HBV was 0.2 percent per year or 0.6 percent over three years (Alter, Gavero, Miller, et al. 1989). The infection rate for dialysis patients who were successfully vaccinated was assumed to be 0.09 percent over three years, based on the ratio of infection in responders and nonresponders reported by Stevens, Alter, Taylor, et al. (1984). The baseline infection rate for unvaccinated dialysis patients may be an underestimate because 14 percent of patients in the dialysis centers surveyed were vaccinated. In addition, the data were self-reported by centers rather than actually observed. Consequently, we also calculated cost-effectiveness ratios using the incidence rate observed in Stevens' large trial of 2 percent for vaccine responders and 13 percent for nonresponders over three years. These incidence rates are probably an upper estimate, since the data are several years old and the centers involved in the trial may have had unusually high average incidences of HBV infection. Finally, we assumed that 75 percent of predialysis patients would require dialysis within three years (Mitch et al. 1976; Rutherford, Blondin, Miller, et al. 1977).

The terminal branches of the decision tree that end with HBV infection were expanded to account for the acute and chronic sequelae of infection. We estimated the frequencies of the four possible clinical outcomes of HBV infection (icteric, fulminant, subclinical, and chronic hepatitis) from natural history studies. The outcomes of HBV infection are different for dialysis patients (London, Drew, Lustbader, et al. 1977; Parfrey, Farge, Forbes, et al. 1985; Shustermann and Singer 1987; Snydman, Bregman, and Bryan 1974) compared to a normal population (Mulley, Silverstein, and Dienstag 1982; King 1982; Kohler, Arnold, Renschin, et al. 1984). There is a higher frequency of subclinical infection and a higher proportion of chronic carriers of the surface antigen among dialysis patients. HBV infection is also associated with hepatocellular carcinoma, but we did not include this outcome in the analysis due to the exceedingly low incidence of this complication in the United States among persons infected as adults. We did not consider severe adverse reactions from the vaccine in the analysis. Such reactions are rare, occurring with fewer than 1 in 100,000 vaccinations (Mulley, Silverstein, and Dienstag 1982). Last, we did not quantitatively account for the fact that HBV-infected dialysis patients may have significantly poorer outcomes after renal transplantation. Including the latter consideration would underestimate the effectiveness of the vaccine. However, there is controversy over whether or not renal transplant patients do indeed have poorer outcomes if they are HBV infected. In addition, only a minority of patients with end-stage renal disease actually receive transplantation, reducing the impact on the analysis (U.S. Renal Data Systems 1990).
Table 2: Frequency of Clinical Outcomes for Patients with
Hepatitis B Infection
 Dialysis Nondialysis
 Patients (%) Patients (%)
Acute infection
Subclinical 92 80
Icteric 6 20
Fulminant 2 |is less than~1(*)
Chronic infection
Recovery 22 90
Carrier 62 5
Chronic persistent 10 4
Chronic active 6 1
Fulminant infection
Recovery 30 30
Death 70 70
*0.1.


COST VARIABLES

The costs associated with diagnosing and treating HBV infection (over and above the costs of treating renal insufficiency) were calculated by multiplying an estimated rate of health services use (number of clinic visits, number of hospital days) by the unit cost of the health service (1987 dollars). Resource use rates (annual number of outpatient visits, length of hospital stays) were based on the assumptions of Mulley and colleagues regarding management of patients with HBV infection (Mulley, Silverstein, and Dienstag 1982). Unit costs of resources were based on Durham Veterans Administration medical cost data. We did not include extra costs for physician follow-up (gastroenterology clinic) for dialysis patients who became surface antigen carriers because dialysis patients already receive intense physician care and they were not perceived to require additional clinic visits. The direct cost of an inpatient day (excluding blood products) was based on average expenditures at the Durham Veterans Administration Medical Center for each bed service (e.g., medicine, medical intensive care). Blood products were calculated separately only for patients with fulminant HBV infection. The costs of blood products, medications, and vaccination were based on the prices paid by the medical center for the products. These prices reflect the actual cost of the products to the hospital. The cost of HBV vaccination (three standard doses at zero, one, and six months) was $114. An additional $38 per person was assigned to the predialysis strategy for the booster given at initiation of dialysis. No labor cost was included in the vaccination cost, since predialysis and dialysis patients receive regular physician follow-up and the additional labor cost associated with administering the vaccine is negligible. We used the Veterans Administration's diagnosis-related group value for a liver biopsy, adjusted for a two-day length of stay ($1,029) as the estimate of the cost of a liver biopsy.

