Recent licensure of the live, attenuated varicella virus vaccine (VARIVAX®; Merck and Company, Inc., Whitehouse Station, New Jersey), coupled with recommendations of the Advisory Committee on Immunization Practices,1 poses questions with respect to cost-effectiveness among selected vaccination target groups. In an effort to help prioritize varicella vaccination initiatives within employee health programs at Department of Veterans Affairs (VA) medical centers, we carefully evaluated the Halloran-Lieu combined model of decision analysis and transmission dynamics2-6, the analytical framework of which can extend to both pediatric and adult outcomes assessment.
Given both the substantial resources and considerable time that are required to generate such a de novo decision analysis model, it was our intent to first determine if and how an existing model could meet our own needs to furnish employee health program guidelines within the VA. On balance, the Halloran-Lieu model provides highly valuable insights that can apply to the vaccination of health care workers who serve an adult patient population. However, we also wish to underscore those model assumptions that do, in fact, blur the transition from pediatric to employee health vaccination approaches.
The Halloran-Lieu model depicts different strategies, and subsequent outcomes, for individuals from different age cohorts (2 to 5, 6 to 11, 12 to 17, and 18 to 35 years, respectively, with data from the last cohort unpublished). Only those individuals who have a negative or uncertain history of varicella are included in the actual model, because those with a positive history have presumed immunity. Thus, from an employee health program perspective, there is some initial underestimation of costs because additional employee visits, surveys, telephone contacts, etc, are necessary unless such information is already present in the employee health database of a facility.
The Halloran-Lieu model strategy that is most transferable to employee health is that of testing all employees who have either a negative or uncertain history for the presence of varicella antibody, and only vaccinating those individuals who demonstrate laboratory evidence of susceptibility. The value of a "no intervention" strategy is to serve as a baseline against which other, incremental cost-effectiveness ratios are calculated. The "vaccinate all" strategy, in the absence of prior antibody testing, will not be considered further because the model showed it to be far more costly than the more selective "test and/or vaccinate" algorithm.
Outcomes in the model included the following: (1) short-term medical costs per case of varicella prevented, which only include direct medical costs associated with antibody determination and vaccine administration, and (2) total costs, which are the summation of both short-term and long-term costs per case prevented. Long-term costs include those costs resulting from any varicella-related inpatient/outpatient medical care (including possible sequelae), as well as projected lost workday costs. Consequently, this methodology reflects a societal perspective rather than simply the institutional perspective, which is associated with short-term costs only.
Figure 1has extracted data (unpublished) from the model's Quattro Pro (Corel, Ottawa, Ontario, Canada) spreadsheet program, comparing "test and/or vaccinate" cost-effectiveness ratios by age cohort, which reflect incremental costs per case prevented relative to the "no intervention" strategy. As the data indicate, there are relatively greater direct medical costs for each cohort, which are offset by averted medical/work-loss costs associated with the actual prevention of varicella cases. In fact, vaccination between the ages of 6 and 11 achieves a net savings compared with no intervention, with respect to total cost per case prevented. As expected, vaccination costs tend to increase with age because of progressively acquired immunity, even among those individuals who cannot either recall or document previous infection.
Furthermore, it must be emphasized that the ratios in Figure 1 are based on the utilization of baseline cost and probability values only, without outcomes that reflect sensitivity analyses (at upper and lower bound values). Although such sensitivity analyses have been conducted,5 it is beyond the scope of this letter to incorporate such modeling permutations into the current dialogue. Rather, suffice it to say that the costs presented in Figure 1 should be treated as trends/approximate ranges rather than static, fixed markers of cost-effectiveness.
Now that there is a frame of reference for assessing cost-effectiveness using this decision analysis model, what can we apply to the employee health setting? First, an 18- to 35-year-old cohort provides a reasonable target group because the majority of health care workers are in this age range as their careers commence. Hence, vaccinating the young health care worker will have expected dividends in both protecting the employee as well as curtailing secondary transmission to susceptible patients. Because the transmission dynamics of the Halloran-Lieu model assume that everyone within an age group has exactly the same exposure to infection and exposes people in exactly the same manner, VA health care workers would presumably fall under the model as originally developed. However, the implications of applying this model to health care workers who treat children are more perplexing, because it does not incorporate assumptions about the interaction of subgroups that have different susceptibility and exposure profiles.
