1309
Vaccine safety versus vaccine efficacy in immunisation programmes
community-wide immunisation programmes against childhood infections there is a conflict between the interests of the individual (vaccine safety and efficacy) and the interests of the community (vaccine uptake and level of herd immunity). Studies suggesting that the complication rate is greater with the high efficacy Urabe Am 9 mumps vaccine than with the lower efficacy Jeryl Lynn vaccine, have led to concern about whether the higher efficacy mumps vaccine should be introduced In
retained in nationwide mass immunisation programmes. We describe the use of a mathematical model to assess benefits and risks to both individual and community, and illustrate this method by reference to immunisation programmes based on these vaccines. On the basis of current epidemiological data on viral transmission and vaccine coverage in England and Wales, data on vaccine-associated and infectionassociated complication rates, and vaccine efficacies estimated from clinical trials, our analyses suggest there is littleto choose between the two vaccines, but that overall performance depends on the level of vaccine coverage in a defined population. In community-based programmes, the greater apparent safety of the Jeryl Lynn vaccine (fewer vaccineinduced complications) is offset by the greater apparent efficacy of the Urabe Am 9 vaccine (fewer complications due to natural infection). The findings suggest that it may not always be in the interests of the community to use the vaccine with the lowest complication rate. or
Lancet 1991; 338:1309-12.
Introduction The aim of community-wide mass vaccination is to reduce or eliminate disease resulting from infection. Apart from its direct protective effect, immunisation has an indirect effect-ie, herd immunity, which reduces the likelihood of infection in unvaccinated individuals to below that prevalent before the introduction of mass vaccination. However, since most vaccines have some small risk of serious complication, any assessment of the benefit from an immunisation programme must include a comparison of the number of cases of serious disease prevented by mass vaccination with that due to the vaccination programme itself. Such calculations are not straightforward because the results depend on the interrelation between various factors--eg, vaccine efficacy and safety, proportion of each cohort of children that are immunised, average age of immunisation, and non-linear effect of mass vaccination on the rate of transmission of the infectious agent (ie, the reduction in the incidence of infection does not change in direct proportion to the number immunised).’
mass
Most vaccines in common use have very good safety records. At the onset of a vaccination programme, when the incidence of infection is high in the community, the risk of serious disease from infection is always orders of magnitude greater than complications associated with vaccination. However, with high uptake, the incidence of disease from infection decreases substantially. Thus, in many industrialised countries nowadays, much attention is given to vaccine-associated complications. Ultimately, if an infectious disease has been nearly eradicated, the risks associated with vaccination are expected to exceed those associated with infection. Hence, there is a conflict of interest between the individual (risk associated with vaccination) and the group (benefits of herd immunity). Another complication is a growing dissatisfaction in some countries (eg, USA) with the failure of existing programmes to eliminate completely transmission in some communities. This failure has led to the development and use of more immunogenic vaccines with increased efficacy; these vaccines provide better protection against infection and can often be given at an earlier age. However, in some instances high efficacy may be associated with increased reactogenicity or lower safety-a finding that further exacerbates the conflict between individual and community interests. Are there circumstances in which a higher efficacy vaccine should be used even though it may lead to more cases of vaccine-associated disease than would a lower efficacy vaccine? In this article we describe a quantitative method for evaluating the benefits and risks associated with mass immunisation, with special emphasis on the comparison of low and high efficacy vaccines. We believe that it is especially timely to address this issue in industrialised countries in view of the high uptake levels of MMR (measles, mumps, and rubella), pertussis, diphtheria, and poliomyelitis vaccines; the recent availability of highefficacy vaccines (eg, for mumps and measles); and the fact that public health policy decisions are often made without analysis of the risks and benefits associated with a certain programme. We show how our method works by reference to a topical subject-namely, the relative merits of two mumps vaccine strains (Jeryl Lynn and Urabe Am 9), which differ in their efficacy and safety (ie, vaccine-induced
complications). Methods We used an age-structured mathematical model to describe the dynamics of mumps virus infection in the population of England and Wales and to assess the potential effect of different vaccination programmes. Mathematical details of the model have been described elsewhere ;2,3 here we focus on data requirements and estimations.