The lifetime cost of treating cases of HBV infection contracted by patients and household contacts during the study period was calculated in present-value terms using a discount rate of 5 percent; we assumed that all costs were incurred at the beginning of each year. The cost of treating infected household contacts was calculated over a 30-year time period. The cost of treating HBV infection in dialysis patients was calculated over a ten-year period due to the reduced life expectancy in this patient population (U.S. Renal Data System 1990).

The cost data are conservative in that they include only the medical care costs associated with diagnosing and treating hepatitis B. The indirect costs associated with loss of productivity for dialysis patients or their infected contacts, and the costs of screening and segregating HBV surface antigen-positive patients in the dialysis unit are excluded.

CALCULATION OF EXPECTED COSTS

Expected costs are used in the numerator of the cost-effective ratios. The calculation of expected costs is facilitated by the decision tree. The expected cost of an outcome is defined as the product of the probability of an outcome and the cost of the outcome. The probability of an outcome equals the joint probability of events leading to that outcome. For example, the probability of developing icteric hepatitis that does not require hospitalization and is fully resolved for strategy 3 (no vaccine) is the product of (1) the probability of needing dialysis in three years (.75); (2) the probability of developing HBV infection in a susceptible dialysis patient over three years (.006); (3) the probability that a dialysis patient will develop icteric hepatitis once HBV infection has been contracted (.06); (4) the probability of not being hospitalized if icteric hepatitis developed (.75); and (5) the probability of recovering fully (.22). The resulting combined probability is .0000447 (.0045 percent). The cost of a case of icteric hepatitis not requiring hospitalization is $615, resulting in an expected cost of $0.03 (.000045 x $615) for that outcome. The expected cost of a strategy equals the sum of the expected costs of all outcomes in the strategy.

The decision tree was evaluated in an iterative fashion, beginning with calculation of the expected carrier cost (the cost of diagnosing and treating HBV infection per household contact). Carrier cost was then used as a cost value in evaluating the entire tree. It was assumed that household contacts have a 50 percent chance of contracting HBV infection in three years (a midrange figure based on three studies) (Szmuness, Prince, Grady, et al. 1974; Irwin, Allen, Dancroft, et al. 1974; Wright 1975), and that infected household contacts would not infect anyone outside the household. Based on an average household size of four, there are three household contacts per patient. Therefore, the expected cost of diagnosing and treating HBV infection for one household member was tripled to account for the three family members at risk for infection. The complete decision tree was evaluated using this estimate of total expected carrier cost.

COST-EFFECTIVENESS ANALYSIS

For each strategy, marginal cost-effectiveness ratios were constructed. The numerator was defined as the additional expected cost of the strategy compared to the baseline strategy of no vaccination. The effectiveness measure was defined as the additional expected cases of HBV infection prevented among patients and household contacts per predialysis patient immunized.

SENSITIVITY ANALYSIS

In sensitivity analysis, selected variables from the decision analysis model are assigned a range of possible values and the cost-effectiveness ratios are recalculated to determine whether the results are influenced by (or sensitive to) the range of values for any selected variable. If the basic results do not change, the analysis is not influenced by (or is insensitive to) changes in the selected variable. If, however, the basic results of the cost-effectiveness ratios change when the values of a selected variable are varied, then the analysis is said to be sensitive to that variable. We used sensitivity analysis to assess which vaccine efficacy or cost variables, HBV infection outcome variables, and HBV infection rates might play a pivotal role in the cost effectiveness of the vaccination strategies. We varied the estimates of response to the predialysis vaccine, total medical care costs (excluding vaccine cost), vaccine cost, HBV infection rates, and probability of progression to dialysis in three years over plausible ranges to assess their impact on the cost-effectiveness ratios. All sensitivity and threshold analyses were performed with the aid of SMLTREE |R~ software.

