Childhood vaccination coverage in Italy: results of a seven-region survey. The Italian Vaccine Coverage Survey Working Group.

Childhood vaccination coverage in Italy: results of a seven-region survey The Italian Vaccine Coverage Survey Working Group1 In Italy few data exist on vaccination coverage and timeliness. We therefore carried out cluster surveys on 12-23-month-olds in nine Italian cities and regions using standard Expanded Programme on Immu- nization methodology. .The study areas accounted for 40% of all live births in Italy in 1991. Coverage levels for the third dose of diphtheria and tetanus toxoids and for oral poliovirus vaccine, which are mandatory, exceeded 90% in all but one area. However, less than two-thirds of the children had com- pleted the primary vaccine series by their first birthday. The commonest reason for failure to complete the series in time was that the child had been sick and was not brought for vaccination. For the two optional vaccines (pertussis and measles) coverage was much poorer, ranging from 8% to 71% for per- tussis and from 9% to 53% for measles. The commonest reason given by the mothers for pertussis non-vaccination was that they had been advised against it, while for measles the commonest reasons were that the child was sick and that they had been advised against it. These findings suggest that although coverage for the mandatory vaccines is high, coverage for pertussis and measles is very low. Additional education of physicians and mothers is needed concerning the true contraindications for vac- cination. Also, in the absence of legislation making pertussis and measles vaccines mandatory, greater efforts are needed to convince physicians and the public about the benefits of their use. Introduction Measuring the coverage and timeliness of vaccine delivery to children is important for identifying groups at risk for low coverage. In addition, it pro- vides a useful means of monitoring the coverage and efficiency of the health care system in delivering paediatric preventive services. In Italy vaccination of all newboms with diph- theria and tetanus toxoids (DT) and with oral polio- virus vaccine (OPV) has been mandatory since the early 1960s (1). These vaccines are provided free of charge by local health units. Pertussis and measles vaccines, though recommended, have always been optional. The policy on these optional vaccines var- ies, even within regions; in some local health units, they are administered free of charge, while in others the parents must purchase the vaccines, which are then administered either by the local health unit or by the child's paediatrician. I Istituto Superiore di Sanita: N. Binkin, M.P. Carrieri, G. Sala- mina, S. Salmaso, and A.E. Tozzi; Abruzzi Region: A. Niccolini; Campania Region: P. D'Argenio, A.R. Maestro, L. Cafaro; Liguria Region: R. LoMonaco; Lombardy Region: L. Sodano; University of Ancona: P. Pandolfi, F. Filippetti, L. lncicchiti, M.M. D'Errico; Molise Region: T.M. Selvaggi; Tuscany Region: Simo- netta Viviani. Requests for reprints should be sent to: Dr S. Salmaso, Laboratorio di Epidemiologia e Biostatistica, Istituto Superiore di Sanita, viale Regina Elena 299, 00161 Rome, Italy. Reprint No. 5545 The current reporting system on vaccinations in Italy covers only those administered within the national health care system. The available data sug- gest that coverage in the last few years has been high (Public Health Service Directorate, Ministry of Health, unpublished data). However, only 11 of the 20 Italian regions provided data in 1991 and many of the data were incomplete. Furthermore, the calcula- tion of vaccine coverage using the surveillance data is difficult because accurate information is rarely available on the number of children seeking vaccina- tion through the national health service, and no indi- cation is given of whether the vaccines were admin- istered in accordance with the recommended schedule. The usefulness of surveillance data in eval- uating levels of coverage of the optional vaccines is even more limited since such vaccines are often administered in the private sector. Two large-scale vaccine coverage surveys were conducted in Italy in the mid- 1980s. The first, in 1984, focused on the coverage of the optional vac- cines, although data were also collected on OPV coverage. The parents of more than 37 000 pre- school and elementary school children aged 3-10 years in 80 communities completed a questionnaire on prior vaccination of and illnesses experienced by their children (2). The results showed that coverage for OPV was over 98% in all areas of the country. In contrast, pertussis vaccine coverage ranged from 12% in the south to 15% in the north: measles vac- cine coverage, from 4% in the south to 13% in the Bulletin of the World Health Organization, 1994, 72 (6): 885-895 © World Health Organization 1994 885 Childhood vaccination coverage in Italy north; and rubella vaccine coverage, from 6% in the south to 29% in the north. There were wide varia- tions in coverage between the 80 communities, sug- gesting that there was considerable variation in the policies of the local health units responsible for deciding local vaccination policy. The second of the two surveys was conducted in 1985-86 among children born in June 1983 in five Italian cities, four of which were in the south of the country (3). This study showed that 60-96% of the children had completed the three-dose polio series; however, even in the cities where coverage was high, less than 70% had been fully vaccinated by their first birthday. One additional recent study that used vaccine sales data found that pertussis vaccine coverage increased from an estimated 22% for the 1985 birth cohort to 50% for the 1989 birth cohort (4). These same data, however, demonstrated a decline in coverage for the 1990 cohort, estimated to be 38%. Since the last surveys were undertaken, cam- paigns have been conducted in several parts of Italy to increase optional vaccine coverage, and a greater number of local health units have begun to offer measles and