Assessment of Delay in Age-appropriate Vaccination Using Survival Analysis

Abstract

Assessment of delay in age-appropriate vaccination provides more information about timeliness of vaccination than up-to-date vaccination coverage. The authors applied survival analysis methods to data from a vaccination coverage survey among children aged 13–59 months conducted in Argentina in 2002. By age 19 months, 43% of children (95% confidence interval (CI): 40, 46) were vaccinated with the fourth dose of diphtheria, tetanus, and pertussis (DTP4). By age 13 months, 55% of children (95% CI: 52, 57) were vaccinated with measles-containing vaccine. By age 7 months, 33% of children (95% CI: 27, 40) were vaccinated with the third dose of hepatitis B. Compared with firstborn children, third children were more likely to be delayed for DTP4 (relative risk (RR) = 1.41, 95% CI: 1.22, 1.62), measles-containing vaccine (RR = 1.54, 95% CI: 1.32, 1.78), and the third dose of hepatitis B (RR = 1.31, 95% CI: 1.03, 1.67). Children whose caregivers had completed secondary school were less likely to be delayed for DTP4 (RR = 0.68, 95% CI: 0.52, 0.90) compared with those whose caregivers had not completed primary school. Survival analysis methods were helpful in measuring vaccine uptake and should be considered in future surveys when assessing delay in age-appropriate vaccination.

health surveys, immunization, preventive health services, survival analysis

CI, confidence interval, DTP, diphtheria, tetanus, and pertussis, DTP4, fourth dose of diphtheria, tetanus, and pertussis, HepB3, third dose of hepatitis B, MCV, measles-containing vaccine, RR, relative risk

Vaccination coverage among children is a basic indicator of vaccination program performance and a measure of compliance with preventive medicine guidelines. Ideally, children receive the scheduled vaccinations as early as possible according to the vaccination schedule, resulting in high age-appropriate vaccination coverage rates. In the United States, Healthy People 2010 sets a national age-appropriate vaccination target of 80 percent for children aged 19–35 months (1). However, a national study published in 2002 found that only 18 percent of children received all vaccinations by the recommended ages (2). Moreover, a US study reported in 2005 that more than one in three children were undervaccinated for more than 6 months during their first 24 months of life (3).

Monitoring age-appropriate vaccination status in the population is of critical importance for establishing disease risk in the population, particularly for those diseases in which age is related to severity or higher complication rates. For example, in the United States, pertussis morbidity and mortality occur primarily among infants (4, 5). Despite high reported vaccination coverage rates, approximately 40 percent of the infants who contracted pertussis during the 1990s in the United States were undervaccinated for their age (4).

A commonly used measure of population vaccination coverage is up-to-date vaccination, which is the proportion of children vaccinated at a certain age (6–9). However, up-to-date vaccination determines only the coverage at specific ages and does not allow measurement of delay in age-appropriate vaccinations (10). Therefore, some children may be considered up-to-date, even if one or more vaccine doses were administered later than recommended. Age-appropriate vaccination can be assessed by determining the age at vaccination for certain vaccine doses, and the measure of delay may be categorized in months (11). However, the age when defined coverage levels are attained is not typically measured. The Kaplan-Meier method is a technique used to describe time-to-event data. This methodology can be useful to assess age-appropriate vaccination. Although this methodology has been used recently (12), it has not been widely applied.

During 2002, a vaccination coverage survey among children aged 13–59 months was conducted in Buenos Aires, the capital of Argentina, with a population of approximately 2.8 million (13). The objectives of the present study were to estimate vaccination uptake (i.e., age-dependent vaccination coverage) by applying survival analysis methods for three vaccine doses included in the routine immunization schedule and to identify sociodemographic factors associated with delayed vaccination.

MATERIALS AND METHODS

We used data from a population-based survey of children aged 13–59 months residing in Buenos Aires that was carried out between March 8 and April 20 of 2002 (14). Census tracts were randomly selected with probability proportional to the size of the census tract (15). The number of interviews for each selected census tract was intended to represent all geographic areas and socioeconomic levels in the city. When census tracts comprised more than one block, census blocks were randomly selected. Face-to-face interviews were conducted with principal caregivers of children living in the selected blocks at the time of the survey. The interviewers started from a preselected point on the selected block and proceeded counterclockwise around this block and, if necessary, around the next nearest blocks, until the required number of interviews was reached. Only one child per family was selected to avoid clustering. When two or more eligible children were present in the same household, the youngest child was selected.