RESULTS

PROBABILITY OF DEVELOPING HEPATITIS

The probability of developing HBV infection for each vaccination strategy over three years is shown in Table 3. Compared with the no-vaccination strategy, the predialysis vaccine strategy reduced the probability of acquiring HBV infection by about 70 percent. Expressed in terms of the number of patients who must be immunized to prevent a single case of HBV infection, the reciprocal of the absolute risk reduction (Laupacis, Sackett, and Roberts 1988), 434 vaccinees are required per case of HBV infection prevented for the predialysis vaccine strategy. The dialysis vaccine strategy would require 625 vaccinees to prevent a single case of HBV infection. The lower efficacy of the vaccine in the dialysis group is reflected by the higher number of vaccinees required to prevent a case of hepatitis B.

MEDICAL CARE COSTS

The expected medical care costs for each vaccination strategy are presented in Table 4. Both vaccine strategies are more costly than the no-vaccination strategy, indicating that the costs saved from the hepatitis B cases prevented do not outweigh the cost of the vaccine. The expected cost of the outcomes differs among strategies because of differences both in the probability of contracting HBV infection and in vaccination costs.

The most significant difference among strategies is the difference in the cost attributable to patients who do not contract HBV infection TABULAR DATA OMITTED (including those who do not progress to dialysis): this cost accounts for 0 percent of expected cost per patient in the no-vaccination strategy, 98 percent of expected cost per patient in the strategy that vaccinates at dialysis, and 99 percent of expected cost per patient in the strategy that vaccinates predialysis patients. Excluding the "no hepatitis" category, two-thirds of the cost accounted for by the hepatitis disease states in each strategy is attributable to the cost of infected dialysis patients who become carriers of HBV and infect their household contacts. Although these patients are already undergoing intensive medical surveillance and use little additional health care resources themselves, they present a risk of infection to their household contacts. The cost of household contacts infected by hepatitis patients decreases progressively from the no-vaccination strategy to the predialysis vaccine strategy. This occurs because each vaccination strategy is effective in reducing the cases of HBV infection in dialysis patients thus reducing the risk to their contacts.
Table 4: Expected Medical Care Costs ($) per Patient for Each
Vaccination Strategy, Using Low-Incidence Estimates(*)
Disease Vaccine at Vaccine before
Outcome No Vaccine Dialysis Dialysis
No hepatitis 0.0 82.47 141.39
Carrier state 1.85 1.09 .70
Other (chronic, 0.96 0.60 0.36
death, recovery)
Total 2.81 84.16 142.45
* Expected value of lifetime treatment costs for all cases of
hepatitis B contracted within the three-year time period
(discount rate = 5 percent). Life expectancies of patients and
household contacts were assumed to be 10 and 30 years,
respectively. Incidence estimates were those of Alter, Favero,
Miller, et al. (1989).


MARGINAL COST AND EFFECTIVENESS OF THE VACCINE STRATEGIES

The marginal cost-effectiveness ratios are presented in Table 5. The marginal ratios indicate that the additional expenditures required to prevent additional cases of HBV infection are considerable. If patients are vaccinated prior to dialysis, the expected cost per additional patient or household contact case of HBV infection prevented is $31,111. This means that an additional $31,111 would be spent on vaccinating and treating predialysis patients for HBV infection to prevent a single case of hepatitis B in those patients who require dialysis within three years. The TABULAR DATA OMITTED expected cost of a dialysis vaccination strategy would be $25,313 per case of hepatitis B prevented. The dialysis vaccination strategy costs slightly less even though it is less efficacious because vaccine is not wasted on patients who are not on dialysis.

To facilitate a comparison of the cost effectiveness of the predialysis and dialysis vaccines estimated in this study with other health care interventions analyzed in the literature, we have calculated the cost of the predialysis vaccine per life saved and per year of life saved over the three-year time period in the analysis. The calculations suggest that the predialysis vaccine strategy costs $4.4 million per life saved, or $583,333 per year of life saved. This compares to a cost of $3.7 million per life saved, or $493,902 per year of life saved for the dialysis vaccine strategy.

We repeated the analysis using the assumption that the probability of contracting HBV infection was 2 percent over three years if successfully vaccinated and 13 percent if unvaccinated; these were the actual observed rates from Stevens' multicenter trial (Stevens, Alter, Taylor, et al. 1984). The expected cost of preventing one case of HBV was $856 ($18,000 per year of life saved) for the predialysis strategy. The expected cost of the dialysis strategy was $679 per case prevented ($14,000 per year of life saved).