Second, one must consider how cost assumptions should be structured in the employee health setting, in contrast to a situation in which the great majority of exposures to varicella occur in the community. Staff exposures to varicella, which are more likely in a pediatric than in a VA setting, may require costly furloughs to the institution, in the absence of a vaccination program. As a result, the "no intervention" strategy should be associated with higher cost-effectiveness ratios, thus making vaccination relatively less expensive. Nevertheless, an offsetting factor to this presumed savings may be the risk of varicella transmission associated with live, attenuated vaccine administration. Although the Centers for Disease Control and Prevention cannot document specific clinical risks of person-to-person transmission,1 employee health programs may elect to furlough those staff who present with a post-vaccination vesicular rash (the approximate prevalence of which equals 5% among adults).
Third, the Halloran-Lieu model cost assumptions may not account for the extra administrative costs that are borne by a facility-based employee health program. Whereas community-based varicella immunizations can most often be scheduled within routine pediatric office visits at a negligible marginal cost, an employee health program may well require additional staff resources to screen for vaccine allergies, obtain consent, and assist with other documentation.
In summary, however, the Halloran-Lieu model provides a useful bench-mark from which occupational medicine specialists can prioritize their own program needs. Although outcomes assessment is not yet a well-charted tool for the comparison of different health interventions, published data on the cost-effectiveness of hepatitis B vaccination provides a pertinent point of reference. Protection against hepatitis B is already considered an established standard of care in occupational medicine, and cost-effectiveness ratios published by Bloom et al7 specify total costs per case prevented for adult vaccinees. For a general adult population with a baseline case probability of 1% infection over a ten-year period, total costs per case prevented were computed to be in the $12,000 to $16,000 range, depending upon discounting methods used. In contrast, for "high-risk" adults (ie, corresponding to ideal health care worker vaccine candidates), with a 50% base case ten-year incidence rate, net savings (not quantified) were realized via the combination of antibody screening and/or vaccination. Although this 50% ten-year incidence rate may not necessarily be an accurate estimate for all designated "high-risk" staff, it still provides a somewhat convenient analytical boundary for the purpose of this discussion.
Consequently, the following policy inference can be made by comparing such cost-effectiveness models (with the caveat that there are differences in how each model has been constructed): An employee health program that invests in varicella vaccination can expect that its cost to the institution (as well as to society) will be substantially greater than protection from hepatitis B, among staff who experience day-to-day potential bloodborne exposures. Conversely, this cost is probably less than vaccinating lower-risk employees for whom hepatitis B protection is, in fact, more of a regulatory imperative than an employee health care-driven option.
We thank Marianne Cloeren, MD, MPH, for her thoughtful comments.
Mitchell I. Burken, MD
VA Maryland Health Care System; Baltimore, MD
Gary A. Roselle, MD
Cincinnati VA Medical Center; Cincinnati, OH
Editor's Note: This article was supported and written by Dr Burken and Dr Roselle while Dr Burken was employed by the Department of Veterans Affairs. No official support or endorsement by the Department of Health and Human Services, Health Care Financing Administration, is intended or should be inferred.
The Greenville Hospital System offers an excellent opportunity for a Board Certified /Eligible Occupational Medicine Physician in our expanding Center for Health and Occupational Services (CHOS). The Center provides a full spectrum of occupational healthcare to include treatment, prevention and rehabilitation services for work injuries/illnesses. No weekends or call required. Affiliated with a state-of-the-art multifacility health system offering a sophisticated network of specialties and technologies. Competitive and progressive salary and benefits package. E.O.E.
Contact Linda Stoddard, Human Resources Greenville Hospital System, 701 Grove Red Greenville, South Carolina 29605 (864) 455-6157
1. Centers for Disease Control and Prevention. Prevention of varicella: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR.
2. Halloran ME, Cochi S, Lieu T, Wharton M, Fehrs LJ. Theoretical epidemiologic and morbidity effects of routine varicella immunization of preschool children with live-virus varicella vaccine in the U.S. Am J Epidemiol.
3. Halloran ME, Struchiner CJ, Watelet L. Epidemiologic effects of vaccines with complex effects in an age-structured population. Math Biosci.
4. Halloran ME, Epidemiologic effects of varicella vaccination. Infect Dis Clin North Am.
5. Lieu TA, Finkler LJ, Sorel ME, Black SB, Shinefield HR. Cost effectiveness of varicella serotesting vs. presumptive vaccination of school-age children and adolescents. Pediatrics.
6. Lieu TA, Cochi SL, Black SB, et al. Cost effectiveness of a routine varicella vaccination program for U.S. children. JAMA.
7. Bloom BS, Hillman AL, Fendrick AM, Schwartz JS. A reappraisal of hepatitis B virus vaccination strategies using cost-effectiveness analysis. Ann Intern Med.
Readers are invited to submit letters for publication in this department. Submit them to: The Editor, Journal of Occupational and Environmental Medicine, PO Box 370, Bryn Mawr, PA 19010. Letters should be typewritten and double spaced and should be designated “For Publication.”