A total of 1,391 children aged 13–59 months were enrolled. The data collected included the following: 1) child sociodemographics, including the date of birth, sex, country of birth, ethnicity (assessed by the country of birth of his/her grandparents), number and age of siblings, child-care attendance, health insurance, and type of immunization provider used (i.e., public, private, or both); and 2) caregiver sociodemographics, including age, marital status, work, and education.

Outcome measures

For the purpose of the study, we selected three vaccines for the analyses: diphtheria, tetanus, and pertussis (DTP), because it is a long-established vaccine with several doses scheduled in the first 2 years of life; measles-containing vaccine (MCV), also a long-established vaccine, but with only one dose recommended at 1 year of age; and hepatitis B vaccine, a vaccine recently introduced with several doses scheduled in the first year of age. Because delay between the recommended age at vaccination and vaccine administration seems to increase with the number of doses in the series (16–18), we examined receipt of the fourth dose of DPT (DTP4) and the third dose of hepatitis B (HepB3).

The recommended childhood National Immunization Schedule in Argentina was used to assess delay in vaccination (19). The National Immunization Schedule for children aged less than 5 years comprises Bacillus Calmette-Guérin at birth; DTP and Haemophilus influenzae type b at 2, 4, 6, and 18 months; measles, mumps, and rubella at 12 months; and hepatitis B vaccine at birth and 2 and 6 months. The most recently introduced vaccines are H. influenzae type b in 1997, a second dose of measles, mumps, and rubella in 1998, and hepatitis B vaccine in November 2000. Delayed vaccination was defined as not having received DTP4 by 19 months, MCV by 13 months, or HepB3 by 7 months.

Only written documentation (i.e., vaccination card or any other written proof of vaccination) was accepted to abstract information on vaccine doses and vaccination dates. Although children who receive an invalid dose must receive an additional dose to be considered adequately immunized, in Argentina they are usually not perceived as unvaccinated because vaccination status is typically assessed by counting the number of doses received, regardless of the vaccination dates. Since this is a distinct category of no vaccination, children with invalid doses were excluded from the analysis. Doses were defined as valid according to the Centers for Disease Control and Prevention criteria for minimum ages and intervals between doses (20). Vaccine doses administered 4 days or less before the minimum age and/or 4 days or less below the minimum interval were counted as valid following the recommendations of the Advisory Committee on Immunization Practices (20). Only DTP doses given after 6 weeks of age were considered valid, and a minimum interval to the next dose of 4 weeks was required for the second and third doses of DTP. For DTP4, a minimum age of 12 months and an interval of 4 months or more after the previous dose were required. Because most health insurance systems in Buenos Aires cover vaccines during the first year of life, children usually receive their MCV between 11 and 12 months of age. Therefore, measles vaccination was considered valid only if any MCV had been administered when the child was more than 11 months of age. Since hepatitis B was introduced into the National Immunization Schedule in November 2000, only children born after that date were assumed eligible to receive three doses of hepatitis B vaccine. The first hepatitis B vaccine dose is recommended at birth in Argentina, so any dose after birth was considered valid; however, a minimum interval of 4 weeks was required between the first and second doses. A minimum age of 6 months and an interval of 8 weeks or more after the second dose and 16 weeks after the first dose were required for HepB3 doses to be counted as valid (20).

Statistical analysis

For analysis purposes, months were considered to have 30.5 days. Therefore, the DTP4 analyses were restricted to 1,012 children who were at least 580 days (19 months) of age by March 8, 2002 (first day of survey), and the MCV analyses were restricted to 1,330 children who were at least 397 days (13 months) of age by March 8, 2002. The HepB3 analyses were restricted to 233 children who were born after November 1, 2000 (when hepatitis B was introduced into the National Immunization Schedule). After exclusion of children with invalid doses, the final sample sizes were 969 children aged 19–59 months for the age-appropriate analyses for DTP4, 1,330 children aged 13–59 months for MCV, and 216 children aged 13–16 months for HepB3.

For each vaccine, vaccination uptake over time (i.e., coverage by age) was estimated by the Kaplan-Meier method with age as the timescale (21). Vaccination coverage at age t was estimated by 1 − S_KM(t), the Kaplan-Meier survival function; 1 − S_KM(t) is the cumulative probability of being vaccinated by age t. The number of days of delay in vaccination was based on the maximum age at which the vaccine was recommended (19). For example, for DTP, the first day of delay was defined as the first day after 19 months.