SENSITIVITY ANALYSIS

The results of three sensitivity analyses comparing the predialysis and no-vaccination strategies are shown in Figure 3. In each panel, the y axis represents the additional medical cost per additional case prevented (relative to the baseline strategy of no vaccine). The first analysis (Panel A) demonstrates that even when all health care costs from HBV infection are increased tenfold the additional cost per additional case prevented is still higher for the predialysis vaccination strategy than the no-vaccination strategy. Given the assumed probability of progressing to dialysis of .75, and a vaccine response rate of 80 percent in predialysis patients, total health care costs must increase 74-fold before the predialysis strategy is preferable to the no-vaccination strategy.

The second analysis (Panel B) shows the sensitivity of the cost-effectiveness ratio to the actual cost of the vaccine. When the cost of the vaccine is decreased tenfold from $114 to $12, the predialysis vaccination strategy is still more expensive than the no-vaccination strategy. Even if the response rate were to increase to 100 percent, the cost of the vaccine would have to be decreased to $1.54 before predialysis strategy would be cost-saving (assuming a response rate of 80 percent). The third analysis (Panel C) shows the insensitivity of the cost effectiveness of the predialysis strategy to the probability of needing dialysis in three years. Even when 100 percent of predialysis patients go on dialysis within three years, implying that no vaccine is "wasted" on patients who will not be at risk of infection, the vaccination strategy is still expensive relative to the no-vaccination strategy when the cost of vaccination is high. The model is also insensitive to variation in carrier cost. The carrier cost must increase more than 140 times if the predialysis and no-vaccination strategies are to be equivalent in terms of expected cost.

The sensitivity of estimated costs per year of life saved to variation in HBV infection rates is shown for the predialysis strategy in Figure 4. The bounds of the analysis were the Alter, Favero, Miller, et al. (1987) data (0.2 percent per year) and the Stevens, Alter, Taylor, et al. (1984) data (4.5 percent per year). The x axis represents the annual HBV infection rates (without vaccination) among patients with renal insufficiency. The y axis represents the additional medical cost per year of life saved. This analysis shows the marked sensitivity of the cost effectiveness of the predialysis strategy to the HBV infection rate. The cost per year of life saved increased dramatically below annual incidence rates of 1 percent. A parallel analysis for the dialysis strategy showed similar results.

DISCUSSION

HBV vaccination has been proposed as a method of further reducing the burden of HBV infection in the dialysis unit below the level obtained by infection control guidelines alone. Previous research indicates that the rate of response to the HBV vaccine can be improved by immunizing patients with moderate renal insufficiency prior to dialysis (Seaworth, Drucker, Starling, et al. 1988). In our analysis, the predialysis vaccine strategy reduces the incidence of HBV infection by 70 percent, although the absolute reduction is small, |is less than~ 1 percent using Alter's current low-incidence estimates (Alter, Favero, Miller, et al. 1987).

We also calculated the direct expected medical care costs for each vaccination strategy. The expected costs per person when patients were immunized prior to dialysis was $142 compared to $3 for patients who were not immunized. This indicates that the vaccine costs far outweigh the costs saved from cases of HBV infection prevented. The additional cost of the predialysis vaccine is reflected in the marginal cost-effectiveness ratios. The marginal ratios indicate the additional expenditures required to prevent the occurrence of additional HBV infection cases. For patients immunized prior to dialysis, the expected cost of preventing an additional patient or household contact from contracting hepatitis B was $31,111. This additional cost translates to approximately $583,333 per year of life saved. These cost estimates are dramatically reduced when the rate of HBV infection is assumed to be higher (13 percent for unvaccinated patients). The expected cost of preventing a case of HBV, assuming higher rates of infection, was $856 for the predialysis strategy, which translates into $18,000 per year of life saved.

We have conducted this analysis from the public health perspective, including health care costs regardless of payment source. We did not include costs for lost productivity of dialysis patients who develop hepatitis B. Dialysis patients are usually not employed because of their dialysis (Evans, Manninen, Garrison, et al. 1985), and any estimates of lost productivity over and above losses already occurring because of dialysis would be low.