ADDRESS Parasite Epidemiology Research Group, Department of Biology, Imperial College of Science Technology and Medicine, London SW7 2BB, UK (D J. Nokes, PhD, Prof R M. Anderson, FRS). Correspondence to Dr D J. Nokes.
1310
Fig 2-Age-dependent risk of complications due
to mumps
virus infection.
Fig 1-(A) Age-dependent rates of mumps infection derived from two serological surveys in England and (B) the proportion of patients attending hospitals (from HIPE) for mumps-associated infection in England and Wales, in each age class, for two time periods. Assuming a normal approximation to binomial probabilities, the proportion of total complications reported for 1979-85 in the age class 0-14 years was significantly higher than the proportion in the same age class for the period 1962-69 (p < 0 0001 )
Rate of mumps transmission The model can mirror the transmission dynamics of mumps infection in a community-for example, oscillatory peaks every 3 years-and highlights the effect of a mass immunisation programme on the age-specific rates or forces of infection (ie, age-specific per caput rate at which susceptible individuals are infected). Our analysis of data from two serological surveys (done in northwest England, 1984/86/ and England, 19865) (statistical methods are detailed elsewhere->-7) shows that estimated infection rates before the introduction of mass immunisation (fig lA) have a pronounced age dependence (characteristic of many childhood viral and bacterial infections); highest transmission rates are in young schoolchildren and lower levels of transmission are in older age classes. The trend is identical in both surveys. To compare possible outcomes we did model simulations with the variables from both sets of data.
Absolute
complication rates after mumps infection
Estimation of age-dependent mumps-case complication rates (ie, the risk of complications per case of clinical or subclinical mumps infection) requires detailed information about risk of infection (clinical and subclinical) by age class-I(a)--per head of population per unit of time, and:about risk of a mumps-related complication (ideally, stratified by complication type) by age class—C(a)—per head of population,per unit of time. Such data should, preferably, be taken from the same population at the same point in time. The absolute risk of mumps complications per case of infection in age class a—R(a)—is defined as C(a) - I(a). The proportion of a population in each age class infected with mumps each year, shown by horizontal age-serological profiles,’"’ is a direct estimate of the finite per caput age-specific rate of infection per year. To overcome the difficulty of fluctuations from age class to age class, we constructed a predicted serological profile2with instantaneous age-specific per caput infection rates derived from the mumps
(A) Calculations based upon serological data from northwest England, 1984/86,° population estimates from England and Wales, 10 and a survey of infectious disease hospitals from 1958 to 19699 (and scaled by HIPE’data over same time period). Best-fit polynomial function given for each type of complications data, bottom curve=central nervous system (CNS) complication (meningitis, encephalitis, meningoencephalitis); middle curve=total complications (mcluding CNS disorders); top curve=all hospital admissions associated with mumps, with or without a defined
complication type (B) As for (A) but based on recent complication data further stratification by complication type was possible
from HIPE. No
serological profiles of northwest England, 1984/86, and England, 19864,6 (as shown in fig lA) and from these rates, we interpolated finite age-specific transmission rates, I(a). Measurement of C(a) in England and Wales is difficult because there was no comprehensive national notification system before the introduction of mass immunisation against mumps. We base our estimates on numbers of mumps-associated complications from two sources. The first source is the Hospital In-Patient Enquiry (HIPE) scheme," which records about 10% of all mumps-related hospital patients in England and Wales (or, from 1982, England only). We used exact conversion factors in our calculations.8 Reports from this scheme provide detailed age-stratified data on total hospital intake associated with mumps infection (International Classification of Diseases [ICD] 3 digit mumps code); the data are also divided into complication type but with poor age-stratification (ICD 5 digit code). As judged by HIPE data (unstratified by complication type) for two periods (1962-69 and 1979-85) most mumps-related hospital patients are in the young age classes (in which there is most infection). There is a