Epidemiological impact of a nationwide measles immunization campaign in Viet Nam: a critical review

Abstract

Objective

To study the impact on measles case incidence of a nationwide measles immunization campaign in Viet Nam, while considering differences in disease surveillance before and after the campaign.

Methods

A nationwide mass immunization campaign was conducted in the north and south of Viet Nam in 2002 and 2003, respectively. During the campaign, a second vaccination dose was given to children aged 9 months to 9 years, and the reported coverage reached 99% in both zones. National measles case-based surveillance data collected during 2001–2006 were reviewed and analysed. Surveillance performance was assessed in terms of case investigation and specimen collection rates and reporting sensitivity for febrile rash cases. Fisher’s exact test was used to test for differences in indicator values before and after the campaign at the national and regional levels; the Wilcoxon signed-rank test was used at the provincial level.

Findings

Despite significant improvements in disease surveillance, a dramatic reduction in observed measles incidence was noted nationwide after the immunization campaign, with a drop in the national incidence of confirmed measles cases per 100 000 population from 5.44 in 2001 to 0.14 after the campaign (i.e. 2003 in the north and 2004 in the south; P < 0.001). Rapid measles resurgence was observed in 2005 and 2006 only in the north-western mountainous region of the country. The north did not show a statistically significant age shift for new cases (median age: 9 years in 2001 versus 8 years in 2003; P = 0.113), whereas the south did (median age: 7 years versus 12 years; P < 0.001).

Conclusion

A campaign approach for controlling measles in developing parts of Asia can prove effective. The swift re-emergence of disease in the north-western region was probably due to suboptimal coverage by the campaign and by the subsequent routine expanded programme on immunization in the north-western mountainous region.

Introduction

Measles is a childhood viral disease associated with a relatively high case-fatality rate, especially in malnourished populations in developing countries.¹ It also causes numerous complications.^2–4 The disease has been successfully controlled in the Americas, where the number of cases decreased dramatically from 2584 in 2002 to 105 in 2003. The incidence has been low ever since.⁵ The turning point was the introduction of nationwide mass measles immunization campaigns, usually called catch-up or knockout campaigns, and periodic follow-up campaigns. This supplementary strategy enabled the maintenance of high coverage by childhood measles immunization and the establishment of case-based surveillance involving a serological laboratory network.⁶

Following success in the Americas, the drive for measles control in other parts of the world was accelerated.⁷ In 2003, the World Health Assembly resolved to halve global measles mortality during 1999–2005 by fully implementing the joint strategic plan set forth by WHO and the United Nations Children’s Fund. The strategy includes strengthening routine immunization, mass immunization campaigns, case-based surveillance with laboratory confirmation and optimal care of infected children, including the administration of vitamin A.⁸ The goal set for 2005 was successfully achieved, with a 60% reduction in global mortality.⁹ In 2005, the WHO Regional Office for the Western Pacific set a regional goal for the elimination of measles, which was defined as terminating the circulation of domestic strains in the region, by 2012.¹⁰

Viet Nam adopted this regional goal and is aiming to achieve it by 2010. Despite a measles immunization coverage rate that has surpassed 93% since 1993 with a one-dose schedule, measles outbreaks have occurred every 7–8 years. This clearly illustrates the limitation of the single-dose approach in interrupting domestic circulation of the measles virus.¹¹ A nationwide mass measles immunization campaign was conducted mainly to provide a second dose of vaccine to children aged 9 months to 9 years. It was carried out in two phases: the north of the country was covered between March and April 2002 and the south between March and April 2003. The campaign involved massive social mobilization and included the participation of entities such as the defence ministry, the Red Cross, local government at all levels and the prime minister. Also, 169 906 health-care workers at all levels and 412 474 volunteers took part in this historic public health project. As a result, the reported coverage reached 6 684 980 out of 6 729 171 individuals (99.3%) in the north and 8 389 067 out of 8 466 868 (99.1%) in the south. This paper critically examines the epidemiological impact of this nationwide measles immunization campaign, while taking into account any changes in surveillance performance from before to after the campaign.

Methodology

In Viet Nam, measles has been designated a notifiable infectious disease since the 1970s. In 2001, a WHO-recommended measles case-based surveillance system was introduced. Initially it focused on investigating and reporting outbreaks of diseases associated with febrile rash; venous blood specimens were collected from a few early cases in order to confirm whether an outbreak was really due to measles. Since 2001, and especially after the immunization campaign, the system rapidly developed into a case-based system in which all suspected measles cases were reported and investigated and specimens were collected. A suspected measles case was defined in one of two ways: (a) any person with fever and a maculopapular (i.e. non-vesicular) rash accompanied by cough, coryza or conjunctivitis; (b) any person suspected by a clinician of having measles.