We used sensitivity analysis to identify and assess those vaccine efficacy or cost variables, and those hepatitis outcome variables, that might play a pivotal role in the cost effectiveness of the vaccination strategies. The most sensitive variable in the analysis was the actual vaccine cost. At current vaccine costs, our analysis suggested that the predialysis vaccination strategy was still far more expensive than the no-vaccination strategy, even when the vaccine response rate was 100 percent. The vaccine cost would have to be reduced from $114 to $1.54 for the predialysis vaccination strategy to become cost effective. Even given the high assumption regarding the incidence of HBV infection, the comparable threshold for achieving cost effectiveness of the predialysis strategy was $33 for the cost of the vaccine.

Our analysis did show significant sensitivity of the cost-effectiveness ratios to the HBV infection rate. When HBV infection rates as low as those reported by Alter, Favero, Miller, et al. (1987) were used in the analysis, neither vaccination strategy was cost effective. If, however, infection rates were as high as those observed in the Stevens, Alter, Taylor, et al. (1984) study, then both vaccination strategies would be significantly more cost effective. This suggests that either vaccine strategy would be more cost effective (in the range of $50,000 per year of life saved) in dialysis centers that experience higher rates of HBV infection (about 2 percent annual incidence), and that local incidence rates should be considered when determining the appropriate vaccination strategy.

In our analysis, we used estimates of VA health care costs rather than costs of care in the private sector. However, even when calculated health care costs were increased tenfold, the cost per case of HBV infection prevented was still higher for both vaccination strategies. Total health care costs would have to be 74-fold times greater before the predialysis vaccination strategy would be preferred to the no-vaccination strategy.

Our analysis was based on the published response rate among dialysis patients to the plasma-derived HBV vaccine available in the United States (Heptavax B -- Merck, Sharp & Dohme) (Stevens, Alter, Taylor, et al. 1984; de Graeff, Dankert, de Zeeuw, et al. 1985; Bommer, Gussendorf, Jilg, et al. 1984). We used an estimated response rate of 55 percent that was obtained from a standard vaccination schedule at zero, one, and six months for dialysis patients. The actual maximum response rate in dialysis patients immunized according to this schedule is 57 percent. Higher response rates have been reported by adding a booster dose at 12 months (72 percent) (de Graeff, Dankert, de Zeeuw, et al. 1985). Two different heat-inactivated vaccines are available in Europe (HEVAC-B, Institut Pasteur Production and CLB, Central Laboratory of the Netherlands Red Cross Blood Transfusion Service). Each of these vaccines yields a response in dialysis patients similar to that obtained from HeptaVax B when given in comparable doses within comparable schedules: 45 percent at 12 months for HEVAC-B and 53 percent at 12 months for CLB (de Graeff, Dankert, de Zeeuw, et al. 1985; Bommer, Gussendorf, Jilg, et al. 1984). Again, higher response rates (82 percent for HEVAC-B and 91 percent for CLB) can be achieved by increasing the dose and frequency of vaccination (Benhamou, Courouche, Jungers, et al. 1984). Even if vaccine efficacy were improved, the significant increased cost of doubling or tripling the vaccine dose would markedly reduce the cost effectiveness of vaccine strategies. Also, the newer vaccines available in the United States have not been marketed at a lower cost.

The cost of an HBV vaccination program can be compared with other medical interventions. We have estimated that, given average reported incidence rates, the predialysis vaccine costs $583,333 per year of life saved, and that the dialysis vaccine strategy costs $493,902 per year of life saved. The cost per year of life saved may be as low as $18,000 for the predialysis strategy and $14,000 for the dialysis strategy if incidence rates are at the high end of the plausible range. These figures compare to a cost of approximately $8,000 per year of life saved for the pneumococcal vaccine in the elderly population (all figures in 1987 dollars) (Sisk and Riegelman 1986; Willems et al. 1980). In contrast, the influenza vaccine is cost saving when administered to people over 65 years of age (Riddiough, Sisk, and Bell 1983). Neither pneumococcal nor influenza vaccines are more efficacious than HBV vaccine in terms of preventing their targeted diseases, but both vaccines are less expensive, costing less than $10 per dose per year, compared to an annual cost of $51 per year for the predialysis HBV vaccine strategy (three initial doses and one booster for three years). Comparisons such as these must be interpreted cautiously due to varying assumptions underlying the calculations. For example, the other vaccine calculations use quality-adjusted life years as the unit of effectiveness, whereas unadjusted life years saved were used in the present analysis. Assuming some morbidity exists among patients during extended years of life, the cost-effectiveness ratios in the present analysis are underestimated relative to those based on quality-adjusted life years. However, these comparisons help place the cost-effectiveness analysis of HBV vaccination into some context with other health care expenditures.