The association between each characteristic and delayed vaccination was evaluated by estimating relative risks for delayed vaccination in a log-binomial regression analysis (22) using the SAS procedure GENMOD (23). Characteristics that were statistically significant in the bivariate analyses were included in a multivariable log-binomial regression model. All analyses utilized SAS statistical software, release 8.02 (23), and statistical tests were two tailed with α = 0.05.

RESULTS

Of the 969 children in the DTP4 analysis, 97 percent were born in Argentina, 61 percent received DTP4 only in the public sector, and 40 percent had no health insurance. Less than half of the children (44 percent) were the first child in their family. The majority of caregivers (78 percent) were married or living together, and 35 percent had not completed secondary school. The distributions of characteristics for the MCV and HepB3 analyses were similar to those for the DTP4 analysis (table 1).

TABLE 1.

Bivariate analysis of delayed vaccination for DTP4,* first dose of MCV,* and HepB3,* by characteristics, Buenos Aires, Argentina, 2002†

Characteristic	DTP4							MCV							HepB3
		Total	Delayed			p value‡	Total	Delayed			p value	Total	Delayed			p value
				No.	%					No.			%				No.	%
Child's age (months)
13–18	N/A*				316	132	42	0.33		N/A
19–24	259	139	54	0.08	268	116	43
25–35	288	160	56		300	144	48
36–47	266	169	64		277	136	49
48–59	156	84	54		169	75	44
Sex
Male	548	316	58	0.62	745	352	47	0.11	107	69	65	0.33
Female	421	236	56		585	251	43		109	77	71
Country of birth
Argentina	938	536	57	0.59	1,278	579	45	0.61	203	135	67	0.38
Other countries	6	4	67		10	6	60		1	1	100
Unknown	25	12	48		42	18	43		12	10	83
Ethnicity
≥2 grandparents born in Argentina	808	462	57	0.99	1,096	497	45	0.83	167	114	68	0.57
≥3 grandparents born in Latin America	110	62	56		155	73	47		27	16	59
Other	22	12	55		34	13	38		11	7	64
Unknown	29	16	55		45	20	44		11	9	82
Position in the family
First child	423	197	47	<0.01	571	206	36	<0.01	97	55	57	0.02
Second child	266	165	62		369	184	50		63	48	76
Third child	218	152	70		297	169	57		39	29	74
Unknown	62	38	61		93	44	47		17	14	82
Child-care attendance
No attendance	396	232	59	0.19	670	298	45	0.75	192	125	65	0.14
Public	300	179	60		336	160	48		5	5	100
Private	243	127	52		274	121	44		5	4	80
Unknown	30	14	47		50	24	48		14	12	86
Provider of vaccine
Public only	593	360	61	0.03	820	391	48	0.19	141	91	65	0.54
Private only	217	109	50		286	118	41		42	30	71
Both public and private	135	71	53		182	76	42		22	16	73
Unknown	24	12	50		42	18	43		11	9	82
Health insurance
No health insurance	387	251	65	<0.01	545	280	51	<0.01	100	67	67	0.34
Union insurance	393	205	52		527	215	41		74	46	62
Private prepaid medicine	156	79	51		204	84	41		29	23	79
Unknown	33	17	52		54	24	44		13	10	77
Principal caregiver age (years)
≤25	154	95	62	0.37	233	113	49	0.52	58	35	60	0.41
26–35	451	245	54		610	268	44		92	67	73
≥36	270	159	59		352	156	44		45	29	64
Unknown	94	53	56		135	66	49		21	15	71
Education
No education/incomplete primary	18	16	89	<0.01	32	14	44	0.01	7	4	57	0.55
Primary complete	320	197	62		435	208	48		67	44	66
Secondary complete	345	177	51		466	217	47		74	47	64
Tertiary/university complete	187	107	57		250	89	36		44	34	77
Unknown	99	55	56		147	75	51		24	17	71
Employment
Not working	529	304	58	0.69	727	331	46	0.91	139	93	67	0.89
<40 hours/week	188	100	53		250	112	45		23	17	74
≥40 hours/week	153	90	59		208	91	44		31	20	65
Unknown	99	58	59		145	69	48		23	16	70
Marital status
Married/living together	759	424	56	0.26	1,034	458	44	0.29	167	108	65	0.21
Not living together	101	65	64		139	71	51		26	21	81
Unknown	109	63	58		157	74	47		23	17	74
Zone
Northern	301	172	57	0.06	415	189	46	0.55	63	48	76	<0.01
Central	334	175	52		447	194	43		78	42	54
Southern	334	205	61		468	220	47		75	56	75
Lateness of dose 1
Not late	835	437	52	<0.01		N/A			122	55	45	<0.01
Late	134	115	86						94	91	97

