Short message service sentinel surveillance of influenza-like illness in Madagascar, 2008–2012
Soatiana Rajatonirina a, Jean-Michel Heraud b, Laurence Randrianasolo a, Arnaud Orelle b, Norosoa Harline Razanajatovo b, Yolande Nirina Raoelina c, Lisette Ravolomanana c, Fanjasoa Rakotomanana a, Robinson Ramanjato c, Armand Eugène Randrianarivo-Solofoniaina c & Vincent Richard d
a. Unité d’Epidémiologie, Institut Pasteur de Madagascar, Antananarivo, Madagascar.
b. National Influenza Center, Institut Pasteur de Madagascar, Antananarivo, Madagascar.
c. Ministère de la Santé Publique, Antananarivo, Madagascar.
d. Unité d’Épidémiologie, Institut Pasteur de Dakar, 36 avenue Pasteur, BP 220, Dakar, Sénégal.
Correspondence to Vincent Richard (e-mail: email@example.com).
(Submitted: 18 October 2011 – Revised version received: 12 March 2012 – Accepted: 13 March 2012.)
Bulletin of the World Health Organization 2012;90:385-389. doi: 10.2471/BLT.11.097816
Four years after the revised International Health Regulations (IHR) went into effect and amidst fears that influenza pandemics and other disease outbreaks could seriously threaten the health of entire populations in developing countries, it became necessary to bolster surveillance capacity, particularly in low-resource settings.1 This was especially important in such settings because they generally lack the health-care infrastructure, laboratory diagnostic capability and number of physicians needed to effectively control emerging epidemic-prone diseases.
The circulation of the highly pathogenic avian influenza A(H5N1) virus in humans and animals in Asia, Europe and Africa since 2006 has raised fears of the possible emergence of a highly pathogenic pandemic influenza strain.2 This has prompted many developing countries to establish surveillance systems for the early detection of the emergence and/or circulation of potential pandemic influenza strains.
Madagascar’s disease surveillance systems have traditionally relied on the passive collection and reporting of data aggregated weekly or monthly. Such systems cannot detect incipient disease clusters for prompt investigation and response. In addition, diagnostic capacity outside the capital city is limited and delays in identifying the causal agents of disease outbreaks have often resulted in an inability to launch an appropriate public health response. Madagascar needed a surveillance system that could quickly identify aberrant patterns of disease activity rather than depend exclusively on case confirmation.
In March 2007 Madagascar’s Ministry of Health, in collaboration with the Institut Pasteur of Madagascar (IPM), launched a sentinel surveillance system designed to detect incipient disease outbreaks through daily collection of syndromic case data in sentinel health centres,3 backed by biological surveillance (e.g. collection and testing of biological specimens) at selected sentinel sites.
The first step in creating the sentinel network was to identify appropriate stakeholders. A steering committee was established with epidemiologists, virologists and managers from the Ministry of Health’s epidemiological surveillance service (Direction de la Veille sanitaire et de la Surveillance épidémiologique, DVSSE) and emerging and neglected disease control units (Direction des Urgences et de la Lutte contre les Maladies Négligées, DULMN). Next, regional health department and district managers were brought into the discussions surrounding the selection of potential sentinel centres.
The sentinel network in Madagascar, which is a large island, began with 6 sentinel sites in 2007 and had expanded to 34 by 2011. Expansion was intended to improve geographic coverage and representativeness and to make it possible to monitor epidemiologic trends in different climate zones.
Sentinel sites are primary health-care centres that are selected on the basis of several criteria, including the presence of at least two general practitioners, the availability of a mobile phone network, and the degree of commitment to actively participate in surveillance activity. Participation is entirely voluntary. Sentinel general practitioners (SGPs) serving on a gratuitous, voluntary basis are the backbone of the system, which currently covers about 8% of the Malagasy population.
To assess the functioning of the surveillance system, we collected data from sentinel general practitioners from 1 January 2008 to 31 January 2012. All patients seen at the sentinel centres were systematically screened by each practitioner, who reported the number of cases that met the case definitions (described later) and the total number of patient visits on each reporting day.
Sentinel general practitioners were expected to transmit encrypted data via cellular telephone (encrypted short message service, SMS) at least once a day, from Monday to Friday, despite occasional clinic closures resulting from routine weekday and weekend schedules. Active data collection was also carried out: if data from a sentinel centre were not received by 10:00 on a particular day, IPM staff would contact the sentinel centre to obtain the missing data.