The results of our analysis were insensitive to changes in most variables included in the model. The cost effectiveness of the predialysis HBV vaccination strategy would improve if several factors originally excluded from the model had been incorporated. First, we assumed that surveillance costs incurred under the no-vaccination strategy could not be reduced following the introduction of the dialysis or predialysis vaccination strategies. While this assumption may be reasonable, the vaccination strategies would become less expensive relative to the no-vaccination strategy if such surveillance costs could be lowered. Second, we assumed that the clinical outcome of HBV infection once contracted by dialysis patients was not altered if a patient had received the vaccine previously. If the consequences of HBV infection were less severe for patients who had received the vaccine, the health care costs associated with HBV infection would be reduced for both vaccination strategies. Third, the indirect costs of morbidity and mortality were excluded from the analysis. The incorporation of these costs, which include the production losses and psychological costs associated with illness and death, would only slightly improve the cost effectiveness of the vaccination strategies. Finally, the effectiveness of the vaccine strategies was underestimated somewhat by limiting the time frame of the analysis to three years. However, even when all patients were assumed to require dialysis in the first year, the vaccination strategies were still considerably more expensive per case prevented than the no-vaccination strategy. Furthermore, given uncertainty regarding the long-term duration and level of immunity provided by lifetime treatment in a vaccination strategy, it was not feasible to extend the analysis to cover the life of dialysis patients.

The predialysis HBV vaccine strategy is the most clinically effective immunization strategy; it is also the most expensive one. Assuming average incidence rates, the predialysis strategy costs $31,111 per case prevented, or $583,000 per year of life saved. The most important factors that adversely influence the cost-effectiveness ratios of the two vaccination strategies is the current high cost of the hepatitis B vaccine and the low prevalence rate of HBV infection. At current costs for HBV vaccine and current average rates of HBV infection, a predialysis vaccination strategy is very expensive. Furthermore, when compared with other immunization programs, a dialysis vaccination strategy is not cost effective. However, the cost effectiveness of both strategies improves significantly as the incidence rises, and may be cost effective in dialysis centers with higher infection rates. In centers with very low infection rates, the cost of the vaccine would have to be drastically lowered for HBV vaccination programs targeting dialysis or predialysis patients to be comparable to other vaccination programs.

APPENDIX

TABULAR DATA OMITTED

TABULAR DATA OMITTED

TABULAR DATA OMITTED

ACKNOWLEDGMENTS

The authors wish to thank Joe Lipscomb, Ph.D. for the review of an earlier draft, Michael Monger for computer assistance, and Sheila Evans and Deborah Chapman for secretarial support.

REFERENCES

Alter, H. J., M. S. Favero, and J. E. Maynard. "Impact of Infection Control Strategies on the Incidence of Dialysis-Associated Hepatitis in the United States." Journal of Infectious Diseases 153 (1986): 1149-51.

Alter, M. J., M. S. Favero, J. K. Miller, L. A. Moyer, and L. A. Bland. "National Surveillance of Dialysis Associated Disease in the U.S., 1987." Transactions of the American Society of Artificial Internal Organs 35, no. 7 (1989): 8202-31.

Benhamou, E., A. M. Courouche, P. Jungers, A. Laplanche, F. Degos, J. Brangier, and J. Crosnier. "Hepatitis B Vaccine: Randomized Trial of Immunogenicity in Hemodialysis Patients." Clinical Nephrology 21, no. 3 (1984): 143-47.

Bommer, J., E. Ritz, K. Andrassy, G. Boomer, F. Deinhardt, W. Jilg, and G. Darai. "Effect of Vaccination Schedule and Dialysis on Hepatitis B Vaccination Response in Uraemic Patients." Proceeding of the European Dialysis and Transplant Association 20 (1983): 161-68.

Bommer, J., M. Gussendorf, W. Jilg, F. Deinhardt, H. G. Koch, G. Darai, G. Boomer, M. Rambausek, and E. Ritz. "Vaccination against Hepatitis B in Patients with Renal Insufficiency." Proceeding of the European Dialysis Transplant Association 21 (1984): 300-305.

