Abstract

Although varicella vaccine is recommended for infants, many physicians and parents have withheld vaccination from infants because of concerns most the vaccine's long-term efficacy. We used a decision-analytic Markov model to examine the effects of decreasing vaccine efficacy on individuals and society. The model incorporated published information on age-specific incidence, morbidity, and bloodshed rates, as well equally data on shifting disease brunt from childhood to adulthood every bit vaccine compliance increases. The effects of 2 vaccination strategies—vaccinating infants at age 12 months and waiting to vaccinate until children are 10 years of age—were compared with the effects of no vaccination. If the efficacy of the vaccine were to decrease by 75%, then 50% compliance with vaccination at age 12 months would save 1800 life-years and 12,800 quality-adjusted life-years annually in the United States. The quality-adjusted life expectancy of individuals vaccinated at age 12 months would be 63 h longer than that of nonvaccinated individuals and would increase to 79 h every bit vaccination compliance increases and the burden of chickenpox shifts to machismo. Varicella vaccination of infants at age 12 months appears to be benign, even if the efficacy of the vaccine declines essentially.

In 1995, a condom [1, 2] and constructive varicella vaccine became available in the United States. Although varicella vaccination is cost-effective [3–10] and has been endorsed by the American Academy of Pediatrics (AAP; Elk Grove Village, IL) [11, 12] and the Centers for Illness Control and Prevention (CDC; Atlanta) [xiii], pediatricians have been boring to embrace universal vaccination [fourteen, fifteen]. As a result, simply 43% of children are currently vaccinated [xvi]. Obstacles to vaccination include (1) the perception that varicella is a mild disease in children, and (2) the concern that the efficacy of the vaccine could potentially wane [12, 14, xv, 17, eighteen] and, ultimately, could lead to an increase in the number of cases of chickenpox in adults, for whom the illness is more severe [19–21]. Proponents of universal vaccination have countered these arguments by emphasizing that studies have suggested that (ane) vaccine efficacy persists for at to the lowest degree 20 years [22], (2) well-nigh hospitalizations and deaths due to varicella occur among children and can be prevented by vaccination [17, 23, 24], and (3) vaccination may also prevent morbidity associated with herpes zoster [24–26].

In whatsoever vaccination program, what is best for the customs may not be what is best for each individual [21]. When considering mandatory vaccination, public health officials try to maximize societal do good [27], whereas parents and wellness care providers are more than concerned about the welfare of individual children. Varicella models published elsewhere have all viewed vaccination in terms of its societal benefit, offer no specific guidance for the individual.

Some pediatricians have proposed, as an alternative to universal vaccination, that vaccination be delayed until children are 10 years of age [nineteen, 20]. This delay would consequence in nearly children acquiring amnesty later on natural infection, with the remainder receiving vaccination before adulthood. To aid individuals and policymakers make up one's mind whether and when to vaccinate, we developed a decision-analytic model that considered possible waning vaccine efficacy. The model considered three vaccination strategies and weighed a postulated future risk of affliction during adulthood confronting the known benefit of the vaccine to children.

Materials and Methods

Conclusion-Analysis Model

Using a standard reckoner program (Decision Maker vii.07; Pratt Medical Group), we constructed a Markov model to examine the effects of three vaccination strategies in a birth accomplice of four million children [28, 29]. The 3 strategies involved either (ane) vaccinating all infants at age 12 months, (ii) delaying vaccination until age 10 years and and so vaccinating but if a child has no history of varicella, or (3) non vaccinating at all. Markov models simulate the natural history of a disease as a cohort progresses through a finite number of "health states" (figure one) [xxx]. Time is represented past yearly cycles during which the current health country of the patients may remain unchanged or may modify on the ground of "transition probabilities," as derived from the literature.

Figure 1

Markov model representing the lifetime of a birth cohort. Circles denote

Markov model representing the lifetime of a birth cohort. Circles denote "health states." Directly arrows denote events resulting in transition to a new health state. Circular arrows bespeak the possibility of maintaining the aforementioned wellness state. The wellness country of all accomplice members is "unvaccinated" (Unvaccinated) at the beginning of life. Depending on the vaccination strategy followed, the health state of some members may be changed to "vaccinated" (Vaccinated) after the first or 10th yearly wheel. Each year, nonvaccinated individuals may contract chickenpox and vaccinated individuals may contract breakthrough varicella. Both of these groups of individuals are and then considered to be "immune" (Immune). Immune individuals tin develop zoster then can be considered immune once more, or, if they experience a permanent complication, they may have unilateral deafness (Deafened), monocular blindness (Blind), or both (not shown). Regardless of an individual's health country, it is possible for an individual to dice directly (Dead) equally a upshot of either chickenpox or zoster or as a outcome of death as it occurs among the full general population.

Figure ane

Markov model representing the lifetime of a birth cohort. Circles denote

Markov model representing the lifetime of a nascency cohort. Circles denote "wellness states." Straight arrows denote events resulting in transition to a new health state. Circular arrows bespeak the possibility of maintaining the same wellness land. The wellness state of all accomplice members is "unvaccinated" (Unvaccinated) at the kickoff of life. Depending on the vaccination strategy followed, the health state of some members may be changed to "vaccinated" (Vaccinated) afterwards the first or 10th yearly wheel. Each yr, nonvaccinated individuals may contract chickenpox and vaccinated individuals may contract breakthrough varicella. Both of these groups of individuals are and then considered to be "immune" (Immune). Immune individuals tin can develop zoster then tin can exist considered immune once again, or, if they experience a permanent complication, they may have unilateral deafness (Deaf), monocular blindness (Bullheaded), or both (not shown). Regardless of an private's health state, it is possible for an private to dice straight (Dead) as a result of either chickenpox or zoster or as a consequence of death as it occurs among the general population.

For this assay, the entire cohort began life as unvaccinated newborns. At age 12 months, the health state of infants who received vaccination was classified as "vaccinated," whereas the health state of those who did not receive vaccination was classified as "unvaccinated." For each year assessed, an age-specific underlying attack rate was multiplied by the number of susceptible individuals in the population, to predict the number of chickenpox cases and the number of major complications and deaths due to chickenpox. Natural bloodshed rates were based on life tabular array statistics.