statistically significant trend (see fig 1B) for a higher proportion of cases in the younger age classes in the more recent data set than in the data set for the 1960s. This finding accords with the suggestion of a decreasing average age of mumps infection in EnglandThe second data source, a retrospective survey of mumps-related hospital admissions, done in 16 infectious disease hospitals in England between 1958 and 1969,9 provides the only data stratified m detail by age and mumps-related complication type. They are not a complete assessment of cases in the population, but may be scaled up for each age class by means of the total estimates derived from the HIPE over the same period.8 Data from these two sources yield estimates of C(a) when divided by population size projections for England and Wales (or, where appropriate, England only),lO stratified by age, over the same time periods. Fig 2 shows the calculated R(a). A polynomial regression
1312
precise predictions of the model are sensitive to the assumptions made about efficacy, complication rates (both vaccine and naturally associated), and vaccine uptake. Evidence from long-term follow-up and outbreak studies on individuals given the Jeryl Lynn vaccine27,28 suggests that estimates based on clinical trials (as in our study) may overestimate the true protective efficacy of this strain. By implication, lower protective immunity would offset the advantage gained by the Jeryl Lynn vaccine at higher coverage levels. Therefore, the efficacy of both strains at generating immunity under routine use in the same population needs to be assessed. Faced with such uncertainties, the decision about which is the "best" mumps vaccine for community immunisation programmes remains unresolved. We wish to emphasise that any assessment of The
the relative benefits of different vaccines must take into account the complexities of the effect of mass immunisation on the age-specific incidences of infection and disease, which are especially important when the risk of severe disease due to infection changes with age.29 Given the choice of two vaccines that differ in safety and efficacy, the individual will naturally favour the one with the lowest risk, especially when, at a stage of high uptake, the incidence of disease from infection is very low. However, at the community level the decision for public health authorities is less clear cut, since both the relative ability of each vaccine to control the incidence of disease from infection (related to efficacy) and the likelihood of vaccineinduced complications (the safety issue) must be taken into account. Our analyses suggest that it may not always be in the interests of the community to use the vaccine with the lowest complication rate. When vaccine uptake is less than that required to block transmission, higher efficacy vaccines are likely to lead to fewer cases of serious disease. D. J. N. is supported by a Royal Society University Research Fellowship, and R. M. A. thanks the Wellcome Trust and the Medical Research Council for financial support. We thank Dr D. Salisbury from the Department of Health for constructive criticism of the analyses.
REFERENCES RM, Nokes DJ. Mathematical models of transmission and control. In: Holland WW, Detels R, Knox G, eds. Oxford textbook of public health. Vol 2. 2nd ed. Oxford Medical Publications, 1991: 225-52. 2. Anderson RM, Crombie JA, Grenfell BT. The epidemiology of mumps in the UK: a preliminary study of virus transmission, herd immunity and the potential impact of immunization. Epidemiol Infect 1987; 99: 65-84. 3. Anderson RM, May RM. Age-related changes in the rate of disease transmission: implications to the design of vaccination programmes. J Hyg (Camb) 1985; 94: 365-436. 4. Nokes DJ, Wright J, Morgan-Capner P, Anderson RM. Serological study of the epidemiology of mumps virus infection in northwest England. Epidemiol Infect 1990; 105: 175-95. 5. Morgan-Capner P, Wright J, Miller CL, Miller E. Surveillance of antibody to measles, mumps, and rubella by age. Br Med J 1988; 297: 770-72. 6. Farrington CP. Modelling forces of infection for measles, mumps and rubella. Stat Med 1990; 9: 953-67. 7. Grenfell BT, Anderson RM. The estimation of age-related rates of infection from case notifications and serological data. J Hyg (Camb) 1985; 95: 419-36. 8. Department of Health and Social ’Security. Reports on Hospital In-Patient Enquiry. London: HM Stationery Office, 1958-85. 9. Report. A retrospective survey of the complications of mumps. J R Coll Gen Pract 1974; 24: 552-56. 10. Office of Population Censuses and Surveys. Registrar General’s Satistical Reviews. Estimated populations. London: HM Stationery Office, 1958-69, 1979-85.