Centers for Disease Control. Control Measures for Hepatitis B in Dialysis Centers. Viral Hepatitis Investigation and Control Services, Atlanta, GA, November 1977.

Crosnier, J., P. Jungers, A. M. Courouche, A. Laplanche, I. Benhamou, F. Degos, B. Lacour, P. Prunet, Y. Cerisier, and P. Guesry. "Randomized Placebo-Controlled Trial of Hepatitis B Surface Antigen Vaccine in French Haemodialysis Units: Part I. Medical Staff." Lancet 1 (28 February 1981a): 455-59.

Crosnier, J., P. Jungers, A. M. Courouche, A. Laplanche, D. Benhamou, F. Degos, B. Lacour, P. Prunet, Y. Cerisier, and P. Guesry. "Randomized Placebo-Controlled Trial of Hepatitis B Surface Antigen Vaccine in French Haemodialysis Units: Part II. Haemodialysis Patients." Lancet 1 (11 April 1981b): 797-800.

de Graeff, P. A., J. Dankert, D. de Zeeuw, C. H. Gips, and G. K. van der Hem. "Immune Response to Two Different Hepatitis B Vaccines in Haemodialysis Patients: A 2 Year Followup." Nephron 40 (1985): 155-60.

Desmyter, J., G. De Groote, J. Colaert, M. Reynders, E. E. Reerink-Brongers, P. N. Lelie, P. J. Dees, H. W. Reesnick, and the Leuven Renal Transplantation Colaborative Group. "Efficacy of Heat-Inactivated Hepatitis B Vaccine in Haemodialysis Patients and Staff." Lancet 2 (10 December 1983): 1323-28.

Evans, R. W., D. L. Manninen, L. P. Garrison, L. G. Hart, C. R. Blagg, R. A. Gutman, A. R. Hull, and E. G. Lowrie. "The Quality of Life of Patients with End-Stage Renal Disease." New England Journal of Medicine 312, no. 9 (1985): 553-59.

Irwin, G. R., A. M. Allen, W. H. Dancroft, et al. "Hepatitis B Antigen and Antibody. Occurance in Families of Asymptomatic HBAg Carriers." Journal of the American Medical Association 227, no. 9 (1974): 1042-45.

King, J. W. "A Clinical Approach to Hepatitis B." Archives of Internal Medicine 142 (May 1982): 925-28.

Kohler, H., W. Arnold, G. Renschin, H. H. Dormeyer, and K. H. Buschenfelde. "Active Hepatitis B Vaccination of Dialysis Patients and Staff." Kidney International 25 (1984): 124-28.

Laupacis, A., D. L. Sackett, and R. S. Roberts. "An Assessment of Clinically Useful Measures of the Consequences of Treatment." New England Journal of Medicine 318, no. 6 (1988): 1728-33.

London, W. J., J. S. Drew, E. D. Lustbader, B. G. Werner, and B. S. Blumberg. "Host Response to Hepatitis B Infection in Patients in a Chronic Hemodialysis Unit." Kidney International 12 (1977): 51-58.

Mitch, W. E., M. Walser, G. A. Burlington, and J. Lemann. "A Simple Method of Estimating Progression of Chronic Renal Failure." Lancet 2 (18 December 1976): 1326-30.

Mulley, A. G., M. D. Silverstein, and J. L. Dienstag. "Indications for Use of Hepatitis B Vaccine, Based on Cost-Effectiveness Analysis." New England Journal of Medicine 307 (9 September 1982): 644-52.

Parfrey, P. S., D. Farge, C. Forbes, R. Dandavino, S. Kenick, and R. D. Guttmann. "Chronic Hepatitis in End Stage Renal Disease: Comparison of HBsAg Negative and HBsAg Positive Patients." Kidney International 28 (1985): 959-67.

Riddiough, M. A., J. E. Sisk, and J. C. Bell. "Influenza Vaccination: Cost-Effectiveness and Public Policy." Journal of the American Medical Association 249, no. 23 (1983): 3189-95.

Rutherford, W. E., J. Blondin, J. P. Miller, A. S. Greenwalt, and J. D. Vavra. "Chronic Progressive Renal Disease: Rate of Change of Serum Creatinine Concentration." Kidney International 11 (1977): 62-70.