Blood specimens were usually centrifuged at a district hospital, and serum was sent to one of two national measles laboratories (the National Institute of Hygiene and Epidemiology in Hanoi or the Ho Chi Minh City Pasteur Institute in Ho Chi Minh City). Cases of measles were confirmed in one of three ways: a laboratory-confirmed case was one in which the patient tested positive for measles-specific immunoglobulin-M (IgM) antibodies on the Enzygnost Anti-Measles Virus IgM enzyme immunoassay (Dade Behring, Marburg, Germany); an epidemiologically confirmed case was one in which there was documented evidence of exposure to a confirmed measles case within the incubation period; and a clinically confirmed case was one that met the clinical case definition despite the absence of a blood test and a history of exposure. Suspected cases in which laboratory test results were negative were discarded as non-measles. The WHO recommends testing all such cases for rubella in countries and areas in which measles has almost been eliminated. The recommendation was slightly modified and followed in Viet Nam, as has been done in the Caribbean, and all suspected measles cases were tested for rubella-specific IgM.¹²

We analysed the 2001–2006 national measles surveillance data for Viet Nam. In assessing the campaign’s epidemiological impact, only measles cases confirmed by laboratory results, epidemiological linkage or clinical compatibility were included. A descriptive epidemiological analysis of confirmed cases was carried out to reveal monthly incidence trends and geographical and age distributions. Pre-campaign cases in the north were defined as those that occurred during 2001, whereas post-campaign cases were those that occurred during 2003. Similarly, 2002 and 2004 were considered the pre- and post-campaign periods in the south. We thus excluded cases that occurred during the campaign phases to avoid possible misclassification and to reflect the same seasons and durations of pre- and post-campaign periods in both north and south. In so doing, we circumvented distortions that could have been introduced into the analysis by underlying seasonal fluctuations in measles incidence. For the geographical distribution of confirmed measles cases (Fig. 1), the pre-campaign period was defined as 2001 for both north and south to illustrate the actual geographical variation in the annual incidence across the nation in a single pre-campaign year.

Fig. 1.

Incidence of confirmed measles cases in Viet Nam by province, before (2001) and (2003 in the north, 2004 in the south) after the mass measles immunization campaign^a

^a The nationwide mass immunization campaign was conducted in the north of Viet Nam in 2002 and in the south in 2003. 2001 data were analysed for before the campaign.

^b The horizontal line on the right-hand map indicates the division of the country into the north and south zones.

Surveillance performance indicators, namely the case investigation rate, the specimen collection rate, the proportion of silent provinces that reported no febrile rash cases in a year, and the number of reported cases discarded as non-measles per 100 000 population, were derived for before and after the campaign and compared. For each province, the Wilcoxon signed-rank test was used to test for differences between the medians of the indicators before and after the campaign; for the national and regional levels, Fisher’s exact test was used (Table 1).

Table 1. Comparison of performance indicators for measles case-based surveillance in Viet Nam before and after the mass measles immunization campaign.

Performance indicator by zone	Before the campaigna	After the campaignb	P-value from provincal analysisc	P-value from regional and nationwide analysesd
Case investigation ratee (%)
North	686 of 9 365 (7)	196 of 228 (86)	0.002	< 0.001
South	1418 of 4 736 (30)	435 of 696 (63)	0.036	< 0.001
Nationwide	2104 of 14 101 (15)	631 of 924 (68)	< 0.001	< 0.001
Specimen collection ratef (%)
North	317 of 9 365 (3)	157 of 228 (69)	< 0.001	< 0.001
South	746 of 4 736 (16)	392 of 696 (56)	< 0.001	< 0.001
Nationwide	1063 of 14 101 (8)	549 of 924 (59)	< 0.001	< 0.001
Febrile rash cases reported but discarded as non-measles per 100 000 populationg
North	0.20	0.41	0.112	NA
South	0.19	1.01	< 0.001	NA
Nationwide	0.20	0.74	< 0.001	NA
Proportion of silent provincesh (%)
North	1 of 28 (4)	9 of 28 (32)	NA	0.005
South	4 of 33 (12)	4 of 33 (12)	NA	1.000
Nationwide	5 of 61 (8)	13 of 61 (21)	NA	0.041

NA, not applicable.
^a 2001 in the north of Viet Nam and 2002 in the south of Viet Nam.
^b 2003 in the north of Viet Nam and 2004 in the south of Viet Nam.
^c P-value derived using the Wilcoxon signed-rank test for the difference between the median values of the indicators before and after the campaign in each province.
^d P-value derived using Fisher’s exact test for the difference in indicator values recorded before and after the campaign.
^e Proportion of cases investigated among the total number of suspected measles cases reported.
^f Proportion of cases for which a serum specimen was collected among the total number of suspected measles cases reported.
^g The figures do not include suspected cases that were reported but not classified because there was no case investigation.
^h Proportion of provinces that did not report even one suspected measles case.