Considering the varicella vaccine is not 100% efficacious, some vaccinated children experience quantum varicella (BV), which is caused by a wild-type virus simply is normally balmy in severity. The health state of individuals who develop either chickenpox or BV is reclassified as "immune." For individuals who die of any cause, the wellness state classification is "expressionless." Both immune and vaccinated groups of individuals are at risk for developing zoster and its complications. Zoster may outcome in death or permanent disability.

The simulation concluded when all cohort members had died. Totaling the amount of time that the health land of individuals was classified equally "vaccinated," "unvaccinated," or "immune" (but not "dead") yielded the life expectancy of the cohort. Subtracting a length of fourth dimension proportional to the corporeality of time that individuals were affected by short- and long-term morbidity produced the quality-adjusted life expectancy (QALE; appendix A). To determine the QALE for an individual who followed a given vaccination strategy—equally opposed to determining the QALE of the entire accomplice, not all of whom chose the aforementioned strategy—we reran the model with the use of age-specific set on rates that were generated by the cohort model for the level of compliance tested.

Data and Assumptions

Table ane presents the baseline values for the model and the ranges for sensitivity analysis. When data were unknown, we always chose the most conservative estimate, thereby purposefully biasing the results confronting vaccination.

Table 1

Baseline values for the Markov model and ranges for sensitivity analysis used in the assessment of varicella vaccination strategies in a birth cohort of 4 million individuals.

Baseline values for the Markov model and ranges for sensitivity assay used in the assessment of varicella vaccination strategies in a nativity accomplice of 4 1000000 individuals.

Table one

Baseline values for the Markov model and ranges for sensitivity analysis used in the assessment of varicella vaccination strategies in a birth cohort of 4 million individuals.

Baseline values for the Markov model and ranges for sensitivity analysis used in the assessment of varicella vaccination strategies in a birth cohort of 4 meg individuals.

Varicella incidence . We derived prevaccine-era, historic period-specific incidence of varicella from data in the Kentucky Behavioral Gamble Factor Survey Study (BRFSS) [31] and the Health Interview Surveys (HIS) [32]. For children aged ⩽15 years, we used age-specific rates from the BRFSS, considering they were reported in 1-year intervals. For children aged µ15 years, we used the most recent HIS rates [32, 33]. To calculate the age-specific annual varicella attack rates, we divided the number of expected cases in a given year past the number of susceptible individuals at the beginning of that year [3].

Vaccination and subsequent induction of herd amnesty may cause the incidence of varicella to be redistributed according to age [34]. With decreased numbers of susceptible children and increased numbers of susceptible adults (equally a result of waning immunity), the varicella attack rate might decrease amidst children and increment among adults. To capture this result of herd immunity, we converted age-specific attack rates into susceptibility-specific attack rates. Furthermore, considering BV probably is less infectious than natural varicella, we decreased the overall attack rate by utilize of the following formula: assail rate = f[n + v · (1-east) · i], where f is the susceptibility-specific assault charge per unit function, due north is the percentage of nonimmune, nonvaccinated individuals in the population, 5 is the pct of vaccinated individuals, eastward is the efficacy of the vaccine, and i is the relative infectivity of BV [6].

Complications due to chickenpox . Major complications associated with chickenpox include encephalitis, pneumonia, and superinfection of the skin [35]. In the United States, earlier introduction of the varicella vaccine, the estimated number of annual hospital admissions of patients with complications due to chickenpox ranged from 3837 [36] to 9300 admissions [37], with almost 1-half of the admissions involving children aged <v years [37]. We used age-specific hospitalization rates from the Commission on Professional and Hospital Activities (CPHA) National Sample File to estimate the likelihood of complications [32]. Lin and Hadler [38, 39] recently reported like hospitalization rates in Connecticut.

Deaths due to chickenpox . We based our age-specific death rates on the most recent CDC data for the period 1985–1994 [35]. These data show an increasing case-fatality rate amidst adults, which will magnify the effect of any vaccine-induced upwardly shift in incidence with age [40].

Incidences of zoster and its complications . Ii population-based studies take examined the age-specific incidences of zoster [41, 42] and complications associated with zoster [42, 43]. We relied on the findings of the more recent studies [41, 43], although all iii studies found similar rates for specific complications as well as increasing complication rates with advancing age. The boilerplate duration of postherpetic neuralgia was based on its observed age-specific natural history [44]. More than contempo studies take reported the median, rather than the mean, duration of pain; in addition, they have non been age-specific and generally take been express to 6 months of follow-up [45–47].

Hospitalizations due to zoster . Recently published information accept shown that rates of hospitalization due to zoster are 4-fold greater than those due to chickenpox [39]. The majority of individuals admitted to the hospital were elderly persons with no underlying medical weather. HIV infection was an underlying condition in xi% of hospitalizations for zoster.

Deaths due to zoster . According to CDC data for 1979–1995, more deaths occur annually as a event of zoster [48] than every bit a result of chickenpox [35]. Deaths due to zoster occur predominantly amidst elderly individuals, whereas deaths due to chickenpox can occur among individuals of all ages.

Vaccine efficacy in the prevention of chickenpox . We defined vaccine efficacy as the reduction in the observed number of cases of chickenpox versus the expected number of cases of chickenpox. We calculated the efficacy of the varicella vaccine, as reported from trials in the Us, for each of the first 7 years after administration [49, 50]. Although i vaccine trial in Nihon reported a twenty-yr follow-upward with an efficacy of 100% after the outset year that vaccine was administered, the study was small and relied on self-reporting of vaccine failure, a reporting method that tends to overestimate efficacy [22]. Other analyses have assumed that efficacy either did not wane [4, 7, 9, 10] or decreased past 15% during a lifetime [three, 6]. For our analysis, we causeless a 75% subtract in efficacy during a lifetime (appendix A) [fifty].