pertussis vaccination free of charge. To assess the current coverage for both mandatory and optional vaccines, the adherence to the vaccine schedule, and the reasons for late or non-vaccination, as well as to evaluate the accuracy of the available data on vaccine coverage, we undertook a vaccine coverage survey. The study, which used standard Expanded Programme on Immunization (EPI) cluster survey methodology, involved nearly 1800 children aged 12-23 months living in seven of the 20 regions of the country. These seven regions accounted for 40% of all live births in Italy in 1991 (5). Methods Selection of regions, cluster sites, and individual children In six of the seven study regions, participants in a 2- year field epidemiology training programme sponsor- ed by the Istituto Superiore di Sanita in Rome, and the Centers for Disease Control and Prevention, Atlanta, GA, USA, organized and conducted the sur- vey in their home regions; in the seventh region, residents of a university-based public health speciali- zation programme conducted the survey. Although not randomly chosen, the regions were geographical- ly dispersed; two were in the north (Lombardy and Liguria), three in the centre (Tuscany, the Marches, and Abruzzi), and two in the south (Campania and Molise). In Lombardy and Campania, separate sur- veys were conducted in the regional capitals (Milan and Naples, respectively) and in the remaining com- munes of the region, resulting in a total of nine study areas. Selection of cluster sites and individual children Except as noted below, standard EPI cluster survey methodology, involving systematic sampling from a cumulative population list, was used to select the 30 cluster sites in each of the study areas.a In five regions, the entire population served as the sampling universe. In Campania, Naples and the remainder of the region were each considered as separate sam- pling universes and 30 clusters were chosen in each. In Milan, a computerized list of births was available, and a random sample was selected directly; in the remainder of Lombardy, 30 clusters were chosen from a cumulative population list. In all cases, the population lists for sampling were based on births in 1991, by commune of residence of the mother, which were obtained from the communal vital record regis- ter. When the number of births was small, adjacent communes were aggregated prior to sampling. After 30 communes had been selected within the sampling universe, the local health unit of the com- mune was asked to obtain a list from the local birth registration office of all children who would be 12-23 months of age by the date that the study was to begin in the region. Random number tables were used to select seven of the children in the eligible age group as well as a list of seven altematives. Survey methods The standard questionnaire for the EPI surveysa was modified to reflect the Italian vaccine schedule and likely reasons for non-vaccination. The resulting questionnaire was pre-tested on a small sample of children in Milan and Tuscany and modifications made prior to training the interviewers. The background of the interviewers varied by region, but included vaccination nurses and physi- cians working in the local health units, postgraduate physicians, and the survey organizers themselves. In each region, a training session was held to explain the goals of the survey, the methods, and to review the questionnaire and its completion. The organizers in each area provided telephone back-up during the survey to deal with any questions or problems that developed. a Training for mid-level managers: the EPI coverage survey. Unpublished document WHO/EPI/MLM/91.10, 1991. 886 WHO Bulletin OMS. Vol 72 1994 Italian Vaccine Coverage Survey Working Group The interviewers attempted to make telephone contact with the mothers on their list to make an appointment. If the mother could not be contacted after three tries or she did not have a telephone num- ber, a visit was made to the household. If no one was at home, an attempt was made to verify with the neighbours that the family still lived there, and a note was left asking the mother to contact the inter- viewer to arrange an appointment. Only when a mother could not be reached after three telephone calls and a visit was one of the children on the list of alternatives used. All interviews were conducted in person, usually at the child's home but occasionally in the local health unit. Definitions For the analysis, children were considered to be fully immunized against diphtheria, tetanus, poliomyelitis and pertussis if they had received three doses of the corresponding vaccines by the time of the interview. In all regions except Tuscany, coverage for hepatitis B vaccine was considered only for those children bom after 16 June 1991, when the law conceming hepatitis B vaccine came into effect. In Tuscany, the corresponding date used was 25 October 1991, when the region sent a circular to the local health units mandating the use of this vaccine. For measles vac- cine, coverage was calculated only for those chil- dren aged .15 months. The timeliness of vaccination was assessed only for DT and OPV, and only for those children who had a vaccination card. Under the vaccination sched- ule (Table 1), the first doses of DT and OPV should be administered within the third month of life, with the second dose given 6-8 weeks later, by the fourth or fifth month of life, and the third dose 6 months after the second dose, i.e., by the tenth or eleventh month of life. Interpretations of the age at which vaccinations are to be conducted vary; within the third month, for example, has been interpreted in some areas to mean before 8 weeks of age, while in others the same recommendation is interpreted to mean at or after 12 weeks of age. Because of these ambiguities, the first dose was considered to be delayed if it had been administered after 4 months (16 weeks) of age; for the second and third doses, respectively, the corresponding cut-offs were 6 months (24 weeks), and the child's first birthday. The mothers of