11. Christenson B, Heller L, Bottiger M. The immunizing effect and reactogenicity of two live attenuated mumps vaccines in Swedish schoolchildren. J Biol Stand 1983; 11: 323-32. 12. Vesikari T, Andre FE, Simoen F, et al. Evaluation in young children of the Urabe Am 9 strain of live attenuated mumps vaccine in comparison with the Jeryl Lynn strain. Acta Paediatr Scand 1983; 72: 32-40. 13. Vesikari T, Andre FE, Simoen E, et al. Comparison of the Urabe Am 9-Schwarz and Jeryl Lynn-Moraten combination of mumps-measles vaccines in young children. Acta Paediatr Scand 1983; 72: 41-46. 14. Popow-Kraupp T, Kundi M, Ambrosch F, Vanura H, Kunz C. A controlled trial for evaluating two live attenuated (Urabe Am 9-Schwarz and Jeryl Lynn-Moraten) in young children. J Med Virol 1986; 18: 69-79. 15. Kakakios AM, Burgess MA, Bransby RD, Quinn AA, Allars M. Optimal age for measles and mumps vaccination in Australia. Med J Aust 1990; 152: 472-74. 16. Vesikari T, Ala-Laurila EL, Heikkinen A, Terho A, D’Hondt E, Andre FE. Clinical trial of a new trivalent measles-mumps-rubella vaccine in young children. Am J Dis Child 1984; 138: 843-47. 17. Lavergne B, Frappier-Davignon L, Quevillon M, Hours C. Clinical trial of Trivirix for measles, mumps and rubella. Can Dis Wkly Rep 1986; 12: 85-88. 18.
Just M, Berger R. Immunatworf auf Impftoffe. Munch Med Wschr 1987;
19.
Just M, Berger R. Trivalent Impfstoffe. Munch Med Wschr 1987; 129:
129: 188-90. 446-47. R. Interference between strains in live virus vaccines. I. Combined vaccination with measles, mumps and rubella vaccine. J Biol Stand 1988; 16: 269-73. 21. Furesz J, Contreras G. Vaccine-related mumps meningitis—Canada. Can Dis Wkly Rep 1990; 16: 253-54. 22. Anonymous. Measles, mumps, rubella. In: Campbell AGM, ed. Immunisation against infectious disease. London: HM Stationery Office, 1990: 51-62. 23. Quast U, Hennessen W, Widmark RM. Vaccine induced mumps-like diseases. Devel Biol Stand 1979; 43: 269-72. 24. Fescharek R, Quast U, Maass G, Merkle W, Schwarz S. Measles-mumps vaccination in the FRG: an empirical analysis after 14 years of use. II. Tolerability and analysis of spontaneously reported side effects. Vaccine 1990; 8: 446-56. 25. Nalin DR. Mumps vaccine complications: which strain? Lancet 1989; ii: 20.
Berger R, Just M, Gluck
1396. 26. Nalin DR.
Mumps, measles, and rubella vaccination and encephalitis. Br Med J 1989; 299: 1219. 27. Christenson B, Bottiger M. Changes of immunological patterns against measles, mumps and rubella. A vaccination programme studied 3 to 7 years after the introduction of a two-dose schedule. Vaccine 1991; 9: 326-29. 28. Kim-Farley R, Bart S, Stetler H, et al. Clinical mumps vaccine efficacy. Am J Epidemiol 1985; 121: 593-97. 29. Nokes DJ, Anderson RM. The use of mathematical models in the epidemiological study of infectious diseases and in the design of mass immunization programmes. Epidemiol Infect 1988; 101: 1-20.
1. Anderson
From The Lancet Oliver Wendell Holmes An editorial article in a recent number of the New York Medical some interesting details of the genial Dr Wendell Holmes, who is said to be fitting himself to compose his poetic survey of long life from a personal experience of the years beyond the Psalmist’s allotment. He has of later years been a close student of the art of personal hygiene, and since his eightieth birthday in 1889 his sanitary vigilance has been incessant. Knowing that bronchitis and pneumonia are the dread enemies of old age, his rooms are furnished with thermometers, barometers, aerometers, and every other kind of instrument that will help him to ward off exposure to chill and cold. He does not rise in the morning in winter until he has ascertained that the temperature of his rooms and bath are at the required height. Dr Holmes governs his life by rule; his meals are the product of much thought and experience, and he allots the amount of time to be expended on literary work, reading, and exercise by fixed and unalterable rules. He was never robust, but his maturer years have found him still hale and uncomplaining, whilst intellectually he is still vigorous.