Take chances of complications due to varicella in vaccinated individuals . BV by and large has been less severe than natural varicella [51–56]. A recent case-control written report reported that the varicella vaccine was 97% constructive against moderately astringent and astringent disease [57]. Our estimate of an 89% reduction in the risk of complications, which was based on the relative reduction in the number of lesions, was lower than the take chances reduction estimates of 95%–99% that were used in previous analyses [3, four, half-dozen, 7, 9, 10].

Vaccine efficacy in the prevention of herpes zoster . Studies of children with leukemia and studies of healthy children and adults have demonstrated that vaccination reduces the incidence of herpes zoster by 75%–82% [58–62]. This reduction may stalk from the vaccine strain's reduced propensity to cause zoster; if so, once an individual contracts BV with wild-type varicella, the protective result may not persist. Furthermore, no study has compared the severity of wild-type zoster with that of the vaccine strain. To be bourgeois, nosotros assumed that the hazard and severity of zoster after BV would be identical to those seen after wild-type varicella—that is, the vaccine would have no future protective effect against zoster.

Quality adjustments . We used a validated utility musical instrument, the Index of Well-Beingness, to determine quality of life for individuals with chickenpox, zoster, complications of chickenpox and zoster, or vaccine side effects (appendix B) [63].

Results

Validation of Model

In the absence of vaccination, the model accurately reproduced the age-specific incidence of chickenpox reported by the BRFSS. Furthermore, the model predicted that 7% of 18-year-old individuals would be susceptible, a finding similar to estimates of 6%–8% reported elsewhere [37, 64]. The predicted cumulative number of hospitalizations (7700) and deaths (81) due to varicella was also similar to information reported by the CPHA (9300 hospitalizations) [37], the CDC (71 deaths) [37], and the Halloran model (9900 hospitalizations and 56 deaths) [six].

Population Decision

Life expectancy . Without vaccination, the life expectancy for the population is 74.8 years. Accomplishment of 50% and 97% compliance with vaccination at historic period 12 months increases life expectancy by four h and eight h, respectively, and increases QALE by 28 h and 54 h, respectively, compared with no vaccination. A strategy that involves vaccination of nonimmune children at age 10 years and achieves 97% compliance increases life expectancy by five h and QALE by 16 h, compared with no vaccination. These gains are similar to gains achieved past other well-accustomed vaccination programs, including programs for vaccination against mumps (vii h) [65] and Haemophilus influenzae type B (21 h) [66]. Moreover, when compliance is 97%, vaccination at historic period 12 months results in a savings of 24,650 quality-adjusted life-years annually in the United states.

Chickenpox cases and chickenpox-related hospitalizations and deaths . Table 2 shows the projected number of chickenpox cases and numbers of chickenpox-related hospitalizations and deaths during a lifetime. Vaccination at age 12 months has a small effect on the total number of chickenpox cases, because waning immunity leaves vaccinated individuals susceptible to BV. Even so, because BV is a milder affliction than chickenpox, vaccinated individuals have a higher quality of life during affliction, in addition to fewer hospitalizations and deaths.

Table 2

Numbers of chickenpox and zoster cases and numbers of hospitalizations and deaths due to chickenpox and zoster during the lifetime of a birth cohort of 4 million individuals.

Numbers of chickenpox and zoster cases and numbers of hospitalizations and deaths due to chickenpox and zoster during the lifetime of a birth accomplice of four million individuals.

Table ii

Numbers of chickenpox and zoster cases and numbers of hospitalizations and deaths due to chickenpox and zoster during the lifetime of a birth cohort of 4 million individuals.

Numbers of chickenpox and zoster cases and numbers of hospitalizations and deaths due to chickenpox and zoster during the lifetime of a nativity cohort of 4 1000000 individuals.

Vaccination at age 12 months prevents almost all babyhood deaths due to chickenpox. Although the number of deaths among adults may increase as a result of vaccination, the total number of deaths due to varicella remains lower with vaccination than without. Delaying vaccination until age 10 years results in the fewest total deaths, mainly by preventing death among adults.

Zoster cases and zoster-related complications and deaths . Vaccination at age 12 months exerts a modest consequence on zoster (table 2). Vaccination at age 10 years would accept almost no result on zoster, because µ90% of children would become infected with varicella before receiving vaccination.

Private Determination

Life expectancy . Current estimates of rates of local vaccination coverage range from six% to 52% [12]. For an individual living in an area with 25% coverage, all three vaccination strategies yield virtually identical life expectancies (74.8 years). Compared with non providing vaccination, vaccinating at age x years adds 12 quality-adjusted hours, whereas vaccinating at 12 months adds 61 quality-adjusted hours. Of the increase in QALE, ∼54% comes from preventing chickenpox, 31% from preventing zoster, and 15% from preventing decease.

Sensitivity Assay

Vaccine efficacy . We began with the nearly pessimistic assumption—that vaccine efficacy would subtract by 75% during a lifetime. The assumption of a constant vaccine efficacy of 92% doubles the vaccine-associated gain in QALE to 115 h. The utilise of information from another large vaccine trial in the United states, in which efficacy seemed to increase over time, yielded a quality-adjusted gain of 139 h [49].

Population compliance with vaccination . As more individuals opt for vaccination, two effects might be observed. Showtime, the varicella set on rate amid nonvaccinated children might decrease, as has already been observed elsewhere [67]. Consequently, more than nonvaccinated children would accomplish adulthood without having amnesty. Second, the assail charge per unit amongst nonimmune adults—including those with waning vaccine immunity—might increase. Figure 2 shows that, as vaccination compliance increases from 0% to 97%, the subtract in the number of cases occurring during childhood is much more dramatic than the increase in cases occurring during adulthood. With increasing vaccination compliance in a population, the life expectancy for unvaccinated individuals decreases, because more of these individuals reach adulthood without immunity (figure 3). For vaccinated individuals, all the same, life expectancy remains adequately constant, because the main determinants of their life expectancy—vaccine efficacy and the rate of complications due to BV—are independent of population compliance. The divergence in life expectancy between vaccinated and nonvaccinated individuals ranges from nine h (with 25% compliance) to 22 h (with 97% compliance).