children who had not completed the three doses of the mandatory vaccinations by their first birthday, of those who had not received pertussis vaccine, and of those who were >16 months of age and had not yet received measles vaccine were interviewed to determine why, using a data col- lection instrument similar to that recommended for use in EPI surveys but modified to reflect the situa- tion in Italy. Use of routine surveillance for monitoring vaccination coverage For four of the regions included in the survey, rou- tine surveillance data on the number of first, second, and third doses of DT, OPV, and pertussis vaccines were available for either 1990 or 1991. Using as denominator the number of resident children under 1 year of age, we estimated the vaccination coverage and compared the results with those obtained in the coverage survey. Health care delivery and the use of optional vaccines In addition to the coverage survey, we also conduc- ted a brief telephone survey in each region of all the local health units from which clusters had been drawn. The data collected included policies on noti- fying mothers to bring their children for a first visit, whether they interpreted the regulations to mean that the first dose should be given during or after the third month of life, whether they offered free pertus- Table 1: The Italian national vaccination schedulea Age administered Vaccineb Interval between doses 3rd month of life OPV + DT (optional P) + HBV 4th-5th month of life OPV + DT (optional P) + HBV 6-8 weeks after first dose 10th-llth month of life OPV + DT (optional P) + HBV 6 months after second dose 15th month of life (Optional measles, mumps) 3rd year of life OPV >1 year after third dose 6th year of life DT 4-5 years after third dose Pre-puberty (Optional rubella for females) 14th year of life T Every 10 years a Source: ref. (1). b OPV = oral poliovirus vaccine; DT = diphtheria-tetanus; P = pertussis; HBV = hepatitis B vaccine; and T = tetanus. WHO Bulletin OMS. Vol 72 1994 887 Childhood vaccination coverage In Italy sis or measles vaccine, and whether they had ever conducted vaccine campaigns and if so for which vaccines. The percentage of local health units engag- ing in each of these behaviours was plotted against the vaccination coverage for that survey universe, and a regression analysis was carried out using cover- age as the dependent variable. When more than one cluster was drawn from a local health unit, the res- ponse of the unit was correspondingly weighted such that the total number of responses was 30. Data entry and statistical analysis Data entry and analysis were performed using Epi Info version 5.0 (6). For each variable, 95% confi- dence intervals that took into account the cluster design of the survey were calculated using an experi- mental Epi Info routine and verified using Epi Table software (Medecins sans Frontieres, Paris). Results Study population A total of 210 mothers were interviewed in each of the eight study universes where cluster sampling was used; in the remaining study universe where random sampling was employed, 100 mothers were inter- viewed. Table 2 shows the the proportion of children in each cluster drawn from the alternative list, and the sex, maternal age, and maternal education levels for the study populations. The proportion drawn from the alternative list ranged from 8.7% in Lom- bardy to 20.0% in Milan. Most replacements occurred because the mother had moved or could not be con- tacted; the vast majority of those contacted agreed to participate. The characteristics of the mothers in the study did not differ significantly from those obtained from birth records from each of the regions for 1990 (ISTAT, unpublished data). Mandatory vaccines Vaccine coverage. Coverage, by region, for the first and third doses of DT and OPV for the entire cohort and for the first and third doses of hepatitis B vac- cine for those born after this vaccine became manda- tory, is shown in Table 3. The coverage was virtu- ally universal (.98%) for the first dose of DT and OPV in all areas studied. By contrast, the levels of the third dose varied considerably, ranging from a high of 99% in the Marches to a low of 77% in Cam- pania. For the newly mandatory hepatitis B vaccine, the coverage for the third dose ranged from 63% in Abruzzi and Naples to 92% in Lombardy. Adherence to vaccine schedule. The proportions of children whose first, second, and third doses of OPV and DT (or diphtheria-tetanus-pertussis (DTP)) vac- cines were delayed are shown in Table 4. The pro- portion of vaccinated children receiving a third dose of DT within the recommended period ranged from 33% in Naples to 69% in Lombardy. For OPV the proportions were similar. In general, the proportion who were vaccinated late increased with the dose number of the series. Reasons for incomplete vaccination. In all the regions studied, the commonest reason given for the delay in completing the primary DT and OPV sched- Table 2: Characteristics of the study population, by region, Italian vaccination coverage survey, 1993 Campania Lombardy Characteristic Abruzzi Naples Other Liguria Milan Other Marches Molise Tuscany Sample size 210 210 210 210 100 210 210 210 210 % drawn from 14.7 14.8 10.0 15.2 20.0 8.7 13.3 11.9 11.0 alternate list % male 50.5 51.9 52.4 51.9 58.0 49.5 54.3 55.2 50.0 Matemal age (years) Mean ± SD 30.0 ± 4.4 29.6 ± 5.2 29.5 ± 5.1 31.4 ± 4.6 32.0 ± 4.9 30.6 ± 4.6 30.8 ± 4.9 30.3 ± 4.9 31.4 ± 5.0 Median 29 (18-44)a 29 (18-44) 29 (16-45) 31 (18-44) 32 (19-42) 31 (20-42) 30 (19-47) 30 (18-44) 31 (19-42) Maternal education level (/04) None 0 1.9 0.5 0.5 1.0 0 0.5 0 0.5 Elementary 5.7 28.6 17.6 1.4 7.0 3.8 4.8 7.6 5.2 Middle school 41.9 31.4 42.4 36.8 20.0 36.7 35.7 42.9 39.0 High school 44.3 33.3 29.5 49.8 53.0 49.0 47.6 42.9 42.9 University 8.1 4.8 10.0 11.5 19.0 10.5 11.4 6.7 12.4 a Figures in parentheses are the range. WHO Bulletin OMS. Vol 72 1994888 Italian Vaccine Coverage Survey Working Group Table 3: Percentage coverage of first and third doses of diphtheria-tetanus, poliovirus, and hepatitis B