*

DTP4, fourth dose of any diphtheria and tetanus toxoids and pertussis vaccines, where children were born between March 1997 and August 2000; MCV, measles-containing vaccine, where children were born between March 1997 and February 2001; HepB3, third dose of hepatitis B vaccine, where children were born between November 2000 and March 2001; N/A, not applicable.

†

This survey was conducted between March 8, 2002, and April 20, 2002.

‡

Two-sided p values for the null hypothesis of no difference among percentages delayed across subgroups of the characteristic.

Vaccination uptake

At the interview, 81, 91, and 67 percent of children were up-to-date for DTP4, MCV, and HepB3, respectively (combining vaccinated in figure 1). In contrast, only 43, 55, and 33 percent were age-appropriately vaccinated for DTP4, MCV, and HepB3, respectively (figure 1). Vaccination uptake for the study vaccine doses using the Kaplan-Meier methodology is shown in figures 2, 3, and 4. Reference lines were drawn in the figures at the age recommended for each vaccine. By age 19 months, 43 percent of children (95 percent confidence interval (CI): 40, 46) were vaccinated with DTP4. By age 13 months, 55 percent of children (95 percent CI: 52, 57) were vaccinated with MCV. By age 7 months, 33 percent of children (95 percent CI: 27, 40) were vaccinated with the third dose of hepatitis B. Using 85 percent coverage as a benchmark and reading from figure 2, we found that there was an 18-month delay between the recommended age of 19 months and the observed age of 37 months for DTP4 vaccine. For MCV, the uptake was steeper, and the 85 percent benchmark coverage was reached by 18 months, or about 6 months after the recommended age (figure 3). HepB3 vaccination uptake was slower and did not reach the 85 percent benchmark by age 16 months (figure 3); however, about 70 percent coverage was attained by 15 months of age, or 9 months after the recommended age (figure 4).

FIGURE 1.

Vaccination status for the fourth dose of diphtheria and tetanus toxoids and pertussis vaccines (DTP4), first dose of measles-containing vaccine (MCV), and third dose of hepatitis B vaccine (HepB3), Buenos Aires, Argentina, 2002. Children in the DTP4 analysis were born between March 1997 and August 2000, in the MCV analysis between March 1997 and February 2001, and in the HepB3 analysis between November 2000 and March 2001.

FIGURE 2.

Proportion of 969 children aged 19–59 months vaccinated with the fourth dose of diphtheria, tetanus, and pertussis vaccine at each age, Buenos Aires, Argentina, 2002. Inverse Kaplan-Meier curve with pointwise 95% confidence intervals. A reference line is drawn at 19 months. Children in this analysis were born between March 1997 and August 2000.

FIGURE 3.

Proportion of 1,330 children aged 13–59 months vaccinated with the first dose of measles-containing vaccine at each age, Buenos Aires, Argentina, 2002. Inverse Kaplan-Meier curve with pointwise 95% confidence intervals. A reference line is drawn at 13 months. Children in this analysis were born between March 1997 and February 2001.

FIGURE 4.

Proportion of 216 children aged 13–17 months vaccinated with the third dose of hepatitis B vaccine at each age, Buenos Aires, Argentina, 2002. Inverse Kaplan-Meier curve with pointwise 95% confidence intervals. A reference line is drawn at 7 months. Children in this analysis were born between November 2000 and March 2001.