Journal gives
(July 11, 1891)
Vaccine safety versus vaccine efficacy in immunisation programmes
community-wide immunisation programmes against childhood infections there is a conflict between the interests of the individual (vaccine safety and efficacy) and the interests of the community (vaccine uptake and level of herd immunity). Studies suggesting that the complication rate is greater with the high efficacy Urabe Am 9 mumps vaccine than with the lower efficacy Jeryl Lynn vaccine, have led to concern about whether the higher efficacy mumps vaccine should be introduced In
retained in nationwide mass immunisation programmes. We describe the use of a mathematical model to assess benefits and risks to both individual and community, and illustrate this method by reference to immunisation programmes based on these vaccines. On the basis of current epidemiological data on viral transmission and vaccine coverage in England and Wales, data on vaccine-associated and infectionassociated complication rates, and vaccine efficacies estimated from clinical trials, our analyses suggest there is littleto choose between the two vaccines, but that overall performance depends on the level of vaccine coverage in a defined population. In community-based programmes, the greater apparent safety of the Jeryl Lynn vaccine (fewer vaccineinduced complications) is offset by the greater apparent efficacy of the Urabe Am 9 vaccine (fewer complications due to natural infection). The findings suggest that it may not always be in the interests of the community to use the vaccine with the lowest complication rate. or
Lancet 1991; 338:1309-12.
Introduction The aim of community-wide mass vaccination is to reduce or eliminate disease resulting from infection. Apart from its direct protective effect, immunisation has an indirect effect-ie, herd immunity, which reduces the likelihood of infection in unvaccinated individuals to below that prevalent before the introduction of mass vaccination. However, since most vaccines have some small risk of serious complication, any assessment of the benefit from an immunisation programme must include a comparison of the number of cases of serious disease prevented by mass vaccination with that due to the vaccination programme itself. Such calculations are not straightforward because the results depend on the interrelation between various factors--eg, vaccine efficacy and safety, proportion of each cohort of children that are immunised, average age of immunisation, and non-linear effect of mass vaccination on the rate of transmission of the infectious agent (ie, the reduction in the incidence of infection does not change in direct proportion to the number immunised).’
mass
Most vaccines in common use have very good safety records. At the onset of a vaccination programme, when the incidence of infection is high in the community, the risk of serious disease from infection is always orders of magnitude greater than complications associated with vaccination. However, with high uptake, the incidence of disease from infection decreases substantially. Thus, in many industrialised countries nowadays, much attention is given to vaccine-associated complications. Ultimately, if an infectious disease has been nearly eradicated, the risks associated with vaccination are expected to exceed those associated with infection. Hence, there is a conflict of interest between the individual (risk associated with vaccination) and the group (benefits of herd immunity). Another complication is a growing dissatisfaction in some countries (eg, USA) with the failure of existing programmes to eliminate completely transmission in some communities. This failure has led to the development and use of more immunogenic vaccines with increased efficacy; these vaccines provide better protection against infection and can often be given at an earlier age. However, in some instances high efficacy may be associated with increased reactogenicity or lower safety-a finding that further exacerbates the conflict between individual and community interests. Are there circumstances in which a higher efficacy vaccine should be used even though it may lead to more cases of vaccine-associated disease than would a lower efficacy vaccine? In this article we describe a quantitative method for evaluating the benefits and risks associated with mass immunisation, with special emphasis on the comparison of low and high efficacy vaccines. We believe that it is especially timely to address this issue in industrialised countries in view of the high uptake levels of MMR (measles, mumps, and rubella), pertussis, diphtheria, and poliomyelitis vaccines; the recent availability of highefficacy vaccines (eg, for mumps and measles); and the fact that public health policy decisions are often made without analysis of the risks and benefits associated with a certain programme. We show how our method works by reference to a topical subject-namely, the relative merits of two mumps vaccine strains (Jeryl Lynn and Urabe Am 9), which differ in their efficacy and safety (ie, vaccine-induced
complications). Methods We used an age-structured mathematical model to describe the dynamics of mumps virus infection in the population of England and Wales and to assess the potential effect of different vaccination programmes. Mathematical details of the model have been described elsewhere ;2,3 here we focus on data requirements and estimations.