vaccines, by region, for children aged 12-23 monthsa Campania Lombardy Vaccineb Abruzzi Naples Other Liguria Milan Other Marches Molise Tuscany DT1 99.5 98.6 99.1 99.5 100 100 100 100 100 (98.6-100)c (97.0-100) (97.8-100) (96.6-100) DT3 95.2 77.1 87.2 96.7 91.0 94.8 98.6 95.2 95.2 (91.1-99.3) (71.0-83.2) (82.5-91.9) (94.1-99.2) (85.4-96.6) (88.8-100) (97.0-100) (92.2-98.3) (91.6-98.9) OPV1 99.5 98.6 98.1 99.5 100 100 100 99.5 100 (98.6-100) (97.0-100) (96.3-99.9) (98.6-100) (98.6-100) OPV3 97.6 77.6 88.2 96.7 91.0 97.6 98.6 94.3 95.2 (95.7-99.6) (71.5-83.7) (83.7-92.6) (94.1-99.2) (85.4-96.6) (95.3-100) (97.0-100) (90.9-97.7) (91.6-98.9) HBV1 89.2 84.4 79.3 94.0 96.6 99.2 95.7 82.3 98.3 (80.9-97.4) (76.4-92.4) (67.5-91.1) (88.7-99.3) (91.8-100) (97.7-100) (89.8-100) (71.6-93.9) (90.7-100) HBV3 62.5 62.3 63.1 85.9 79.3 92.2 87.1 68.3 81.0 (51.3-74.8) (53.9-70.7) (50.4-75.7) (78.7-93.2) (68.8-89.8) (87.9-96.0) (78.7-95.6) (56.3-80.2) (68.3-93.8) a For hepatitis B vaccine only for children born after 16 June 1991, when vaccine was made mandatory (25 October 1991 for Tuscany). b DT = diphtheria-tetanus; OPV = oral poliovirus vaccine; HBV = hepatitis B vaccine. c Figures in parentheses are the 95% confidence intervals. ule was that the child was sick and was not brought for vaccination, although the proportion of mothers who gave this reason varied from 37% in Milan to 69% in Molise. Other reasons included the follow- ing: lack of awareness that the vaccination should be completed by a certain deadline; the child was sick and brought in, but vaccination was deferred; and they were not notified by the local health unit to come or the appointment given for the third dose was after the child's first birthday. It was not possible to determine whether the decision not to bring a sick child for vaccination was made in consultation with the child's paediatrician. However, for those regions that collected data on the illnesses reported when the child should have been vaccinated, many were not true contraindications to vaccination. When we examined the health care factors asso- ciated with delayed vaccination, little correlation was seen between the following: the timely completion of the third dose and the percent of the local health units that sent out letters for the first visit (r = 0.22); whether these units considered that the first dose should be given during or after the third month of life (r = 0.11); or whether they performed active fol- low-up of the children who did not come for their first visit (r = -0.21). Table 4: Percentage of children receiving first, second, and third doses of diphtheria-tetanus and poliovirus vac- cines within one month of the deadline, by region, for 12-23-month-oldsa Campania Lombardy Vaccineb Abruzzi Naples Other Liguria Milan Other Marches Molise Tuscany DT1 87.1 54.1 68.4 90.4 83.7 93.2 89.1 87.6 87.6 (80.6-93.6)C (45.1-63.1) (60.8-76.1) (87.1-93.8) (74.7-90.7) (90.0-96.5) (83.9-94.2) (82.2-93.0) (83.0-92.2) DT2 81.8 49.0 59.2 87.0 72.4 87.8 81.8 72.5 81.4 (75.0-88.7) (41.3-56.7) (50.3-68.1) (83.0-90.9) (62.4-81.4) (81.8-93.8) (74.3-89.2) (62.4-82.6) (76.4-86.5) DT3 54.5 33.3 43.1 63.4 54.4 68.7 55.3 51.8 56.5 (43.3-65.6) (26.4-40.2) (34.2-52.3) (56.7-70.0) (44.1-64.7) (58.9-78.5) (45.9-64.6) (40.8-62.7) (48.0-65.0) OPV1 86.6 54.6 69.6 90.0 83.7 92.7 87.5 84.7 88.1 (80.1-93.1) (45.7-63.5) (61.9-77.4) (86.6-93.3) (74.7-90.7) (89.5-95.9) (81.6-93.4) (78.8-90.7) (83.8-92.4) OPV2 81.3 50.5 59.3 87.9 77.3 87.8 80.7 71.3 83.8 (74.0-88.7) (42.5-58.4) (49.8-68.8) (84.0-91.9) (67.3-85.3) (81.0-92.6) (73.1-84.4) (61.1-81.6) (79.0-88.6) OPV3 55.6 34.6 44.8 62.1 56.7 65.2 55.3 52.4 55.6 (45.3-66.0) (27.2-41.9) (34.9-55.4) (56.2-67.9) (46.5-66.9) (55.6-74.8) (46.2-64.3) (41.2-63.6) (45.3-66.0) a In Italy the following deadlines are specified by law: 4 months of age for the first dose; 6 months for the second; and 12 months for the third dose. b DT = diphtheria-tetanus; OPV = oral poliovirus vaccine. c Figures in parentheses are the 95% confidence intervals. WHO Bulletin OMS. Vol 72 1994 889 Childhood vaccination coverage in Italy Optional vaccines Vaccine coverage. The coverage for the optional vaccines was considerably lower than that for the mandatory vaccines (Table 5). For pertussis vaccine, the coverage for the third dose ranged from 7.6% in Molise to 71% in Lombardy. For measles vaccine, the coverage was also low, ranging from 8.6% in Naples to 53% in Lombardy. Reasons for non-use of measles and pertussis vac- cines. The reasons given by mothers for not having had their children vaccinated against measles and pertussis are given in Table 6. For pertussis, the commonest reasons included the following: they had been advised against it; the vaccine had undesirable side-effects; they did not know the vaccine existed; it was better to get the disease; and the disease was not dangerous. In the areas with higher coverage, the first two of these reasons were commoner; in those with lower coverage, the lack of knowledge about the vaccine's existence was more dominant. The rea- sons for not vaccinating against measles were more varied. The commonest, however, included that the child was sick and was not brought for vaccination, that they had been advised against it or told to wait until the child was older, that it was better to get the disease, and that the vaccine had undesirable side- effects. Except in Naples and Campania, the cost of vaccine was rarely cited as a reason for non-vaccina- tion for either pertussis or measles. Of note is that 14% of mothers in Naples and 7% of those in Cam- pania stated that their child had already had measles; in the other areas, the corresponding values were 0-3.5%. A strong correlation was observed between the proportion of local health units in each study uni- verse providing pertussis vaccine free of charge and the actual coverage (r = 0.95; 95% CI = 0.77-0.99; P = 0.001). Because so few of the local health units had conducted pertussis vaccine campaigns (0-23%, depending on the region), we did not examine the correlation of vaccine coverage with such