Effigy 2

Annual number of cases of chickenpox, by age, for different levels of population compliance with vaccination at age 12 months. The horizontal and vertical axes show the age of the cohort and the annual number of chickenpox cases, respectively. The 2 curves denote 2 levels of population compliance with vaccination at age 12 months. The solid line denotes 0% compliance and the dotted line 97% compliance. Under the assumption that vaccine efficacy wanes at a rate of 1% annually, as population compliance with vaccination increases, cases of disease shift from childhood to adulthood. At 97% compliance, the incidence of disease occurring during childhood declines, resulting in more-susceptible adults and a higher incidence of disease among adults.

Annual number of cases of chickenpox, by age, for different levels of population compliance with vaccination at age 12 months. The horizontal and vertical axes bear witness the historic period of the cohort and the annual number of chickenpox cases, respectively. The ii curves denote 2 levels of population compliance with vaccination at age 12 months. The solid line denotes 0% compliance and the dotted line 97% compliance. Nether the assumption that vaccine efficacy wanes at a rate of 1% annually, as population compliance with vaccination increases, cases of disease shift from childhood to machismo. At 97% compliance, the incidence of disease occurring during babyhood declines, resulting in more-susceptible adults and a college incidence of disease among adults.

Figure 2

Annual number of cases of chickenpox, by age, for different levels of population compliance with vaccination at age 12 months. The horizontal and vertical axes show the age of the cohort and the annual number of chickenpox cases, respectively. The 2 curves denote 2 levels of population compliance with vaccination at age 12 months. The solid line denotes 0% compliance and the dotted line 97% compliance. Under the assumption that vaccine efficacy wanes at a rate of 1% annually, as population compliance with vaccination increases, cases of disease shift from childhood to adulthood. At 97% compliance, the incidence of disease occurring during childhood declines, resulting in more-susceptible adults and a higher incidence of disease among adults.

Annual number of cases of chickenpox, past historic period, for dissimilar levels of population compliance with vaccination at historic period 12 months. The horizontal and vertical axes bear witness the age of the cohort and the annual number of chickenpox cases, respectively. The 2 curves denote two levels of population compliance with vaccination at age 12 months. The solid line denotes 0% compliance and the dotted line 97% compliance. Nether the supposition that vaccine efficacy wanes at a rate of one% annually, as population compliance with vaccination increases, cases of disease shift from childhood to adulthood. At 97% compliance, the incidence of disease occurring during childhood declines, resulting in more than-susceptible adults and a higher incidence of disease amongst adults.

Figure 3

The effects of increasing population compliance with vaccination on the unadjusted life expectancy of individuals choosing each of the 3 vaccination strategies. The horizontal axis shows the proportion of the population complying with vaccination at age 12 months. The 3 curves denote 3 individuals, each of whom follows a different vaccination strategy. The solid line denotes an individual who chooses no vaccination, the dotted line denotes an individual vaccinated at age 12 months, and the dashed line denotes an individual vaccinated at age 10 years. The vertical axis denotes the gain or loss in life expectancy associated with choosing a particular strategy, compared with the prevaccination life expectancy of the population. Individual life expectancy is a function of both the strategy selected and the proportion of the general population that complies with vaccination. The distance between the curves at a given level of population compliance denotes the benefit of vaccination for an individual in that population. When 25% of the population complies with vaccination, individuals vaccinated at 12 months gain 10 h of life expectancy and nonvaccinated individuals neither gain nor lose hours of life expectancy. At 97% compliance, individuals vaccinated at 12 months gain 9 h, but nonvaccinated individuals lose 18 h, for a total benefit of 27 h.

The effects of increasing population compliance with vaccination on the unadjusted life expectancy of individuals choosing each of the 3 vaccination strategies. The horizontal axis shows the proportion of the population complying with vaccination at age 12 months. The 3 curves announce iii individuals, each of whom follows a different vaccination strategy. The solid line denotes an individual who chooses no vaccination, the dotted line denotes an individual vaccinated at age 12 months, and the dashed line denotes an individual vaccinated at age ten years. The vertical axis denotes the gain or loss in life expectancy associated with choosing a particular strategy, compared with the prevaccination life expectancy of the population. Individual life expectancy is a function of both the strategy selected and the proportion of the general population that complies with vaccination. The distance betwixt the curves at a given level of population compliance denotes the do good of vaccination for an individual in that population. When 25% of the population complies with vaccination, individuals vaccinated at 12 months gain 10 h of life expectancy and nonvaccinated individuals neither gain nor lose hours of life expectancy. At 97% compliance, individuals vaccinated at 12 months proceeds 9 h, merely nonvaccinated individuals lose xviii h, for a total benefit of 27 h.

Figure 3

The effects of increasing population compliance with vaccination on the unadjusted life expectancy of individuals choosing each of the 3 vaccination strategies. The horizontal axis shows the proportion of the population complying with vaccination at age 12 months. The 3 curves denote 3 individuals, each of whom follows a different vaccination strategy. The solid line denotes an individual who chooses no vaccination, the dotted line denotes an individual vaccinated at age 12 months, and the dashed line denotes an individual vaccinated at age 10 years. The vertical axis denotes the gain or loss in life expectancy associated with choosing a particular strategy, compared with the prevaccination life expectancy of the population. Individual life expectancy is a function of both the strategy selected and the proportion of the general population that complies with vaccination. The distance between the curves at a given level of population compliance denotes the benefit of vaccination for an individual in that population. When 25% of the population complies with vaccination, individuals vaccinated at 12 months gain 10 h of life expectancy and nonvaccinated individuals neither gain nor lose hours of life expectancy. At 97% compliance, individuals vaccinated at 12 months gain 9 h, but nonvaccinated individuals lose 18 h, for a total benefit of 27 h.