ADDRESS Parasite Epidemiology Research Group, Department of Biology, Imperial College of Science Technology and Medicine, London SW7 2BB, UK (D J. Nokes, PhD, Prof R M. Anderson, FRS). Correspondence to Dr D J. Nokes.
1310
Fig 2-Age-dependent risk of complications due
to mumps
virus infection.
Fig 1-(A) Age-dependent rates of mumps infection derived from two serological surveys in England and (B) the proportion of patients attending hospitals (from HIPE) for mumps-associated infection in England and Wales, in each age class, for two time periods. Assuming a normal approximation to binomial probabilities, the proportion of total complications reported for 1979-85 in the age class 0-14 years was significantly higher than the proportion in the same age class for the period 1962-69 (p < 0 0001 )
Rate of mumps transmission The model can mirror the transmission dynamics of mumps infection in a community-for example, oscillatory peaks every 3 years-and highlights the effect of a mass immunisation programme on the age-specific rates or forces of infection (ie, age-specific per caput rate at which susceptible individuals are infected). Our analysis of data from two serological surveys (done in northwest England, 1984/86/ and England, 19865) (statistical methods are detailed elsewhere->-7) shows that estimated infection rates before the introduction of mass immunisation (fig lA) have a pronounced age dependence (characteristic of many childhood viral and bacterial infections); highest transmission rates are in young schoolchildren and lower levels of transmission are in older age classes. The trend is identical in both surveys. To compare possible outcomes we did model simulations with the variables from both sets of data.
Absolute
complication rates after mumps infection
Estimation of age-dependent mumps-case complication rates (ie, the risk of complications per case of clinical or subclinical mumps infection) requires detailed information about risk of infection (clinical and subclinical) by age class-I(a)--per head of population per unit of time, and:about risk of a mumps-related complication (ideally, stratified by complication type) by age class—C(a)—per head of population,per unit of time. Such data should, preferably, be taken from the same population at the same point in time. The absolute risk of mumps complications per case of infection in age class a—R(a)—is defined as C(a) - I(a). The proportion of a population in each age class infected with mumps each year, shown by horizontal age-serological profiles,’"’ is a direct estimate of the finite per caput age-specific rate of infection per year. To overcome the difficulty of fluctuations from age class to age class, we constructed a predicted serological profile2with instantaneous age-specific per caput infection rates derived from the mumps
(A) Calculations based upon serological data from northwest England, 1984/86,° population estimates from England and Wales, 10 and a survey of infectious disease hospitals from 1958 to 19699 (and scaled by HIPE’data over same time period). Best-fit polynomial function given for each type of complications data, bottom curve=central nervous system (CNS) complication (meningitis, encephalitis, meningoencephalitis); middle curve=total complications (mcluding CNS disorders); top curve=all hospital admissions associated with mumps, with or without a defined
complication type (B) As for (A) but based on recent complication data further stratification by complication type was possible
from HIPE. No
serological profiles of northwest England, 1984/86, and England, 19864,6 (as shown in fig lA) and from these rates, we interpolated finite age-specific transmission rates, I(a). Measurement of C(a) in England and Wales is difficult because there was no comprehensive national notification system before the introduction of mass immunisation against mumps. We base our estimates on numbers of mumps-associated complications from two sources. The first source is the Hospital In-Patient Enquiry (HIPE) scheme," which records about 10% of all mumps-related hospital patients in England and Wales (or, from 1982, England only). We used exact conversion factors in our calculations.8 Reports from this scheme provide detailed age-stratified data on total hospital intake associated with mumps infection (International Classification of Diseases [ICD] 3 digit mumps code); the data are also divided into complication type but with poor age-stratification (ICD 5 digit code). As judged by HIPE data (unstratified by complication type) for two periods (1962-69 and 1979-85) most mumps-related hospital patients are in the young age classes (in which there is most infection). There is a