cam- paigns. As was the case for pertussis vaccine, the corre- lation between the proportion of local health units in each study universe that provided measles vaccine free of charge and the actual coverage was high, but not as high as for pertussis (r = 0.75; 95% CI = 0.17-0.94; P = 0.02). Although many of the local health units stated that they had carried out vaccine campaigns, there was no correlation between the pro- portion offering measles vaccine campaigns and the coverage (r = 0.16; 95% CI = -0.56 to 0.75). Use of routine surveillance for monitoring vaccina- tion coverage. Data from routine surveillance were available for four of the seven study regions for either 1990 or 1991 (Table 7). For the mandatory vaccines, the surveillance system appears to have underestimated the coverage in two of the regions, with the surveillance estimates being less than the lower 95% confidence limit of the survey value. In the Marches, the number of each dose of DT and OPV administered exceeded the number of 1-year- olds, resulting in coverages greater than 100%. In the remaining regions, the survey and surveillance coverages were similar. For pertussis vaccine, the coverage obtained from the surveillance system dif- fered substantially from that obtained in the survey in all four regions. Discussion Although the EPI cluster survey methodology has been more commonly applied in developing country settings, the present study confirms the feasibility of using it to obtain timely information on the vaccina- tion status of children in a developed country, rapid- ly and at reasonably low cost, using existing health care personnel. In countries such as Italy that have a more extended vaccine schedule than that used in developing countries, consideration should be given, however, to performing such surveys on slightly Table 5: Percentage coverage of first and third doses of pertussis vaccine and of measles vaccine, by regiona Campania Lombardy Vaccine Abruzzi Naples Other Liguria Milan Other Marches Molise Tuscany Pertussis 1 25.7 15.2 19.4 43.8 62.0 79.0 55.7 11.0 30.5 (1 6.2-35.2)b (9.1-21.4) (18.1-26.7) (31.4-56.2) (51.0-72.0) (70.9-87.2) (45.0-66.4) (5.1-16.8) (19.4-41.5) Pertussis 3 21.4 12.4 15.2 37.6 56.0 71.0 51.0 7.6 26.2 (18.1-29.8) (7.2-17.5) (10.2-20.2) (26.3-49.0) (45.0-66.0) (61.9-80.0) (40.7-61.2) (3.7-11.6) (15.1-37.3) Measles 18.7 8.6 16.1 32.3 43.3 52.6 35.8 11.0 44.4 (11.6-25.9) (3.1-14.2) (8.3-23.8) (24.5-40.2) (30.3-56.3) (42.8-62.5) (26.8-44.9) (5.6-16.4) (35.2-53.5) a The data for measles vaccine are for children >15 months of age. b Figures in parentheses are the 95% confidence intervals. WHO Bulletin OMS. Vol 72 1994890 Italian Vaccine Coverage Survey Working Group Table 6: Distribution of the reasons given by the mothers for non-vaccination with pertussis and measles vaccine, by region Campania Lombardy Abruzzi Naples Other Liguria Milan Other Marches Molise Tuscany Pertussis vaccine n = 165 n = 184 n = 178 n = 65a n = 41 n = 52 n = 103 n = ggb n = 155 Were advised against it (%) 26.7 20.7 16.3 69.2 39.0 48.1 35.9 22.2 14.5 Undesirable side-effects (%) 26.7 10.9 13.5 0 14.6 15.4 16.5 10.1 15.5 Didn't know there was 7.3 21.2 20.8 1.5 2.4 5.8 10.7 29.3 19.4 a vaccine (%) Better to get the disease (%) 10.9 8.7 12.4 0 7.3 0 1.9 2.0 3.9 Disease not dangerous (%) 2.4 5.4 4.5 0 2.4 1.9 2.9 1.0 1.3 Measles vaccine n = 118 n = 140 n = 130 n = 59c n = 37 n = 71 n = 105 n = 141 n = 86 Sick, not brought in (%) 13.6 9.3 6.9 18.6 10.8 23.9 29.5 12.1 26.7 Were advised against it (%) 7.6 17.9 10.0 37.3 21.6 11.3 23.8 6.4 9.3 Better to get the disease (%) 15.3 12.9 15.4 0 13.5 2.8 3.8 10.6 5.8 Undesirable side-effects (%) 19.5 5.0 16.2 0 8.1 4.2 12.0 5.7 2.3 Physician suggested waiting (%) 0.8 0.7 1.5 10.2 2.7 2.8 3.8 0.7 31.4 a Data for 65 of the 131 unvaccinated or incompletely vaccinated. b Data for 99 of the 194 unvaccinated or incompletely vaccinated. c Data collected for 59 of the 123 unvaccinated. older children to avoid problems with incomplete denominators for measles vaccination and to deter- mine more accurately the duration of delays in receiving the final dose of the three-dose vac- cines without resorting to the use of more complicated person-time denominators. Our study may have overestimated the rate of vaccine coverage. The percentage of mothers drawn from the altemative list in most areas was relatively low, and we had virtually a 100% response rate among those mothers who were contacted success- fully. None the less, for practical reasons, our sam- pling was based on birth registration data for the communes included in the survey. Although registra- tion is obligatory within 10 days of birth in order to receive health benefits and for school attendance, undocumented immigrants, who may also be less likely to be immunized, were not included in the study sample. Also excluded were any families who had moved into the area after the child was bom, although geographical mobility in Italy is low (5). These deficiences would most probably have had the greatest effect on the coverage estimates for Milan and Lombardy, since most immigration and in-coun- try migration are to the more heavily industrialized north (5). It is difficult to estimate the impact of this exclusion, but it is unlikely that it dramatically affected the reported coverage levels. Although the use of several interviewers might have affected the consistency of the data collection, information on the vaccination dates was obtained directly from vaccination cards, minimizing any dif- ficulties with interviewing technique and with mater- nal recall. However, use of several interviewers might have caused some inconsistencies in recording the reasons for non-vaccination. The high coverage levels against diphtheria, teta- nus, and poliomyelitis are reflected in the virtual Table 7: Comparison of the percentage vaccination coverage for the third doses of diphtheria-tetanus (DT), oral poliovirus (OPV), and pertussis vaccines from the vaccination survey and from routine surveillance Source of surveillance data Lombardy Regional health (except Milan) authority Ministry of Health Regional health authority Ministry of Health Year 1990 1991 1990 1991 DT 3 (%) OPV 3 (%) Survey Surveillance Survey Surveillance 94.8 97.8 98.6 95.2 97.6 97.8 Pertussis 3 (%) Survey Surveillance 71.0 58.1 >100 98.6 >100 51.0 42.6 89.9 94.3 76.0 7.6 1.3 95.2 90.7 95.2 96.4 26.2 13.4 a Based on 6 months of data. WHO Bulletin OMS. Vol 72 1994 Region Marches Molise Tuscanya 891 Childhood vaccination coverage In Italy absence of these diseases among young children in Italy. In 1992, no cases of tetanus or diphtheria were notified among under-5-year-olds, and the last docu- mented case of poliomyelitis in a child was an imported one in 1988 (Ministry of Health, unpub- lished data). The levels of vaccine coverage in all the areas studied except Campania and its capital city, Naples, approach or exceed the district-specific goals set by WHO for 1997;b they also compare favour- ably with coverage rates elsewhere in Western Europe (7)c and are higher than the levels reported for many parts of the USA (7, 8).d Although vaccination coverage was high, there were some delays in vaccine delivery, especially for the third dose. The rates of failure to adhere to the vaccine schedule were particularly high in the com- mune of Naples, where two-thirds of the children had not received the third dose of DT and OPV by their first birthday. At least part of the delay in some of the study areas appears to have been due to late initiation to the entire vaccination series. Such delays may have resulted from the failure of the local health unit to receive a list of recent births from the com- munal birth register, failure of the health unit to noti- fy the parents in time to appear for the first appoint- ment, or delays in parents bringing their child once they had been notified. Another important factor leading to delays was the postponing of vaccination due to intercurrent illness by either the mother or health personnel, which may have varied from area to area depending on the local health policies. Fre- quently, further questioning of mothers revealed that many of the children had illnesses that were not truly contraindications for vaccination. Thus it appears that improving the system of notifying the mother about the first vaccination visit, combined with pro- viding physicians and vaccination staff, as well as mothers, with clear guidelines about true contraindi- cations would probably increase the timeliness of vaccination coverage. Educating physicians and vac- cination staff may perhaps be best accomplishSd via circulars, the medical literature, and scientific/meet- ings. For educating mothers, it may be helpful to develop an information sheet to accompany the vacci- nation booklet and to have paediatricians and vaccina- tion staff discuss contraindications with mothers dur- ing routine health care visits and vaccination sessions. b Meeting of the Expanded Programme on Immunization, WHO Regional Office for Europe, Milan, 23-25 June 1992. c Expanded Programme on Immunization Information System. Summary for the WHO European Region. Unpublished docu- ment, WHO/EPI/CEIS/93.2, 1993. d Stoltenow CL et al. Early childhood vaccination in two rural Nebraska counties. Paper presented at: EIS (Epidemic Intelli- gence Service) conference, Atlanta, GA, USA, 19-23 April 1993. In contrast to the high levels of vaccine cover- age for the obligatory vaccines, the levels for the optional vaccines were very low. Although caution is required in generalizing our results to the entire country, the levels of pertussis and measles vaccina- tion in Italy appear to be the lowest in industrialized Europe.d Reported levels of pertussis vaccine cover- age range from 54% in Germany to 100% in San Marino; for measles vaccine, coverage ranges from 38% in Austria to 100% in San Marino.d Not surpris- ingly, pertussis and measles remain common illness- es in the face of the relatively low coverage rates, with the mean annual number of notified cases among under-5-year-olds for the period 1985-89 being as follows: 7731 for pertussis (range: 3680- 10 732; rate: 266 per 100 000) and 12 022 for mea- sles (range: 2892-22 737; rate: 414 per 100 000) (9-13). The experience in Italy with the optional vac- cines contrasts sharply with that in several other European countries in which some or all of the child- hood vaccines are optional rather than mandatory, but the coverage is none the less quite high (14). The low coverage in Italy appears to be multifactorial and involves the following factors: the access to free vaccine on site; the willingness of local physicians to recommend the optional vaccines when their admin- istration is not legally mandated (15); the attitudes of the mothers themselves about the safety of the vac- cines and the perceived severity of the diseases; absence of ongoing educational efforts; and perhaps the importance of autonomy and independent choice in the Italian culture (16). Mothers of children who elected not to have their children vaccinated against measles and pertus- sis frequently cited as a reason that the family paedi- atrician had not recommended such a course or, for measles vaccine, that such action should be delayed until the child was older. However, two recent sur- veys of paediatricians, vaccination doctors, and gen- eral practitioners suggest that, although their lack of advocacy may play a role in this respect, it does not fully explain the low coverage. A random sample telephone survey of the knowledge, attitudes, and practices of 132 paediatricians and 48 vaccination doctors conducted in nine regions of Italy in 1991 showed that only two out of three paediatricians and four out of five vaccination doctors recommended that children in their practices be vaccinated against pertussis (17). A similar level (65%) was found in a mail survey of paediatricians and general practition- ers throughout the country conducted in 1991 (15). A greater proportion recommended measles vaccina- tion: in the nine-region telephone survey, 96% of paediatricians and 100% of vaccination doctors