The monthly cost of data transmission per centre was less than 2 United States dollars (US$).
Standard WHO case definitions were used for comparability. Fever (an axillary temperature ≥ 37.5 °C) was the first syndromic symptom targeted. Three febrile illnesses were selected for surveillance: malaria (only confirmed cases), influenza-like illness (ILI) and arbovirus infection. Cases were defined as follows: malaria, fever plus a positive result on the rapid diagnostic test; ILI, fever with a cough and sore throat; arbovirus infection, fever without respiratory symptoms plus at least two of the following: headache, arthralgia, myalgia-like backache, skin rash, retro-orbital pain, haemorrhage. Diarrhoea syndrome was defined as three or more abnormally loose stools during the previous 48 hours. A malaria rapid diagnostic test was performed on every febrile patient with the kit provided by the national malaria programme.
Data were available for analysis shortly after each patient’s initial visit. The data obtained daily from the SMS were entered into a specifically designed MS Access® database as soon as they arrived at the IPM. Thus, the turnaround time between data collection and receipt of the data, even from remote areas, never exceeded 24 hours. The data received through the SMS system included sentinel site code, date of data collection, number of outpatient consultations, number of malaria cases , ILI, dengue-like illness and diarrhoea, and number of consultations by age group.
Data were checked for variation from day to day. Analyses of temporal distribution were conducted at each sentinel centre for each syndrome category of interest. Hence, various health service-based indicators were monitored daily. They included the percentage of cases with fever, confirmed malaria, ILI and suspected arbovirus infection, as well as the percentage of diarrhoea cases with and without fever.
Surveillance data were analysed and presented using easy-to-interpret tables and graphs that illustrated the number of cases of each syndrome under surveillance. In addition, a daily and weekly baseline (i.e. average number of cases in the previous years) was calculated for each syndrome and plotted against current observations to identify aberrant patterns. Increases were reported immediately by telephone to the Ministry of Health to monitor events and response and decide whether to initiate an outbreak investigation. The first step in any outbreak investigation was to confirm the signals before responding to an epidemic.
SGPs reported information on each patient visit and listed every new febrile patient’s chief complaint or symptoms on a special paper form sent weekly to the IPM. The staff of the IPM was therefore able to validate all information previously sent via mobile phone.
An external report summarizing weekly and monthly trends in centre activity were distributed to state and regional health staff and was shared with participating SGPs.
The data collected daily over the study period from the 34 sentinel centres corresponded to 862 585 patient visits. Overall, 86.7% of the data were transmitted within 24 hours.
A total of 95 401 cases (11.1%) presented with fever. The special form was completed for 80 691 of these patients (84.6%). The male to female ratio among patients with a fever syndrome was 0.89:1. Age was available for 79 651 patients (98.7%). The mean age was 12.8 years (95% confidence interval, CI: 12.7–12.9).
ILI accounted for 20 933 (21.9%) of all cases of febrile illness.
The results obtained by the sentinel surveillance network in Madagascar are in accordance with the aim pursued by the country’s Ministry of Health. Sentinel surveillance systems offer advantages over passive surveillance, which is limited by incomplete reporting.4,5 They provide a way to overcome the coverage limitations inherent in the passive surveillance of disease. However, efficient and accurate sentinel surveillance systems need to rely on strong communication systems.
A “real-time” sentinel surveillance system may be essential for the detection of infectious disease outbreaks and the launching of a prompt public health response (Box 1). The advantage of using syndromic data for outbreak detection lies in the speed with which a response can be launched.6,7 Mobile phone technology, which is cheap and universally available, has been recognized as a useful tool in health care.8 However, published studies on mobile phone use in health-care settings have been conducted almost exclusively in countries where a shortage of resources and the need to provide care in remote locations are the main reasons for using mobile phones.9 The cost of transmitting data on a daily basis in Madagascar’s sentinel surveillance network is minimal; it costs less than US$ 2 per month per sentinel centre, and each centre’s mobile phone equipment costs a mere US$ 10. However, the costs of maintaining the system need to be more accurately estimated, both in terms of monetary expense and the person–hours involved in responding to system alerts. Although the Ministry of Health endorses sentinel general practitioners’ voluntary participation, some practitioners are asking for incentives. Motivation has been maintained not through monetary payment but through the provision of medical equipment and training opportunities, including an annual epidemiology workshop and an annual meeting of staff in all sentinel sites. We have noted high staff turnover and, to address the problem, health district officers are being trained to train and supervise new staff.
Box 1. Summary of main lessons learnt
- