a Epidemic Intelligence Service, Epidemiology Program Office,
b National Immunization Program, Centers for Disease Control and Prevention, Atlanta, USA.
c Formerly, Expanded Programme on Immunisation, Department of Health, Pretoria, South Africa.
d Current affiliation: World Health Organization Office for East Africa, Nairobi, Kenya.
e Communicable Disease Control, Mpumalanga Department of Health, Nelspruit, South Africa.
f Department of Family Medicine and Primary Health Care, Pretoria, South Africa.
g Communicable Disease Control, Western Cape Department of Health, Cape Town, South Africa.
h World Health Organization Regional Office for Africa, Harare, Zimbabwe.
Dr Amra Uzicanin, National Immunization Program, Mailstop E-05, Centers for Disease Control and Prevention, 1600 Clifton Road, Atlanta, GA 30333, USA. E-mail: auzicanin{at}cdc.gov
Abstract
Background In South Africa, as part of an effort to eliminate indigenous measles by 2002, vaccination campaigns were conducted in 19961997 targeting all children aged 9 months to 14 years; coverage was estimated at 85%. The impact of the campaigns on measles disease burden was evaluated in 1999.
Methods We analysed routine measles surveillance data and undertook a retrospective review of hospital registers in two of South Africas nine provinces.
Results In Mpumalanga in the pre-campaign years (19921996), 4498 measles cases and 6 deaths were reported; 182 cases and no deaths were reported in 19971998. Hospital registers showed 1647 measles hospitalizations and 11 deaths in the pre-campaign period, and 60 hospitalizations and no deaths after the campaign (1997April 1999). In Western Cape in pre-campaign years (19921997), 5164 measles cases and 19 deaths were reported; 132 cases and no deaths were reported in 1998. Hospital registers showed 736 measles hospitalizations and 23 deaths in the pre-campaign period, and 29 measles hospitalizations and no deaths post-campaign (1998July 1999).
Conclusions Study findings indicate that reported measles cases, measles-related hospitalizations and deaths were considerably reduced in both provinces after the campaign compared with the pre-campaign period. Longer observation is needed to evaluate the long-term impact of the campaigns.
Keywords Measles, vaccination, South Africa
Accepted 13 June 2002
In South Africa, routine administration of measles vaccine to infants was first introduced in 1975; since then, various vaccination schedules and vaccine types have been used.1,2 The current schedule was introduced in 1995 and includes two doses of measles vaccine (Schwartz strain) administered to children aged 9 months (first dose) and 18 months (second dose). In 1994, the first national vaccination coverage survey estimated that 85% of children aged 1223 months had received at least one dose of measles vaccine; the coverage in the nine provinces of South Africa ranged from 72% to 95%.3
Measles has been a notifiable disease in South Africa since 1979. During the 1980s, 15 000 to 20 000 measles cases and 250 to 500 deaths were reported each year (Figure 1). During the 1990s, measles remained endemic and epidemics continued to occur periodically, but the case fatality ratio sharply declined at the beginning of the decade and remained at low levels. In contrast to the 1980s when 60% to 70% of reported measles cases were among children aged <5 years, most cases reported during the 1990s occurred among school-aged children and young adults.
|
In 19961997, measles vaccination campaigns were implemented in all nine provinces of South Africa targeting all children aged 9 months14 years (Table 1). Overall campaign coverage was estimated at 85%.4
|
Materials and Methods
Study provinces
Mpumalanga is a predominantly rural province in the northeast of South Africa. The population is 2.8 million; Blacks/Africans represent the largest ethnic group (89%).5 In 1994, the province incorporated a number of densely populated areas formerly designated as homelands for Blacks/Africans during the former apartheid regime. These areas were historically neglected in terms of economic and infrastructure development, and most of their residents still live in poverty. In the 1994 national survey, routine measles vaccination coverage among children aged 1223 months in Mpumalanga was 79%.3
Western Cape Province is urbanized with good infrastructure. The provinces population of 3.9 million consists of diverse ethnic groups, with South African Mixed race (in South African official statistics referred to as Coloureds) being the largest (54%), followed by Blacks/Africans and Whites (each 21%).5 Most of the population resides in the provincial capital of Cape Town and in surrounding densely populated townships. Routine measles vaccination coverage is the highest in the country and was estimated by the 1994 national survey at 95%.3
Together, these two provinces comprise approximately 17% of South Africas population of 41 million.
Measles surveillance
From 1979 to 1998, measles surveillance in South Africa involved reporting of physician-diagnosed measles cases and deaths on a standard notification form to the National Department of Health in Pretoria through the respective Provincial Departments of Health. Reported data include the International Classification of Diseases, Ninth Revision (ICD-9) code of the disease (055 for measles), age, sex, race, date of disease onset, magisterial district, province, and outcome (alive/died). The new measles surveillance system with requirements for epidemiological investigation and laboratory confirmation of suspected measles cases was implemented in late 1998, but it was not fully functional throughout the nation at the time of this study.
Hospital record review
The Provincial Departments of Health in Mpumalanga and Western Cape identified the acute care hospitals in which patients with measles could have been cared for during 19921998. A questionnaire was sent by fax to these hospitals to request their participation in the study and to obtain baseline information about the availability of hospital records for the study period, annual number of measles hospitalizations, and hospital policy on admitting suspected measles cases. Hospitals that had a policy not to admit suspected measles cases and those without hospital registers available for review were excluded from the study. Measles-related hospitalization was defined as a patient diagnosed with measles on admission and/or on discharge from 1 January 1992 through the date of the hospital record review (MarchApril 1999 in Mpumalanga, and July 1999 in Western Cape). Hospital nurses were trained to review the admission registers and compile a line-list of case-patients on a standard form. The following patient information was abstracted: age, gender, ethnicity, admission and discharge dates and diagnoses, and the outcome of hospitalization (died in the hospital or discharged alive). Admission and discharge diagnoses were abstracted exactly as they appeared in the hospital registers.
To estimate the rate of measles-related complications we counted the number of measles patients for whom one or more of the following diagnoses were recorded on hospital admission and/or on discharge: pneumonia, gastrointestinal complications (diarrhoea, vomiting), otitis media, and neurological complications (convulsions, meningitis and encephalitis). In Mpumalanga, the rate of recorded complications was calculated on a subset of patient records (N = 1090), because some records were listed only in the measles book which contained no other diagnoses. Hospitalized measles patients for whom the hospitalization outcome in the hospital register was recorded as died were considered measles-related deaths.
We could not assess the accuracy of the clinical diagnosis of measles in either the surveillance reports or in the hospital study because diagnoses were not routinely confirmed by serology before October 1998. Also, in the hospital study it was not possible to assess the appropriateness of other recorded diagnoses and the accuracy of the recorded hospitalization outcome.
Data analysis
Routine measles surveillance data for 19801998 gathered through the disease notification system were available as an EpiInfo summary file, and were used to analyse demographic characteristics and plot the time series of measles cases reported during the study period in both provinces.
Data from the hospital study were entered in an EpiInfo database. Analyses were conducted using EpiInfo software6 and statistical significance was evaluated using the 2 test or the Kruskal-Wallis test for two groups. Evaluation of the difference between medians was done by the median test using SAS software.7
Results
Mpumalanga Province
Routine surveillance
In Mpumalanga in 19801998, 10 371 measles cases and 101 deaths were reported through the routine surveillance system. During 19801996, 10 189 measles cases and 101 deaths were reported; the lowest pre-campaign annual number of measles cases (135) was reported in 1990 and 1991, and the highest in 1992 (2583 cases). During the 5-year pre-campaign period (19921996), 4498 measles cases were reported, an annual average of 900 cases (Table 2). After the 1996 campaign, no measles deaths and record low numbers of measles cases were reported in 1997 and 1998 (92 and 90 cases, respectively). Both before and after the 1996 campaign, most reported cases occurred among Blacks/Africans; after the campaign, the proportion of reported measles cases in this group increased significantly.
|
|
|
|
|
|
To our knowledge, this is the first study undertaken in Africa to evaluate impact of a catch-up vaccination campaign on measles disease burden by using data from two mutually independent sourcesroutine measles surveillance and province-wide hospital record review.
South Africa has a well-developed, predominantly hospital-centred health care system in which approximately 80% of the population receives their health care mainly in the public sector.8 Even though racial inequalities in access to health care services inherited from the previous apartheid system still persist in some areas, in general one may assume that measles-related hospital admissions would occur if there was ongoing measles transmission in surrounding communities. The completeness of measles reporting in South Africa was not formally evaluated but is thought to be low because of the passive nature of the routine surveillance system that has multiple deficiencies.9 Nonetheless, the routine surveillance system was an adequate method for monitoring overall trends in disease incidence, as evidenced by a close parallelism between the routine surveillance data and the hospital admission data. In the absence of a fully functional case-based surveillance system with laboratory confirmation of all suspected measles cases, the review of the hospital registers in Mpumalanga and Western Cape was useful in verifying the reduction in measles morbidity and mortality observed in the routine surveillance system.
Several limitations should be considered when interpreting results of this study. First, diagnosis of measles in the routine surveillance system and in South African hospitals during 19921998 was entirely clinical; no laboratory confirmation was performed. Second, hospital registers were not complete for the entire study period in some participating hospitals, and alternative sources with less complete patient information (e.g. nursing turn-over records, measles record books, infection control registers) were used to supplement admission registers where available. In Mpumalanga, two hospitals located in the Eastern Highveld Region were excluded from the study because hospital registers were unavailable. This resulted in under-representation of that part of the Province. In addition, seven hospitals, including the three largest, either did not have the admission registers for 19921993 available for review or the registers were grossly incomplete. In Western Cape, the hospital registers from a major Cape Town hospital that was closed permanently in 1994 were not available for review. This hospital was a referral centre for paediatric measles-related admissions and was reportedly admitting hundreds of measles cases each year in the early 1990s. However, in both provinces, most participating hospitals had complete admission registers for recent years (19961999), providing a more precise assessment of the measles-associated hospitalizations in the post-campaign period. Third, the 19961997 measles vaccination campaigns coincided with an increase in reported measles incidence in both provinces, and it was not possible to separate the effect of the campaigns from that of the 19951996 measles epidemic on the subsequent reduction in measles morbidity and mortality. Fourth, since information on vaccination was not collected either in the routine surveillance system or in hospital admission registries, the vaccination history of measles cases could not be ascertained. Lastly, the one-time retrospective nature of the hospital record review and the short post-campaign observation time limit inferences about the duration of the reduction in measles disease burden resulting from the campaigns.
It is unlikely that our study overestimated the post-campaign reduction of measles-related hospitalizations and hospital-based deaths given the high participation rate of public hospitals in both provinces and the availability of generally complete hospital registers for post-campaign years. Moreover, it is likely that the proportion of patients incorrectly diagnosed with measles may have increased following the 19961997 campaigns compared to the pre-campaign period, due to decreased predictive value positive of the clinical diagnosis of measles in the presence of reduced disease incidence.10,11 In both study provinces, the proportion of hospitalized measles cases diagnosed with pneumonia was significantly lower after the campaign. This finding may reflect misdiagnosis of milder illnesses (e.g. rubella) as measles or admission of milder measles cases after the campaign.
Currently available data from the new case-based measles surveillance system that requires laboratory confirmation suggest that measles virus does not cause the majority of clinically diagnosed measles cases in South Africa reported after the 19961997 campaigns. Of 904 suspected measles cases reported in 1999, serum specimens were taken for analysis and test results were available for 817 (90%); of these 79 (10%) tested positive for measles-specific IgM antibodies.12 In 2000, of 1449 suspected measles cases, 1303 (90%) had laboratory test results available; of these, 77 (6%) tested positive.13 From January to December 2001, 901 suspected cases were reported; of these, 859 (99%) were tested and only 8 (1%) were confirmed as measles.14 This consistently low proportion of reported measles cases confirmed by the laboratory during the period 19992001 is consistent with a sustained reduction of measles virus transmission in South Africa.
The implications of the ongoing HIV/AIDS pandemic on measles control have not yet been fully elucidated, but concerns have been raised that HIV/AIDS might impede measles elimination efforts due to reduced vaccine effectiveness in infected children.15 We did not find evidence that this was the case in South Africa, a country with one of the worst HIV/AIDS epidemics in the world.16 However, our study was not specifically designed to investigate the interaction between the HIV/AIDS epidemic and measles elimination efforts.
For continued success of measles elimination activities in South Africa, vaccination strategies aimed at preventing a major accumulation of susceptibles in new birth cohorts that could result in measles resurgence will be critically important. Experience in the Americas has shown that routine measles immunization of infants can be successfully complemented, but not replaced, with mass vaccination campaigns while pursuing measles elimination.17
In Mpumalanga in 1998, a province-wide vaccination coverage survey conducted at district level estimated the routine measles vaccination coverage among children aged 1223 months at 71%, with a marked heterogeneity of district-specific coverage, indicating the need for additional supplemental immunization activities to achieve and sustain measles elimination.18 In the same study, a strong negative correlation was observed between each districts rank for vaccination coverage achieved in the1996 catch-up campaign and the rank for routine measles vaccination coverage. This suggests that children residing in the areas where routine coverage was low particularly benefited from an additional opportunity for immunization offered through the campaign.18 Earlier studies in South Africa found that the impact of previous supplemental measles vaccination campaigns was short-lived in rapidly growing peri-urban informal settlements.19,20 The statistically significant increase in the proportion of Blacks/Africans among measles cases reported in Mpumalanga during 19971998 may possibly be explained by limited measles virus circulation in some communities among remaining susceptibles; however, in the absence of laboratory confirmation for these cases, this could not be ascertained. While continuing efforts to improve routine childhood measles vaccination coverage with two doses of measles vaccine to 95% in all districts are essential, South Africa will need to implement follow-up campaigns at 3- to 5-year intervals. In June 2000 all provinces of South Africa successfully implemented a follow-up measles vaccination campaign targeting all children aged 959 months.
Our study has documented a short-term reduction of measles disease burden in Western Cape and Mpumalanga following the catch-up campaigns implemented 19961997 to historic low levels, but a longer observation period is needed to evaluate success of the measles elimination strategies implemented in South Africa. The ultimate goal is to strengthen the surveillance system so as to enable routine laboratory confirmation and detailed epidemiological investigations of all suspected measles cases and all outbreaks, and generate information on proportion of imported cases, distribution of outbreak sizes, and distribution of the duration of outbreaks (useful for assessing the progress toward elimination of indigenous measles).21 Molecular epidemiology techniques based on virus isolation and genomic sequencing may be useful in evaluating origins of measles virus importations. Lastly, further research is needed to evaluate interaction between the HIV/AIDS epidemic and measles elimination strategies.
KEY MESSAGES
|
Acknowledgments
This study was funded by the Centers for Disease Control and Prevention, Atlanta, USA. The authors thank Ms Ursheila Matai, formerly with the Expanded Programme on Immunization of the Department of Health, Pretoria, South Africa for her contribution to the study design and implementation, the nurses in Mpumalanga and Western Cape Provinces for their assistance with data collection, Ms Karen Wooten of the National Immunization Program, CDC Atlanta, for statistical support during the data analysis, and Ms Mary McCauley of the National Immunization Program, CDC Atlanta, for assistance in editing an earlier version of this paper.