reported advising mothers to have their children vac- 892 WHO Bulletin OMS. Vol 72 1994 Italian Vaccine Coverage Survey Working Group cinated against measles (17); in the national mail survey, 95% of paediatricians and 73% of general practitioners favoured the use of measles, mumps, and rubella vaccine (15). In both surveys, however, physicians may have overstated their advocacy role, and in the mail survey, the low response rate (21%) may have influenced the results. Furthermore, nei- ther of the surveys determined whether physicians were actively counselling their patients or providing their opinions only when specifically asked. Among the physicians who did not recommend use of the optional vaccines, misconceptions about their efficacy and adverse effects as well as about the severity of the diseases they prevent appeared to be common. In the telephone survey, both paediatri- cians and vaccination physicians tended to underesti- mate the efficacy of pertussis vaccination and more than half were worried about adverse reactions asso- ciated with the vaccine (17); in the mail survey, phy- sicians reported that the potential post-vaccination complications of pertussis and measles vaccines were worse than those associated with the diseases themselves (15). Physicians also seem to be con- cerned about the medico-legal consequences of administering optional vaccines. Recently, legisla- tion was passed in Italy that provides for compensa- tion for severe adverse events resulting from the administration of the mandatory vaccines, but no provisions are made for the optional vaccines. The vaccination campaigns conducted at the local health unit or regional levels appear to have had somewhat limited results. There may be several reasons for this. First, the number of local health units that have launched campaigns may not have been enough to have had a substantial effect, particu- larly for pertussis, where only 26% of the local health units participating in the national survey have carried out campaigns (15). Second, in some areas, the campaigns have been limited to sending letters to parents and/or physicians, which may have restricted their effectiveness. Finally, the measles campaigns conducted to date have been largely aimed at provid- ing complete coverage for children aged 1-8 years and much of their efforts have focused on reaching children in the school setting. This strategy was adopted to increase rapidly the vaccine coverage in the country, and thereby decrease measles transmis- sion, and also to prevent development of a suscept- ible pool of older children and adolescents. Although it does appear to have decreased the transmission of measles (18), this strategy has not resulted in sub- stantially higher coverage among the age group included in our survey, perhaps because of the lack of continuity implicit in a campaign and the lack of institutionalization of immunization for young chil- dren in the local health units. Our study findings showed that the availability of free measles and pertussis vaccine was positively correlated with vaccine coverage. Although the costs of vaccine at the time of the study were not high (6000 lira (US$ 4) for a three-dose DTP series; 11 000 lira (US$ 7.30) for measles vaccine alone; and 17 000 lira (US$ 11.30) for measles, mumps, and rubella vaccine (19)), the inconvenience and time spent in buying the vaccines at a pharmacy and visiting the paediatrician or local health unit to have them administered may be substantial. Caution, how- ever, is required in attributing our findings entirely to the availability of free vaccine, since those health units that provide vaccines free of charge may also advocate their use more actively. Whether high pertussis and measles vaccine coverage can be achieved in Italy without making these vaccines mandatory is an open question. Cer- tainly this is one solution, as evidenced by experi- ence in Italy with hepatitis B vaccine. For the chil- dren in our cohort born prior to 16 June 1991, when hepatitis B vaccine was still optional, the levels of coverage were considerably lower than those born after this date when its administration became man- datory. The coverage will probably increase even further, since in some of the regions, it took time to implement the new policy and obtain adequate sup- plies of vaccines. Indeed, the government has delib- erated for many years about the advisability of mak- ing the currently optional vaccines mandatory, and if the efficacy of the new acellular pertussis vaccines can be demonstrated, and the safety issues related to pertussis vaccine are no longer an obstacle, the cur- rent policy may change. In the meantime, it would appear worthwhile to implement several measures to improve vaccine coverage for the optional vaccines similar to those that have worked successfully elsewhere. These might include the following: provision of vaccines free of charge; provision of adequate continuing edu- cation and guidelines for paediatricians and vaccina- tion centre physicians and nurses; and provision of financial incentives for paediatricians and vaccina- tion centre staff (20). In addition, it may be helpful to carry out broad-based publicity campaigns that are aimed at emphasizing the benefits of vaccination and dispelling unrealistic concerns about the risks. It may also be helpful to set up an outreach programme sim- ilar to that in England and Wales, where mothers are counselled at antenatal visits about the benefits of vaccination and a visit is made to each mother short- ly after the birth of her child to encourage her to seek vaccination (20). Mothers' focus groups might prove useful for determining in greater depth why they are not having their children vaccinated and thus for developing new approaches and educational materi- als to deal with these perceptions. WHO Bulletin OMS. Vol 72 1994 893 Childhood vaccination coverage in Italy One of the secondary goals of our survey was to determine whether the existing surveillance system provides data of sufficient accuracy for monitoring coverage levels