Pre- and post-campaign periods in the north and south were defined in the same way as in the epidemiological analysis. In analysing the change in reporting sensitivity from before to after the campaign, only reported suspected measles cases that were eventually discarded as non-measles were included. This was done to eliminate the effect of the reduction in measles incidence due to the campaign and to ensure that pre- and post-campaign performances were comparable. Even though measles was controlled after the campaign, it is unlikely that the incidence of similar diseases such as rubella would have been affected. The estimated mid-year population was used to calculate measles incidence and the number of cases discarded as non-measles that were reported per 100 000 population during each year. The surveillance data set was originally maintained using Microsoft Access software (Microsoft Corp., Redmond WA, United States of America (USA)) and was analysed using SPSS 10.0J (SPSS Inc., Chicago IL, USA).

Results

Impact on monthly incidence

Fig. 2 and Fig. 3 depict the number of confirmed measles cases reported each month from 2001 to 2003 in the north and south of Viet Nam through the measles-case-based surveillance system before and after the immunization campaign. The seasonal peak in incidence occurred from January to May. The campaign targeted this period of high disease transmission and effectively reduced case incidence. In 2002 in the north, disease incidence was significantly reduced during and after the campaign, while in the south, the usual seasonal surge was noted but was reduced in 2003, and incidence remained low after the campaign. Nationally, confirmed measles cases dropped steadily and dramatically from 2001 to 2004: 3708, 1806, 855 and 95, respectively. Of the 95 confirmed measles cases observed in 2004, five were confirmed upon laboratory testing of 900 specimens, and 90 were confirmed on clinical grounds. Clearly, measles was very effectively suppressed in 2004.

Fig. 2.

Change in the number of confirmed measles cases reported^a in the north of Viet Nam before and after the mass measles immunization campaign^b

^a Cases reported through the measles case-based surveillance system.

^b The arrow indicates the starting date for the immunization campaign.

Fig. 3.

Change in the number of confirmed measles cases reported^a in the south of Viet Nam before and after the mass measles immunization campaign^b

^a Cases reported through the measles case-based surveillance system.

^b The arrow indicates the starting date for the immunization campaign.

Impact on geographical and age distributions

Nationally, the incidence of confirmed measles cases per 100 000 population decreased from 5.44 in 2001 to 0.14 after the campaign (P < 0.001). Fig. 1 shows the population-based incidence of confirmed measles cases in each of the 61 provinces in Viet Nam both before the campaign (2001) and after (i.e. 2003 in the north and 2004 in the south). In 2001, the incidence was greater than 10 per 100 000 in 12 provinces. After the campaign, most had an incidence of less than 1 per 100 000, proof that effective suppression of measles had been attained in each province. However, the mountainous north-west and the central highlands continued to have a higher incidence than other parts of the country after the campaign.

Fig. 4 and Fig. 5 show that measles in children aged under 10 years of age was more successfully reduced in the south than the north. After the campaign, the median age for the entire country shifted from 8 to 11 years (P < 0.001 by Mann–Whitney test). On analysing the north and south separately, the former did not show a statistically significant age shift (median age: 9 years in 2001 versus 8 years in 2003; P = 0.113), whereas the latter did (median age: 7 years versus 12 years; P < 0.001). There was also an incidence peak in children aged under 1 year both before and after the campaign, probably due to the neutralizing effect of transferred maternal antibodies on vaccine-induced seroconversion.

Fig. 4.

Age distribution of measles cases before and after the mass measles immunization campaign in the north of Viet Nam

Fig. 5.

Age distribution of measles cases before and after the mass measles immunization campaign in the south of Viet Nam

Change in surveillance performance

Both the case investigation rate and the specimen collection rate were significantly better, at the 95% confidence level, after the campaign at the provincial, regional and national levels (Table 1). The reporting sensitivity of non-measles febrile rash cases also increased 3.5-fold after the campaign. Using the figures for each province, the change in reporting sensitivity was statistically significant both nationally and in the south (P < 0.001 for both using the Wilcoxon signed-rank test). Of concern was the increased number of “silent” provinces seen in the north after the campaign. Eight of the 28 northern provinces became silent and, consequently, the proportion of silent provinces increased significantly (P = 0.041 using the χ² test). Thus, the reporting sensitivity varied between provinces against a background of nationwide improvement. In summary, although the proportion of silent provinces increased, in general the capacity of case-based surveillance to detect and confirm measles cases was significantly better after the campaign.

Long-term impact

Despite the success of the immunization campaign, local measles outbreaks were detected as early as 2005 in Lai Chau province and in Lao Cai province, both in the north-western mountainous region. In total, 293 cases were reported in Lai Chau province, of which 136 were laboratory confirmed and 157 were epidemiologically confirmed. Of these 293 cases, 150 (51%) were in individuals aged over 15 years. In Lao Cai province, 91 cases were reported; 15 were laboratory confirmed and 76 were epidemiologically confirmed. Of these 91 cases, 42 (46%) were in individuals aged over 15 years.

In 2006, there was a larger resurgence of measles involving 1978 confirmed cases, of which 1607 were either laboratory or epidemiologically confirmed. Of these, 1883 cases occurred in Dien Bien province, which was part of the former Lai Chau province in the north-western corner of the country bordering China until 2004, when the latter province was divided into two smaller provinces: Dien Bien and Lai Chau. Of the cases that occurred in Dien Ben, 153 were laboratory confirmed and 1355 were epidemiologically confirmed. Only 248 (14%) of these confirmed cases were in individuals aged over 15 years, while 1289 (71%) were in children aged under 10 years. In the same year, a second routine immunization, given at school entry, was introduced into the province. Throughout 2005 and 2006, the main focus of the resurgence was in the territory comprised by the former Lai Chau province. As can be seen in Fig. 1, the incidence in this area did not decrease much after the campaign, which suggests that the campaign was compromised locally. There was no death due to measles in either the 2005 or 2006 outbreak.

Fig. 6 shows the annual number of measles cases between 1986 and 2006. Two measles surveillance systems co-existed during 2001–2002: notifiable diseases surveillance and measles case-based surveillance. The former system confirmed more cases based on physicians’ diagnosis than the latter, which primarily confirmed cases either by laboratory tests or epidemiologically. Consequently, the former system may be considered less specific than the latter because of possible overdiagnosis by physicians. The peak incidence in 2006 was low. In 2007, there were only 17 confirmed measles cases. Cases mainly occurred in the north-western mountainous provinces and did not lead to a long-term increase in incidence.

Fig. 6.

Annual number of measles cases reported in Viet Nam between 1986 and 2006^a

^a In 2001, only the north of Viet Nam had implemented the measles case-based surveillance system.

Discussion

Findings and limitations

The incidence of measles in Viet Nam was significantly reduced after the 2002–2003 nationwide immunization campaign, particularly in 2004. Comparison of the performance of the surveillance system before and after the campaign clearly indicates that the system’s ability to detect and confirm measles cases improved. The reduced measles incidence observed is, therefore, genuine. However, without the swift introduction of a follow-up second immunization, the reduction in incidence was followed by the rapid resurgence of measles, even though routine second-dose immunization at school entry commenced after 2006.

A notable limitation of the study is the underrepresentation of measles cases that were not investigated and laboratory tested before the campaign. It is difficult to determine whether the numerous uninvestigated and untested cases were measles or not. These cases were clinically diagnosed by physicians and reported through the notifiable infectious disease surveillance system rather than the measles case-based surveillance system. Although the reporting sensitivity of non-measles febrile rash cases increased at the national, regional and provincial levels after the campaign, the analysis did not include those uninvestigated and untested cases, and this may have compromised the validity of both the sensitivity estimate and the epidemiological analysis for the pre-campaign period, especially in the north.

A global context

Between 2000 and 2003, 23 of the 45 countries with the highest measles mortality, including Viet Nam, implemented nationwide measles immunization campaigns. This undoubtedly contributed to the dramatic fall in measles mortality worldwide by 2005.^9,13 The campaigns in three west African nations (Burkina Faso, Mali and Togo) have been documented¹⁴ and the annual number of cases after the campaigns ranged from 333 in Togo to 1712 in Burkina Faso. A reduction in incidence also occurred after the 2002 campaign in Afghanistan¹⁵ but was smaller than reductions obtained in the Americas and Viet Nam.

The reduced incidence of measles following a campaign offers an opportunity to introduce a case-based surveillance system involving case investigation and specimen testing. The WHO Regional Office for the Western Pacific¹⁶ and the global consensus meeting on monitoring measles elimination¹⁷ both have as a criterion for elimination that the surveillance system should be sensitive enough to report at least one suspected measles case per 100 000 population per year in at least 80% of all districts. In Viet Nam, the reporting sensitivity was 0.74 per 100 000 after the campaign (Table 1), just below the level required to certify elimination.

Discussions with preventive medicine staff indicate that the increase in silent provinces in the north after the campaign was probably caused by bias and complacency. Bias was due to preventive medicine staff not reporting suspected measles cases because of a fear it would be interpreted as a failure of the immunization campaign and thus damage their reputation. Complacency was due to the belief that there would not be many febrile rash cases after the campaign because of the reduced measles incidence.

In the Region of the Americas, estimated reporting rates of suspected measles cases between 2003 and 2006 ranged from 3.55 to 4.34 per 100 000.^18–21 During the same period, confirmed measles cases in the region ranged from 75 to 187 per year (incidence rate per 100 000 population from 0.01 to 0.02), indicating near elimination. Compared to the America’s sensitive system, Viet Nam’s measles surveillance system was still suboptimal at the time of the study. However, the reporting rate reached 5.19 per 100 000 in 2006, by far the highest among Western Pacific countries.

In the Western Pacific, the Republic of Korea was the first country to declare measles elimination in 2006, with only 25 new cases (an incidence of 0.52 per 1 million population).²² However, the reporting rate of non-measles febrile-rash cases in the country was 0.18 per 100 000 population, indicating still suboptimal reporting sensitivity.²³ Some Pacific island countries have demonstrated limited measles virus transmission since national campaigns in 1997 and 1998, although sporadic outbreaks have been reported,¹¹ and the reporting rate of non-measles febrile-rash cases was 0.20 per 100 000 population.²³ In the Hong Kong Special Administrative Region of China, the incidence of measles was reduced to 0.9 per 100 000 population in 1998.²⁴ In 2006, it was 1.49 per 100 000 population.²³

Challenges faced after the campaign

A possible explanation of the rapid resurgence of measles in north-western Viet Nam during 2005–2006 is that older children remained vulnerable because the age group targeted for vaccination was narrower (i.e. 9 months to 9 years) than recommended (i.e. up to 14 years).²⁵ The change in age distribution after the campaign observed in the south (Fig. 5) and in outbreaks in Lai Chau province and Lao Cai province in 2005 suggest that the virus may continue to circulate among teenagers. Conversely, children aged under 10 years were mainly affected in the larger 2006 outbreak focused on Dien Bien province, which instead indicates suboptimal immunization coverage. Also, 474 of the 1508 confirmed cases (31%) in the Dien Bien outbreak occurred in children aged under 1 year, pointing to suboptimal coverage by the expanded programme on immunization since 2003. Accordingly, the resurgence was probably due to suboptimal coverage of both the immunization campaign and, thereafter, the expanded programme on immunization. This implies that increasing the upper age limit from 10 to 15 years would not necessarily be effective and would certainly not be cost-effective.

An important strategic question is how to maintain the population immunity attained in 2004 given the variation in immunity across the country. Viet Nam has the advantage that coverage of the routine first dose of measles vaccination is generally high: 98% and 94% in the north and the south, respectively, before the campaign, and 97% and 96%, respectively, after. This background no doubt contributed to the very effective suppression of measles after the campaign. In 2006, the country introduced the second routine immunization dose at school entry in 43 out of 64 provinces, with a coverage exceeding 98%.

The county is reluctant to implement nationwide follow-up campaigns mainly because of concern about the cost. In the nationwide catch-up campaign described here, US$ 2.0 million was spent on operational costs in addition to the cost of the necessary vaccines, syringes and safety boxes provided by Japan. This expenditure was unprecedented for a mass public health intervention in the country. A follow-up campaign targeting 1- to 4-year-olds will require nearly half the money spent on the catch-up campaign. In November and December 2007, an immunization campaign was carried out in 17 mountainous provinces in the north targeting 6- to 20-year olds (1- to 20-year-olds in four provinces with the highest risk) in response to the 2005 and 2006 outbreaks.

Viet Nam is aiming to achieve the level of immunity required to eliminate measles by maintaining a high coverage of routine first vaccinations in infants, routine second vaccinations at school entry and supplementary local campaigns in high-risk districts. The effectiveness of this strategy should be closely monitored since the country is a rare example of a tropical developing country outside the Americas that is pursuing measles elimination following a catch-up campaign, which this critical epidemiological review found to be very effective. ■