statistically significant trend (see fig 1B) for a higher proportion of cases in the younger age classes in the more recent data set than in the data set for the 1960s. This finding accords with the suggestion of a decreasing average age of mumps infection in EnglandThe second data source, a retrospective survey of mumps-related hospital admissions, done in 16 infectious disease hospitals in England between 1958 and 1969,9 provides the only data stratified m detail by age and mumps-related complication type. They are not a complete assessment of cases in the population, but may be scaled up for each age class by means of the total estimates derived from the HIPE over the same period.8 Data from these two sources yield estimates of C(a) when divided by population size projections for England and Wales (or, where appropriate, England only),lO stratified by age, over the same time periods. Fig 2 shows the calculated R(a). A polynomial regression
1312
precise predictions of the model are sensitive to the assumptions made about efficacy, complication rates (both vaccine and naturally associated), and vaccine uptake. Evidence from long-term follow-up and outbreak studies on individuals given the Jeryl Lynn vaccine27,28 suggests that estimates based on clinical trials (as in our study) may overestimate the true protective efficacy of this strain. By implication, lower protective immunity would offset the advantage gained by the Jeryl Lynn vaccine at higher coverage levels. Therefore, the efficacy of both strains at generating immunity under routine use in the same population needs to be assessed. Faced with such uncertainties, the decision about which is the "best" mumps vaccine for community immunisation programmes remains unresolved. We wish to emphasise that any assessment of The
the relative benefits of different vaccines must take into account the complexities of the effect of mass immunisation on the age-specific incidences of infection and disease, which are especially important when the risk of severe disease due to infection changes with age.29 Given the choice of two vaccines that differ in safety and efficacy, the individual will naturally favour the one with the lowest risk, especially when, at a stage of high uptake, the incidence of disease from infection is very low. However, at the community level the decision for public health authorities is less clear cut, since both the relative ability of each vaccine to control the incidence of disease from infection (related to efficacy) and the likelihood of vaccineinduced complications (the safety issue) must be taken into account. Our analyses suggest that it may not always be in the interests of the community to use the vaccine with the lowest complication rate. When vaccine uptake is less than that required to block transmission, higher efficacy vaccines are likely to lead to fewer cases of serious disease. D. J. N. is supported by a Royal Society University Research Fellowship, and R. M. A. thanks the Wellcome Trust and the Medical Research Council for financial support. We thank Dr D. Salisbury from the Department of Health for constructive criticism of the analyses.
REFERENCES RM, Nokes DJ. Mathematical models of transmission and control. In: Holland WW, Detels R, Knox G, eds. Oxford textbook of public health. Vol 2. 2nd ed. Oxford Medical Publications, 1991: 225-52. 2. Anderson RM, Crombie JA, Grenfell BT. The epidemiology of mumps in the UK: a preliminary study of virus transmission, herd immunity and the potential impact of immunization. Epidemiol Infect 1987; 99: 65-84. 3. Anderson RM, May RM. Age-related changes in the rate of disease transmission: implications to the design of vaccination programmes. J Hyg (Camb) 1985; 94: 365-436. 4. Nokes DJ, Wright J, Morgan-Capner P, Anderson RM. Serological study of the epidemiology of mumps virus infection in northwest England. Epidemiol Infect 1990; 105: 175-95. 5. Morgan-Capner P, Wright J, Miller CL, Miller E. Surveillance of antibody to measles, mumps, and rubella by age. Br Med J 1988; 297: 770-72. 6. Farrington CP. Modelling forces of infection for measles, mumps and rubella. Stat Med 1990; 9: 953-67. 7. Grenfell BT, Anderson RM. The estimation of age-related rates of infection from case notifications and serological data. J Hyg (Camb) 1985; 95: 419-36. 8. Department of Health and Social ’Security. Reports on Hospital In-Patient Enquiry. London: HM Stationery Office, 1958-85. 9. Report. A retrospective survey of the complications of mumps. J R Coll Gen Pract 1974; 24: 552-56. 10. Office of Population Censuses and Surveys. Registrar General’s Satistical Reviews. Estimated populations. London: HM Stationery Office, 1958-69, 1979-85.