References
1 Olive JM, Biellik R. Elimination of Measles in the Republic of South AfricaProposed Plan of Action. Pretoria, January 1996.
2 Department of Health, Pretoria, South Africa. Epidemiological Comments 1992;19:11226.
3 Department of Health, Pretoria, South Africa. Epidemiological Comments 1995;22:190214.
4 Centers for Disease Control and Prevention, Atlanta, USA. Progress toward measles eliminationSouthern Africa, 19961998. MMWR 1999;48:58589.[Medline]
5 Statistics South Africa. The People of South Africa, Population Census, 1996: Census in Brief. Pretoria, South Africa: Statistics South Africa, 1996 (Report No. 030111[1996]).
6 Dean AG, Dean JA, Coulumbier D et al. Epiinfo Vesion 6: A Word Processing, Database and Statistics Program for Public Health on IBM-Compatible Microcomputers. Centers for Disease Control and Prevention, Atlanta, GA, USA, 1996.
7 SAS Institute Inc. SAS/STAT Software: Changes and Enhancements through Release 6.12. Cary NC: SAS Institute Inc., 1997.
8 McIntyre D, Bloom G, Doherty J, Brijal P. Health Expenditure and Finance in South Africa. Published jointly by the World Bank and the Health Systems Trust, Durban, South Africa, 1995.
9 Durrheim DN, Harris BN, Speare R, Billinghurst K. The use of hospital-based nurses for the surveillance of potential disease outbreaks. Bull World Health Organ 2001;79:2227.[ISI][Medline]
10 Brown DWG, Ramsey MEP, Richards AF, Miller E. Salivary diagnosis of measles: a study of notified cases in the United Kingdom, 19919. BMJ 1994;308:101517.
11 Ferson MJ, Young LC, Robertson PW, Whybin LR. Difficulties in clinical diagnosis of measles: proposal for modified clinical case definition. Med J Aust 1995;163:36466.[ISI][Medline]
12 World Health Organization Regional Office for Africa, Harare, Zimbabwe. Southern Africa Integrated Disease Surveillance Network Feedback Bulletin Vol. I, Issue 4, 1999.
13 World Health Organization Regional Office for Africa, Harare, Zimbabwe. Southern Africa Integrated Disease Surveillance Network Feedback Bulletin Vol. II, Issue 4, Dec 2000.
14 World Health Organization Regional Office for Africa, Harare, Zimbabwe. Southern Africa Integrated Disease Surveillance Network Feedback Bulletin Vol. III, Issue 1, Jan 2002.
15 Moss WJ, Cutts F, Griffin D. Implications of the human immunodeficiency virus epidemic for control and eradication of measles. Clin Infect Dis 1999;29:10612.[ISI][Medline]
16 Williams B, Campbell C. Understanding the epidemic of HIV in South Africaanalysis of the antenatal clinic survey data. S Afr Med J 1998; 88:24751.[ISI][Medline]
17 Hersh BS, Tambini G, Noguiera AC et al. Review of regional measles surveillance data in the Americas, 199699. Lancet 2000;355:194348.[CrossRef][ISI][Medline]
18 Durrheim DN, Ogunbanjo GA. Measles eliminationis it achievable. Lessons from an immunisation coverage survey. S Afr Med J 2000;90:13035.[ISI][Medline]
19 Kearney M, Yach D, Van Dyk D, Fisher SA. Evaluation of mass measles immunisation campaign in a rapidly growing peri-urban area. S Afr Med J 1989;76:15759.[ISI][Medline]
20 Berry DJ, Yach D, Hennink MH. An evaluation of the national measles vaccination campaign in the new shanty areas of Khayelitsha. S Afr Med J 1991;79:43336.[ISI][Medline]
21 De Serres G, Gay NJ, Farrington CP. Epidemiology of transmissible diseases after elimination. Am J Epidemiol 2000;151:103948.[Abstract]