The effects of increasing population compliance with vaccination on the unadjusted life expectancy of individuals choosing each of the three vaccination strategies. The horizontal axis shows the proportion of the population complying with vaccination at historic period 12 months. The 3 curves denote 3 individuals, each of whom follows a dissimilar vaccination strategy. The solid line denotes an individual who chooses no vaccination, the dotted line denotes an individual vaccinated at historic period 12 months, and the dashed line denotes an individual vaccinated at age x years. The vertical centrality denotes the gain or loss in life expectancy associated with choosing a item strategy, compared with the prevaccination life expectancy of the population. Private life expectancy is a function of both the strategy selected and the proportion of the general population that complies with vaccination. The distance between the curves at a given level of population compliance denotes the benefit of vaccination for an individual in that population. When 25% of the population complies with vaccination, individuals vaccinated at 12 months gain 10 h of life expectancy and nonvaccinated individuals neither gain nor lose hours of life expectancy. At 97% compliance, individuals vaccinated at 12 months gain 9 h, just nonvaccinated individuals lose xviii h, for a total benefit of 27 h.

Risk of varicella complications among vaccinated individuals . To date, only several hundred cases of BV take been reported, and none has had associated complications [53, 67–69]. Given the small sample sizes reported, however, it is not possible to conclude that BV is associated with fewer complications than is chickenpox. From postmarketing surveillance, nosotros know that at to the lowest degree 1 child died of chickenpox 21 months subsequently vaccination [1]. Nosotros tested a range of vaccine efficacies against complications, from 0% (denoting that BV is as prone to complications equally is natural infection) to 100% (denoting that vaccination prevents all complications due to varicella; base case efficacy, 89%). Under all assumptions, vaccination at age 12 months provided the longest QALE.

We also examined the effect on unadjusted life expectancy of the vaccine'due south efficacy against complications. In a location with 25% compliance with vaccination, if the vaccine prevents µ73% of complications, then vaccination at historic period 12 months provides the longest unadjusted life expectancy. Otherwise, vaccination at age ten years offers the longest life expectancy.

Discussion

Despite endorsement past the AAP [11, 12] and the CDC [thirteen], varicella vaccine has not gained widespread credence amongst providers and parents considering of concern about the long-term furnishings of vaccination coupled with the belief that complications due to chickenpox in babyhood are rare. Other analyses made favorable assumptions about vaccine efficacy, assuming lifelong amnesty and almost universal protection against complications [4, half-dozen, vii]. Moreover, those studies expressed uncomplicated morbidity due to chickenpox only in economic terms, specifically equally it affects parental earnings [6]. With the exception of Halloran's model [34], all were economic analyses. Each concluded that vaccination is price-saving only if i includes the indirect cost of parental work loss. Judging by the low compliance rate, this economic argument has not proved persuasive to individual determination-makers. None of the studies expressed outcomes in QALE, nor did they offering any guidance to individuals.

To address the concerns of opponents of varicella vaccine, nosotros tested pessimistic assumptions regarding long-term efficacy, using the model to ask "what if?" What if immunity wanes over fourth dimension? What if BV is more than severe than we conceptualize? If so, the benefits of vaccination to the population may be more modest than originally was thought: although almost all deaths occurring during babyhood however would exist prevented, deaths occurring during adulthood might increase, especially amid unvaccinated individuals, as demonstrated elsewhere [34].

Even so, the results favor vaccination, particularly for the individual, whose gain in life expectancy is not diminished past deaths amidst unvaccinated persons. At worst, vaccination would increase QALE; if vaccine efficacy does not reject or if it rises, the benefits would be much greater. Moreover, without vaccination, there would be hundreds of deaths, thousands of complications, and millions of cases occurring annually. Because near of the varicella-associated morbidity would occur among children <10 years of age [13], a program aimed at vaccinating x-year-one-time children would be of limited do good. Should BV prove to exist more severe than we anticipate, vaccination at age 10 years might offer a minor survival benefit, compared with vaccination at age 12 months; notwithstanding, it would not provide a greater QALE, because the amass morbidity from millions of cases of chickenpox associated with the sometime vaccination strategy would outweigh the excess complications from waning immunity.

How should ane translate these results? Practise a few hours or days really matter? It is of import to realize that these hours represent an boilerplate proceeds for the population and are not only tacked on to the end of each individual'southward life. Near preventative interventions yield similarly pocket-size gains, because we take a large benefit (like avoiding death) that accrues to but a few people and average it across the entire population [70]. Thus, preventing 100 deaths during babyhood in a cohort of 4 million results in an average gain of only 15 h. If the 100 deaths are significant, then the fifteen h that represent those deaths are as pregnant. Accordingly, preventive programs appeal most to policymakers because, unless the disease is mutual, whatever given individual is unlikely to benefit.

For the private, vaccination is justified by the modify in QALE, which is nigh guaranteed, rather than past the tiny modify in the mortality run a risk. The nearly universal experience of unvaccinated individuals, including confinement to the home for 7 days and development of hundreds of itchy lesions, represents 56 h of QALE, a benefit that is µ3 times greater than that afforded by removing all adventure of expiry due to chickenpox. Of perhaps more importance, individuals choosing to forgo vaccination must be aware of their neighbors' choices. In communities that opt for mass vaccination, nonvaccinated children may face a much greater hazard of morbidity and bloodshed than they do currently.

Our study does have several limitations. The role of natural boosting in the persistence of immunity and the level of vaccination compliance necessary to induce herd immunity are unknown. We tested a wide range of assumptions regarding the exposures that an individual is likely to face. Although we predict that vaccinated individuals will exist better off than nonvaccinated individuals, we are less certain virtually the overall benefit to society. As an increasing proportion of the population chooses vaccination, unvaccinated individuals might be adversely affected, because they would be protected by herd immunity equally children, when the vaccine was effective in their peers, and then accomplish machismo with no immunity at all. On the other hand, we omitted at least 1 potential benefit of herd immunity: the protection offered to immunocompromised individuals. It is also possible that, by providing a constant reservoir of virus to unvaccinated children, infectious cases of zoster would forbid the age redistribution seen in our model [71]. Both possibilities but strengthen the case for vaccination. The vaccine'south efficacy in preventing complications also remains uncertain. Considering varicella complications occur so rarely, assessment of the efficacy and the durability of the varicella vaccine will require the follow-upwardly of millions of vaccinees for decades. If efficacy does wane, there remains the choice of administering a booster to adults [72].