11. Christenson B, Heller L, Bottiger M. The immunizing effect and reactogenicity of two live attenuated mumps vaccines in Swedish schoolchildren. J Biol Stand 1983; 11: 323-32. 12. Vesikari T, Andre FE, Simoen F, et al. Evaluation in young children of the Urabe Am 9 strain of live attenuated mumps vaccine in comparison with the Jeryl Lynn strain. Acta Paediatr Scand 1983; 72: 32-40. 13. Vesikari T, Andre FE, Simoen E, et al. Comparison of the Urabe Am 9-Schwarz and Jeryl Lynn-Moraten combination of mumps-measles vaccines in young children. Acta Paediatr Scand 1983; 72: 41-46. 14. Popow-Kraupp T, Kundi M, Ambrosch F, Vanura H, Kunz C. A controlled trial for evaluating two live attenuated (Urabe Am 9-Schwarz and Jeryl Lynn-Moraten) in young children. J Med Virol 1986; 18: 69-79. 15. Kakakios AM, Burgess MA, Bransby RD, Quinn AA, Allars M. Optimal age for measles and mumps vaccination in Australia. Med J Aust 1990; 152: 472-74. 16. Vesikari T, Ala-Laurila EL, Heikkinen A, Terho A, D’Hondt E, Andre FE. Clinical trial of a new trivalent measles-mumps-rubella vaccine in young children. Am J Dis Child 1984; 138: 843-47. 17. Lavergne B, Frappier-Davignon L, Quevillon M, Hours C. Clinical trial of Trivirix for measles, mumps and rubella. Can Dis Wkly Rep 1986; 12: 85-88. 18.
Just M, Berger R. Immunatworf auf Impftoffe. Munch Med Wschr 1987;
19.
Just M, Berger R. Trivalent Impfstoffe. Munch Med Wschr 1987; 129:
129: 188-90. 446-47. R. Interference between strains in live virus vaccines. I. Combined vaccination with measles, mumps and rubella vaccine. J Biol Stand 1988; 16: 269-73. 21. Furesz J, Contreras G. Vaccine-related mumps meningitis—Canada. Can Dis Wkly Rep 1990; 16: 253-54. 22. Anonymous. Measles, mumps, rubella. In: Campbell AGM, ed. Immunisation against infectious disease. London: HM Stationery Office, 1990: 51-62. 23. Quast U, Hennessen W, Widmark RM. Vaccine induced mumps-like diseases. Devel Biol Stand 1979; 43: 269-72. 24. Fescharek R, Quast U, Maass G, Merkle W, Schwarz S. Measles-mumps vaccination in the FRG: an empirical analysis after 14 years of use. II. Tolerability and analysis of spontaneously reported side effects. Vaccine 1990; 8: 446-56. 25. Nalin DR. Mumps vaccine complications: which strain? Lancet 1989; ii: 20.
Berger R, Just M, Gluck
1396. 26. Nalin DR.
Mumps, measles, and rubella vaccination and encephalitis. Br Med J 1989; 299: 1219. 27. Christenson B, Bottiger M. Changes of immunological patterns against measles, mumps and rubella. A vaccination programme studied 3 to 7 years after the introduction of a two-dose schedule. Vaccine 1991; 9: 326-29. 28. Kim-Farley R, Bart S, Stetler H, et al. Clinical mumps vaccine efficacy. Am J Epidemiol 1985; 121: 593-97. 29. Nokes DJ, Anderson RM. The use of mathematical models in the epidemiological study of infectious diseases and in the design of mass immunization programmes. Epidemiol Infect 1988; 101: 1-20.
1. Anderson
From The Lancet Oliver Wendell Holmes An editorial article in a recent number of the New York Medical some interesting details of the genial Dr Wendell Holmes, who is said to be fitting himself to compose his poetic survey of long life from a personal experience of the years beyond the Psalmist’s allotment. He has of later years been a close student of the art of personal hygiene, and since his eightieth birthday in 1889 his sanitary vigilance has been incessant. Knowing that bronchitis and pneumonia are the dread enemies of old age, his rooms are furnished with thermometers, barometers, aerometers, and every other kind of instrument that will help him to ward off exposure to chill and cold. He does not rise in the morning in winter until he has ascertained that the temperature of his rooms and bath are at the required height. Dr Holmes governs his life by rule; his meals are the product of much thought and experience, and he allots the amount of time to be expended on literary work, reading, and exercise by fixed and unalterable rules. He was never robust, but his maturer years have found him still hale and uncomplaining, whilst intellectually he is still vigorous.
Journal gives
(July 11, 1891)
![[Human papillomavirus vaccine. Efficacy and safety].](/assets/img/hold.png)