Factors associated with lack of age-appropriate vaccination

Characteristics associated with delayed vaccination for each of the three vaccine doses are presented in table 1 without adjustment for other measured variables. All characteristics that were statistically significantly associated in the crude analysis were retained in a multivariable relative risk model, except for having a late first dose of a given vaccine. This variable was excluded because children receiving a late first dose were at higher risk for being late on their last dose. Therefore, the inclusion of this variable in the multivariable analyses could overadjust, particularly because the other independent variables are time invariant. After controlling for other characteristics, we found that not being the first child of the family remained significantly associated with delayed vaccination for the three vaccines. For instance, when the first child was used as the referent group, being the third child was significantly associated with delayed vaccination for DTP4 (relative risk (RR) = 1.41, 95 percent CI: 1.22, 1.62), MCV (RR = 1.54, 95 percent CI: 1.32, 1.78), and HepB3 (RR = 1.31, 95 percent CI: 1.03, 1.67) (table 2). Having union health insurance was associated with a somewhat lower risk of delayed vaccination for DTP4 (RR = 0.86, 95 percent CI: 0.75, 0.97) and MCV (RR = 0.82, 95 percent CI: 0.72, 0.93). Children whose caregivers had completed secondary school were less likely to be delayed for DTP4 (RR = 0.68, 95 percent CI: 0.52, 0.90) compared with those whose caregivers had not completed primary school. In addition, living in the central zone of the city was associated with a reduction in delayed vaccination for HepB3 (RR = 0.71, 95 percent CI: 0.56, 0.90). Further adjustment for a late first dose did not appreciably alter the results for DTP4; however, not being the first child and not living in the central zone of the city were not significantly associated with delayed vaccination for HepB3.

TABLE 2.

Multivariable analysis of delayed vaccination for DTP4,* first dose of MCV,* and HepB3,* by selected characteristics,† Buenos Aires, Argentina, 2002‡

Characteristic	DTP4			MCV			HepB3
		Relative risk	95% confidence interval	Relative risk	95% confidence interval	Relative risk	95% confidence interval
Position in the family
First child	1	Referent	1	Referent	1	Referent
Second child	1.32	1.15, 1.51	1.40	1.20, 1.62	1.34	1.08, 1.65
Third child	1.41	1.22, 1.62	1.54	1.32, 1.78	1.31	1.03, 1.67
Unknown	1.49	1.17, 1.90	1.30	0.97, 1.74	1.29	0.95, 1.75
Provider of vaccine
Public only	1	Referent
Private only	0.94	0.79, 1.12
Both public and private	0.90	0.75, 1.07
Unknown	0.70	0.33, 1.49
Health insurance
No health insurance	1	Referent	1	Referent
Union insurance	0.86	0.75, 0.97	0.82	0.72, 0.93
Private prepaid medicine	0.86	0.70, 1.06	0.91	0.75, 1.09
Unknown	0.86	0.46, 1.59	0.72	0.48, 1.08
Education
No education/incomplete primary	1	Referent	1	Referent
Primary complete	0.74	0.57, 0.96	1.13	0.77, 1.66
Secondary complete	0.68	0.52, 0.90	1.18	0.81, 1.73
Tertiary/university complete	0.81	0.60, 1.08	0.95	0.63, 1.43
Unknown	0.78	0.57, 1.07	1.4	0.93, 2.11
Zone
Northern					1	Referent
Central					0.71	0.56, 0.90
Southern					0.96	0.78, 1.17

*

DTP4, fourth dose of any diphtheria and tetanus toxoids and pertussis vaccines, where children were born between March 1997 and August 2000; MCV, measles-containing vaccine, where children were born between March 1997 and February 2001; HepB3, third dose of hepatitis B vaccine, where children were born between November 2000 and March 2001.

†

Only significant characteristics in the bivariate analysis shown in table 1, except for having received the first dose late, are included in the multivariable analysis.

‡

This survey was conducted between March 8, 2002, and April 20, 2002. Entries without values were not included in the multivariable model.

DISCUSSION

As demonstrated here, assessments of up-to-date vaccination provide coverage rates that can be considerably higher than those obtained by age-appropriate vaccination status. The Kaplan-Meier method provided a useful way to visualize the increase in vaccination uptake over time and allowed an estimate of the proportion of children immunized at any given age. Significantly higher rates of age-appropriate vaccination were found among firstborn children for DTP4 and MCV.

Measuring vaccine coverage at the community level is an important strategy to improve coverage rates (24, 25). However, a major objective of public and private sector immunization programs is not only to increase coverage rates but also to achieve the highest levels of protection against preventable diseases at the youngest possible age (1). Therefore, assessment of timeliness of vaccine receipt provides additional valuable information even in programs with high up-to-date vaccination coverage. Differences between up-to date and age-appropriate vaccination, similar to those described in this study, were found in previous studies (11, 26), especially when minimum age and dose interval criteria were used (27). Because children with invalid doses were excluded from our study, all the differences between up-to-date and age appropriate were due to vaccination delay.