In the absence of information on long-term efficacy, both pediatricians and policymakers must make up one's mind today whether or not to vaccinate. Our model favors vaccination, fifty-fifty when conservative assumptions are used. Furthermore, for the individual, the relative do good of vaccination increases every bit others also choose vaccination. Although the present benefit is small, ironically, equally more individuals voluntarily have reward of vaccination, information technology will become increasingly urgent to mandate it for everyone else.

Appendix

Appendix A

For our base of operations instance assumption, we used efficacy information from the vaccine trial done by Kuter et al. [l] in the U.s.a.. Tabular array A1 shows the observed number of cases in each twelvemonth of the study. Nosotros calculated the predicted number of cases by applying historic period-specific incidence rates to the age distribution of study participants. We divers efficacy for each yr by use of the formula [ane — (no. of observed cases/no. of predicted cases)]. For the first 2 years of the report, there was a command group that received placebo, and it was possible to calculate truthful efficacy, which closely correlated with the theoretical efficacy.

Table A1

Theoretical efficacy of varicella vaccine, according to observed and predicted cases from one vaccine trial [50].

Theoretical efficacy of varicella vaccine, co-ordinate to observed and predicted cases from one vaccine trial [l].

Table A1

Theoretical efficacy of varicella vaccine, according to observed and predicted cases from one vaccine trial [50].

Theoretical efficacy of varicella vaccine, according to observed and predicted cases from one vaccine trial [50].

We then extrapolated hereafter efficacy by fitting a linear regression through the yearly efficacy points, yielding the 1% almanac pass up. Nosotros should note that a logarithmic regression more than closely approximates the shape of the decline (R ii = .49 for the logarithmic function vs. R 2 = .44 for the linear function), with efficacy never declining below 85%. However, we chose to use the more conservative linear model, in keeping with our bias against vaccination.

Appendix B

The Index of Well-Being (IWB), which is derived from the social preferences of interviewed respondents, provides a quality-of-life measurement for any known health state by combining levels of mobility, concrete action, social activity, and symptom complexes [73]. For example, chickenpox is associated with a health state in which mobility is confined to the business firm, concrete activity is unlimited, social activeness is limited to the performance of cocky-care (merely not to work or schoolhouse activities), and the symptom complex includes a "burning or itching rash on large areas of face up, torso, arms or legs." Such a state of wellness corresponds to an IWB score of 0.6659. Whenever an private experienced one of these events, a short-term morbidity value obtained past the formula [(1 — IWB score) · duration of disease] was subtracted from the QALE. Considering chickenpox normally lasts seven days [53, 55, 63], information technology should shorten QALE by 2.34 days, according to the formula [(1 — 0.6659) · vii].

References

1

,  ,  , et al.

Postlicensure safety surveillance for varicella vaccine

,

JAMA

,

2000

, vol.

284

 (pg.

1271

-

9

)

ii

,  ,  , et al.

Postmarketing evaluation of the safety and effectiveness of varicella vaccine

,

Pediatr Infect Dis J

,

1999

, vol.

18

 (pg.

1041

-

6

)

3

,  ,  ,  ,  .

Costs and benefits of routine varicella vaccination in High german children

,

J Infect Dis

,

1996

, vol.

174

Suppl 3

(pg.

335

-

41

)

4

,  ,  ,  .

Childhood vaccination against chickenpox: an analysis of benefits and costs

,

J Pediatr

,

1994

, vol.

124

 (pg.

869

-

74

)

5

,  ,  .

An economic evaluation of varicella vaccine for pediatric liver and kidney transplant recipients

,

Clin Infect Dis

,

1993

, vol.

17

 (pg.

441

-

7

)

vi

,  ,  , et al.

Toll-effectiveness of a routine varicella vaccination plan for US children

,

JAMA

,

1994

, vol.

271

 (pg.

375

-

81

)

7

,  ,  ,  ,  .

A benefit-cost analysis of a childhood varicella vaccination programme

,

Postgrad Med J

,

1985

, vol.

61

Suppl iv

(pg.

17

-

22

)

viii

.

Varicella: complications and costs

,

Pediatrics

,

1986

, vol.

78

 (pg.

728

-

35

)

9

,  ,  ,  .

The toll-effectiveness of introducing a varicella vaccine to the New Zealand immunisation schedule

,

Soc Sci Med

,

1999

, vol.

49

 (pg.

763

-

79

)

10

,  ,  .

The cost-effectiveness of varicella vaccine programs for Commonwealth of australia

,

Vaccine

,

1999

, vol.

xviii

 (pg.

407

-

15

)

11

American Academy of Pediatrics Commission on Infectious Diseases

.

Recommendations for the use of alive attenuated varicella vaccine

,

Pediatrics

,

1995

, vol.

95

 (pg.

791

-

6

)

12

American Academy of Pediatrics

.

Commission on Infectious Diseases. Varicella vaccine update

,

Pediatrics

,

2000

, vol.

105

Pt. ane

(pg.

136

-

41

)

13

Informational Committee on Immunization Practices

.

Prevention of varicella: update recommendations of the Informational Commission on Immunization Practices

,

MMWR Morb Mortal Wkly Rep

,

1999

, vol.

48

 (pg.

ane

-

five

)

14

,  .

Reactions of pediatricians to the recommendation for universal varicella vaccination

,

Arch Pediatr Adolesc Med

,

1998

, vol.

152

 (pg.

792

-

vi

)

xv

,  .

Varicella immunization practices and the factors that influence them

,

Arch Pediatr Adolesc Med

,

1999

, vol.

153

 (pg.

357

-

62

)

sixteen

National vaccination coverage levels among children aged 19–35 months—United States, 1998

,

MMWR Morb Mortal Wkly Rep

,

1999

, vol.

48

 (pg.

829

-

30

)

17

Centers for Affliction Control and Prevention

.

Varicella-related deaths among children—United States, 1997

,

MMWR Morb Mortal Wkly Rep

,

1998

, vol.

47

 (pg.

365

-

8

)

xviii

.

Varicella: to vaccinate or not vaccinate?

,

Curvation Dis Child

,

1999

, vol.

81

 (pg.

279

-

80

)

xix

,  .