of the mandatory vaccines. In some of the regions where routine reporting was available, we found that the levels were similar to those obtained through our survey, implying that the sur- veillance data may be useful for monitoring vaccina- tion trends. In others, however, the surveillance system either overestimated or underestimated the coverage. If surveillance rather than periodic surveys is to be used for monitoring the use of mandatory vaccines, it may be useful to modify vaccination reporting forms to provide more accurate denomina- tors and to obtain information on whether the vac- cine schedule had been completed on time. Further- more, efforts are needed to improve both the timeliness and completeness of the system if it is to be used for routine monitoring, especially since only 11 of the 20 regions provided data for 1991 and even fewer have reported in 1992. For the optional vac- cines, periodic surveys are likely to remain necessary since many are administered in the private sector, and establishing a reliable reporting system would be difficult. Acknowledgements We thank the many physicians, nurses, and health assis- tants in the seven participating regions who served as interviewers during the survey. We also wish to thank D. Greco and S. Wassilak for their support and assistance in the preparation and review of this manuscript. Resume Couverture vaccinale chez 1'enfant en Italie: resultats d'une enquete realisee dans sept r6gions En Italie, il n'existe que peu de donnees sur la couverture vaccinale et le respect du calendrier de vaccination. C'est pourquoi nous avons r6alis6 des enquetes par sondage chez des enfants ages de 12 a 23 mois dans neuf zones d'6tude cou- vrant sept r6gions d'italie, en utilisant la methodo- logie normalis6e du Programme elargi de vaccina- tion. Les zones 6tudiees couvraient 40% de 1'en- semble des naissances vivantes recensees en Italie en 1991. Les taux de couverture pour la troisieme dose des anatoxines dipht6rique et t6tanique et pour le vaccin antipoliomy6litique oral, qui sont obligatoires, d6passaient 90% dans toutes les zones d'6tude sauf une. Toutefois, moins des deux tiers des enfants avaient requ la serie complete de vaccins avant l'age d'un an. La raison la plus fr6quente de la non-vaccination 6tait que l'enfant avait 6te malade et n'avait pas 6t6 conduit au centre de vaccination. Pour les deux vaccins facultatifs (contre la coqueluche et la rou- geole), le taux de couverture 6tait beaucoup plus faible, allant de 8% a 71% pour la coqueluche et de 9% a 53% pour la rougeole. Dans le cas du vaccin anticoquelucheux, la raison la plus souvent avancee par les meres pour la non-vaccination 6tait qu'on la leur avait d6conseill6e; pour la rou- geole, la raison la plus fr6quente etait que l'enfant 6tait malade et, 6galement, qu'on leur avait de- conseill6 la vaccination. D'apres les r6sultats des enquetes, la couverture pour les vaccins obliga- toires est forte, mais elle est tres faible pour les vaccins anticoquelucheux et antirougeoleux. II est n6cessaire d'am6liorer l'information des m6decins et des meres quant aux contre-indications vraies a la vaccination. De plus, en I'absence d'une le- gislation rendant obligatoire la vaccination contre la coqueluche et la rougeole, il faudra intensifier les efforts visant a convaincre les medecins et le public de ses avantages. References 1. Pozzi G. [The distribution of vaccines in Italy at the end of 1968]. Journal of the American Medical Association (Italian edition), 1990, 2: 329-330 (in Italian). 2. Salmaso S, Stazi MA, Luzi S. Immunization cover- age in Italy. Bulletin of the World Health Organiza- tion, 1987, 65: 841-846. 3. Grandolfo M, Santoro R. [Surveillance of poliomye- litis in Italy]. Notizario dell'lstituto Superiore di Sanita 1988, vol. 1 (No. 9): 1-4 (in Italian). 4. Binkin NJ et al. Epidemiology of pertussis in a developed country with low vaccination coverage: the Italian experience. Pediatric infectious disease journal, 1992, 11: 653-661. 5. Compendio statistico italiano. Edizione 1992. ISTAT, Rome, 1992 (in Italian). 6. Dean AG et al. Epi Info version 5.0: a word pro- cessing, database, and statistics program for epi- demiology on microcomputers. Stone Mountain, GA, USD, Inc., 1990. 7. Williams BC. Immunization coverage among pre- school children: the United States and selected European countries. Pediatrics, 1990, 36(suppl.): 1052-1055. 8. Hinman AR. Immunizations in the United States. Pediatrics, 1990, 36(suppl.): 1064-1066. 9. Statistiche della sanita anno 1985. Annuario n. 1. Edizione 1990. ISTAT, Rome, 1990 (in Italian). 894 WHO Bulletin OMS. Vol 72 1994 Italian Vaccine Coverage Survey Working Group 10. Statistiche della sanita anno 1986. Annuario n. 2. Edizione 1990. ISTAT, Rome, 1990 (in Italian). 11. Statistiche della sanita anno 1987. Annuario n. 3. Edizione 1990. ISTAT, Rome, 1990 (in Italian). 12. Statistiche della sanita anno 1988. Annuario n. 4. Edizione 1991. ISTAT, Rome, 1991 (in Italian). 13. Statistiche della sanita anno 1989. Annuario n. 5. Edizione 1992. ISTAT, Rome, 1992 (in Italian). 14. Hinman AR, Orenstein WA. Immunisation prac- tice in developed countries. Lancet, 1990, 335: 707-71 0. 15. Addis A et al. [Optional vaccinations for children: observational epidemiological study at the basic health care level]. Medico e bambino, 1993, 6: 404-408 (in Italian). 16. Hofmann P. That fine Italian hand. New York, Henry Holt & Co., 1990: 213-226. 17. Tozzi AE et al. [Antipertussis vaccination: knowl- edge, attitudes, and practice in paediatrics and among vaccination doctors]. Medico e bambino, 1994, 13: 34-38 (in Italian). 18. Grandolfo ME. [The programme to eliminate measles in Italy: a practical example of the efficacy of vaccine prevention]. In: Vierrucci A et al., eds. Le vaccinazioni in pediatria. Milan, CSH, 1991: 37-42 (in Italian). 19. [Handbook 1990-91: pharmacies]. Medico e paziente, 1990; 16(suppl.) (in Italian). 20. Goodwin S. Preventive care for children: immuni- zation in England and Wales. Pediatrics, 1990, 36(suppl.): 1056-1060. WHO Bulletin OMS. Vol 72 1994 895