The Kaplan-Meier method applied to this coverage survey allows easy estimation of the percentage of the study population falling within any category of vaccination delay by subtraction of two points on the curve. The inverse Kaplan-Meier curves were useful to determine the age at which a certain vaccination coverage rate was reached in a population, which is required when assessing the performance of the program in reaching target coverage rates. We found important delays in vaccination with the analyzed doses of vaccine. Over 40 percent of children received DTP4 or MCV vaccine by the recommended age, while only 30 percent were vaccinated on schedule with HepB3. Although low, the percentage of children vaccinated on time with DTP4 in Argentina seems to be higher than the proportion of children receiving the booster dose within the recommended schedule in Germany (12). In addition, this methodology is useful to visualize how schedules are implemented in an area. In our study, a steeper vaccination uptake was observed for long-established vaccines, such as MCV or DTP4, than for hepatitis B, which was more recently incorporated into the vaccination schedule. Surveillance of delay trends over time may be an accurate and useful tool to assess changes in vaccination status of the population that would not be revealed through the traditional vaccination coverage methodologies. This method could let program managers estimate the potential burden of disease based on vaccination coverage at any age and do interventions in a timely manner.

Identification of characteristics associated with delayed vaccination can be helpful when developing strategies to improve age-appropriate vaccination coverage. In our study, children who were late in beginning their vaccinations tended not to complete the recommended immunizations on time. These results are similar to those previously reported in other settings (8, 28, 29). Some factors related to low socioeconomic status (e.g., low parental education or annual income below the poverty level) were found to be associated with an increased risk of vaccination delay in the United States (11, 30). Our coverage survey did not measure poverty; however, some variables such as lack of health insurance or lower education of the caregiver, which may be surrogates of low socioeconomic status, were associated with vaccination delay. These factors may be related to lower access to health-care services. A low educational level of the caregiver and not being the firstborn were also associated with incomplete immunization in previous studies (31, 32).

The results of this study are subject to limitations. Only children with vaccination cards were included in the survey. Having a vaccination card may be associated with a greater probability of being vaccinated (33, 34). A retrospective analysis of US children in the National Health Interview Survey showed that children with vaccination records tended to be White race/ethnicity, from families not living in poverty, covered by insurance, and with two parents. Moreover, all these characteristics were associated with reduced risk of vaccination delay (30). This likely resulted in an underestimate of vaccination delay in our study. However, our choice to count only vaccine doses that were documented on a vaccination card was the only way to obtain accurate vaccination dates. Although this was the most accurate method to obtain vaccination dates, parental cards can be incomplete (9), and some dates may have not been written in the vaccination cards, resulting in an overestimate of the prevalence of vaccination delay. Children who died or moved were not included in the survey, and they might have been vaccinated before death or moving; hence, there may be a potential bias in estimates of the cumulative proportion of vaccination. However, the study intended to estimate coverage rates among children living in the area at the time of the survey, so information on children who had died or moved was not collected.

Our findings provide the proportion of the age-specific population protected by vaccination. This proportion is essential information for public health policy makers, as well as for health-care providers trying to identify subgroups at risk for delayed vaccination and to set priorities for immunization efforts. Thus, from the public health point of view, survival analysis methods can be used to monitor more precisely person-time at risk for disease and should be considered in assessing delay in age-appropriate vaccination in other sources of information about immunization coverage, such as the US National Immunization Survey (35). Estimating the proportion of the population vaccinated at a certain age, which can be easily done through use of Kaplan-Meier curves, is of particular relevance when the achievement of national vaccination goals requires a more refined mechanism to determine population immunity against disease. In addition, information provided by this methodology can be used by health-care providers to reduce missed opportunities for vaccination.

In summary, survival analysis techniques applied to a vaccination coverage survey were useful to measure vaccine uptake and provided clinically and epidemiologically relevant information regarding timeliness of vaccination and person-time at risk for vaccine-preventable diseases. Survival analysis methods should be considered for analysis of data from coverage surveys when assessing delay in age-appropriate vaccination.

This study was funded by the surveillance and disease control program (VIGI+A) of the Ministry of Health in Argentina.

The authors thank Dr. Zulma Ortíz and Dr. Hugo Fernández for their invaluable support in this study. They also thank John Stevenson, Vance Dietz, Phil Rhodes, and Lance Rodewald for their fruitful discussion of the statistical analysis of the data.

Conflict of interest: none declared.