Risks and benefits of varicella vaccine

,

JAMA

,

1998

, vol.

279

 (pg.

993

-

iv

)

20

,  .

Universal vaccination against varicella

,

Due north Engl J Med

,

1998

, vol.

338

 pg.

683

 

21

,  .

Immunisation against chickenpox

,

BMJ

,

1995

, vol.

310

 (pg.

2

-

3

)

22

,  ,  , et al.

Feel and reason: xx-yr follow-upwardly of protective immunity of the Oka strain live varicella vaccine

,

Pediatrics

,

1994

, vol.

94

 (pg.

524

-

six

)

23

.

Universal vaccination against varicella

,

Northward Engl J Med

,

1998

, vol.

338

 pg.

684

 

24

,  .

Vaccination for varicella—just exercise it!

,

JAMA

,

1997

, vol.

278

 (pg.

1529

-

30

)

25

.

Varicella: to vaccinate or not to vaccinate?

,

Arch Dis Child

,

1998

, vol.

79

 (pg.

470

-

1

)

26

,  .

Immunisation against chickenpox: good argument exists for universal vaccination

,

BMJ

,

1995

, vol.

310

 pg.

873

 

27

.

Compensating children with vaccine-related injuries

,

N Engl J Med

,

1987

, vol.

316

 (pg.

1383

-

8

)

28

,  .

Clinical decision analysis by personal computer

,

Curvation Intern Med

,

1981

, vol.

141

 (pg.

1831

-

7

)

29

,  ,  .

Decision Maker iii.0: improved conclusion analysis past personal computer

,

Med Decis Making

,

1983

, vol.

3

 (pg.

39

-

43

)

30

,  .

Markov models in medical decision making: a practical guide

,

Med Decis Making

,

1993

, vol.

13

 (pg.

322

-

38

)

31

,  ,  ,  ,  .

Age-specific incidence of chickenpox

,

Public Health Rep

,

1994

, vol.

109

 (pg.

750

-

five

)

32

,  ,  ,  .

Population-based studies of varicella complications

,

Pediatrics

,

1986

, vol.

78

 (pg.

723

-

7

)

33

,  ,  ,  .

The epidemiology of varicella and its complications

,

J Infect Dis

,

1995

, vol.

172

 (pg.

706

-

12

)

34

.

Epidemiologic furnishings of varicella vaccination

,

Infect Dis Clin North Am

,

1996

, vol.

x

 (pg.

631

-

55

)

35

,  ,  ,  .

Varicella mortality: trends before vaccine licensure in the United states, 1970–1994

,

J Infect Dis

,

2000

, vol.

182

 (pg.

383

-

90

)

36

,  ,  ,  .

Chickenpox hospitalizations amidst residents of Olmsted County, Minnesota, 1962 through 1981: a population-based study

,

Am J Dis Child

,

1984

, vol.

138

 (pg.

1055

-

7

)

37

.

The epidemiology of varicella-zoster virus infections

,

Infect Dis Clin North Am

,

1996

, vol.

10

 (pg.

571

-

81

)

38

,  .

The epidemiology and burden of varicella related disease, Connecticut 1986–1995 [abstract 33]

,

Abstracts of the 31st Centers for Disease Control and Prevention (CDC) National Immunization Conference (Detroit)

,

1997

Atlanta

CDC

pg.

150

 

39

,  .

Epidemiology of main varicella and canker zoster hospitalizations: the pre-varicella vaccine era

,

J Infect Dis

,

2000

, vol.

181

 (pg.

1897

-

905

)

40

,  ,  .

Health bear on of varicella in the1980s [abstract 1138]

,

Abstracts of the 30th Interscience Conference on Antimicrobial Agents and Chemotherapy (Atlanta)

,

1990

Washington, DC

American Society for Microbiology

pg.

276

 

41

,  ,  ,  .

The incidence of herpes zoster

,

Arch Intern Med

,

1995

, vol.

155

 (pg.

1605

-

nine

)

42

,  ,  ,  ,  .

Population-based study of herpes zoster and its sequelae

,

Medicine

,

1982

, vol.

61

 (pg.

310

-

6

)

43

,  ,  ,  .

The sequelae of herpes zoster

,

Arch Intern Med

,

1997

, vol.

157

 (pg.

1209

-

13

)

44

,  .

The outcome of patients with herpes zoster

,

Curvation Dermatol

,

1957

, vol.

75

 (pg.

193

-

half dozen

)

45

.

Prospective epidemiologic written report of painful and neurologic sequelae induced by herpes zoster in patients treated early with oral acyclovir

,

Neurology

,

1995

, vol.

45

Suppl viii

(pg.

78

-

ix

)

46

,  ,  ,  .

Postherpetic neuralgia: bear upon of famciclovir, age, rash severity, and acute pain in herpes zoster patients

,

J Infect Dis

,

1998

, vol.

178

Suppl 1

(pg.

76

-

80

)

47

,  ,  ,  .

Treatment of acute herpes zoster: consequence of early (<48 h) versus late (48–72 h) therapy with aciclovir and valaciclovir on prolonged pain

,

J Infect Dis

,

1998

, vol.

178

Suppl 1

(pg.

81

-

4

)

48

Centers for Illness Control and Prevention (CDC)

.,

Mortality query using the post-obit ICD codes: [053–053.9]

,

2001

Atlanta

CDC

49

,  ,  ,  ,  ,  .

Over five-year follow-upwardly of Oka/Merck varicella vaccine recipients in 465 infants and adolescents

,

Pediatr Infect Dis J

,

1995

, vol.

14

 (pg.

874

-

9

)

50

,  ,  , et al.

Oka/Merck varicella vaccine in healthy children: final written report of a 2-year efficacy study and vii-twelvemonth follow-upwardly studies

,

Vaccine

,

1991

, vol.

nine

 (pg.

643

-

7

)

51

.

Modification of craven pox in family unit contacts by administration of gamma globulin

,

North Engl J Med

,

1962

, vol.

267

 (pg.

369

-

76

)

52

,  ,  , et al.

Acyclovir treatment of varicella in otherwise healthy children

,

J Pediatr

,

1990

, vol.

116

 (pg.

633

-

9

)

53

.

Modified varicella-like syndrome

,

Infect Dis Clin North Am

,

1996

, vol.

10

 (pg.

617

-

29

)

54

,  ,  , et al.

A controlled trial of acyclovir for chickenpox in normal children

,

N Engl J Med

,

1991

, vol.

325

 (pg.

1539

-

44

)

55

,  ,  , et al.

Clinical survey of natural varicella compared with breakthrough varicella afterwards immunization with live attenuated Oka/Merck varicella vaccine

,

Pediatrics

,

1993

, vol.

92

 (pg.

833

-

7

)

56

, et al.

Modified cases of chickenpox afterwards varicella vaccination: correlation of protection with antibody response

,

Pediatr Infect Dis J

,

1992

, vol.

eleven

 (pg.

xix

-

23

)

57

,  ,  ,  ,  ,  .

The effectiveness of the varicella vaccine in clinical practice

,

Northward Engl J Med

,

2001

, vol.

344

 (pg.

955

-

60

)

58

,  ,  ,  .

The risk of zoster after varicella vaccination in children with leukemia

,

N Engl J Med

,

1988

, vol.

318

 (pg.

543

-

8

)

59

,  ,  ,  .

The incidence of zoster later immunization with live attenuated varicella vaccine: a study in children with leukemia. Varicella Vaccine Collaborative Study Grouping

,

N Engl J Med

,

1991

, vol.

325

 (pg.

1545

-

50

)

threescore

,  ,  ,  .

Epidemiology of canker zoster in children and adolescents: a population-based study

,

Pediatrics

,

1985

, vol.

76

 (pg.

512

-

7

)

61

.

Varicella vaccine reflux

,

Pediatrics

,

1992

, vol.

89

 (pg.

353

-

4

)

62

,  .

Prevention of canker zoster

,

Infect Dis Clin North Am

,

1996

, vol.

ten

 (pg.

657

-

75

)

63

,  ,  .

Health status index: category rating versus magnitude estimation for measuring levels of well-being

,

Med Care

,

1979

, vol.

17

 (pg.

501

-

25

)

64

,  ,  ,  .

The risk of measles, mumps, and varicella amid immature adults: a serosurvey of US Navy and Marine Corps recruits

,

Am J Public Wellness

,

1993

, vol.

83

 (pg.

1717

-

20

)

65

,  ,  .

Benefits, risks and costs of immunization for measles, mumps and rubella

,

Am J Public Health

,

1985

, vol.

75

 (pg.

739

-

44

)

66

,  ,  .

Immunization of U.s.a. children with Hemophilus influenzae type b polysaccharide vaccine: a cost-effectiveness model of strategy cess

,

JAMA

,

1985

, vol.

253

 (pg.

521

-

9

)

67

,  ,  ,  ,  .

Postlicensure study of varicella vaccine effectiveness in a day-care setting

,

Pediatr Infect Dis J

,

1999

, vol.

18

 (pg.

1047

-

l

)

68

,  ,  ,  .

Modified chickenpox in children immunized with the Oka/Merck varicella vaccine

,

Pediatrics

,

1993

, vol.

91

 (pg.

17

-

22

)

69

,  ,  ,  .

Varicella outbreaks after vaccine licensure: should they brand you chicken?

,

Pediatrics

,

1999

, vol.

104

 (pg.

561

-

3

)

lxx

,  .

Gains in life expectancy from medical interventions—standardizing data on outcomes

,

N Engl J Med

,

1998

, vol.

339

 (pg.

380

-

6

)

71

,  ,  .

Mass vaccination to control chickenpox: the influence of zoster

,

Proc Natl Acad Sci U.s.

,

1996

, vol.

93

 (pg.

7231

-

5

)

72

,  ,  , et al.

The elimination of indigenous measles, mumps, and rubella from Republic of finland past a 12-year, two-dose vaccination plan

,

N Engl J Med

,

1994

, vol.

331

 (pg.

1397

-

402

)

73

,  ,  .

Health condition: types of validity and the index of well-existence

,

Wellness Serv Res

,

1976

, vol.

11

 (pg.

478

-

507

)

74

,  ,  .

Recurrent varicella-zoster virus infections in apparently healthy immunocompetent children

,

Pediatr Infect Dis J

,

1991

, vol.

x

 (pg.

569

-

75

)

75

,  ,  ,  .

Humoral amnesty and clinical reinfections following varicella vaccine in salubrious children

,

Pediatrics

,

1989

, vol.

84

 (pg.

418

-

21

)

76

,  .

Efficacy of loftier-titer live attenuated varicella vaccine in healthy young children

,

J Infect Dis

,

1996

, vol.

174

Suppl

(pg.

330

-

4

)

77

,  ,  , et al.

Oral acyclovir in the treatment of acute canker zoster ophthalmicus

,

Ophthalmology

,

1986

, vol.

93

 (pg.

763

-

70

)

78

,  ,  ,  ,  ,  .

Treatment of Ramsay Chase syndrome with acyclovir-prednisone: significance of early diagnosis and treatment

,

Ann Neurol

,

1997

, vol.

41

 (pg.

353

-

7

)

79

,  ,  .

Ramsay Hunt facial paralysis: clinical analyses of 185 patients

,

Otolaryngol Head Neck Surg

,

1986

, vol.

95

 (pg.

292

-

7

)

eighty

.

Varicella vaccine: the Japanese experience

,

J Infect Dis

,

1996

, vol.

174

Suppl

(pg.

310

-

3

)

Financial support: National Library of Medicine (grant LM 07092-07; 1000.R. was a National Library of Medicine Medical Informatics Research Beau) and National Institutes of Allergy and Infectious Diseases (midcareer clinical investigator award AI/HDO1671-01 [to M.L.B]).

Presented in part: 20th Annual Meeting of the Society for Medical Conclusion Making, Cambridge, Massachusetts, October 1998 (abstract 96).

Submit a annotate

You have entered an invalid code

Thank you lot for submitting a comment on this article. Your comment will be reviewed and published at the journal's discretion. Please check for farther notifications by email.