Seaworth, B., J. Drucker, J. Starling, S. Drucker, C. Stevens, and J. Hamilton. "Hepatitis B Vaccines in Patients with Chronic Renal Failure before Dialysis." Journal of Infectious Diseases 157, no. 2 (1988): 332-37.

Shustermann, N., and I. Singer. "Infectious Hepatitis in Dialysis Patients." American Journal of Kidney Disease 9, no. 6 (1987): 447-55.

Sisk, J. E., and R. K. Riegelman. "Cost Effectiveness of Vaccination against Pneumococcal Pneumonia: An Update." Annals of Internal Medicine 104, no. 1 (1986): 79-86.

Snydman, D. R., D. Bregman, and J. A. Bryan. "Hemodialysis-associated Hepatitis in the U.S., 1974." Journal of Infectious Diseases 135, no. 4 (1977): 687-91.

Stevens, C. E., H. J. Alter, P. Taylor, E. A. Zang, E. F. Harley, W. Szmuness, and the Dialysis Vaccine Trial Study Group. "Hepatitis B Vaccine in Patients Receiving Hemodialysis." New England Journal of Medicine 311, no. 8 (1984): 496-501.

Szmuness, W., A. Prince, G. Grady, M. K. Mann, R. W. Sevine, E. A. Friedman, M. J. Jacobs, A. Hosephson, S. Ribot, F. L. Shapior, K. H. Stenzeh, W. N. Suki, and G. Vyas. "Hepatitis B Infection: A Point-Prevalence Study in 15 U.S. Dialysis Centers." Journal of the American Medical Association 227, no. 8 (1974): 901-906.

Szmuness, W., C. E. Stevens, E. J. Harley, E. A. Zang, W. R. Olesko, D. C. William, R. Sadovsky, J. M. Morrison, and A. Kellner. "Hepatitis B Vaccine Demonstration of Efficacy in a Controlled Clinical Trial in a High Risk Population in the U.S." New England Journal of Medicine 303, no. 15 (1980): 833-41.

U.S. Department of Health and Human Services. "Protection against Viral Hepatitis." Morbidity and Mortality Weekly Report 39 (1990): 1-25.

U.S. Renal Data System. USRDS 1990 Annual Data Report. Bethesda, MD: National Institutes of Health, August 1990.

Willems, J. S., C. R. Sanders, M. A. Riddiough, and J. C. Bell. "Cost Effectiveness of Vaccination against Pneumococcal Pneumonia." New England Journal of Medicine 303, no. 10 (1980): 553-59.

Wright, R. A. "Hepatitis B and the HBsAg Carrier. An Outbreak Related to Sexual Contact." Journal of the American Medical Association 232, no. 6 (1975): 717-21.

Eugene Z. Oddone, M.D. is Assistant Professor in the Division of General Internal Medicine, Duke University Medical Center, Durham, NC; and the Health Services Research and Development (HSR&D) Field Program, Durham VA Medical Center. Patricia A. Cowper, Ph.D. is Research Health Science Specialist in the HSR&D Field Program, Durham VA Medical Center and the Research Service, Durham VA Medical Center. John D. Hamilton, M.D. is Professor in Medicine in the Division of Infectious Disease, Duke University Medical Center; and the Research Service, Durham VA Medical Center. John R. Feussner, M.D. is Chief of the Division of General Internal Medicine, Duke University Medical Center, the HSR&D Field Program, Durham VA Medical Center; and the Research Service, Durham VA Medical Center. Address correspondence and requests for reprints to Eugene Z. Oddone, M.D., Center for Health Services Research in Primary Care, Durham VA Medical Center, Durham, NC 27705. The article, submitted to Health Services Research on March 21, 1991, was revised and accepted for publication on March 27, 1992.
COPYRIGHT 1993 Health Research and Educational Trust
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Oddone, Eugene Z.; Cowper, Patricia A.; Hamilton, John D.; Feussner, John R.
Publication:Health Services Research
Date:Apr 1, 1993
Words:6642
Previous Article:Ecological analysis of the first generation of community clinical oncology programs.
Next Article:Measuring quality of care in psychiatric emergencies: construction and evaluation of a Bayesian index.
Topics:

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters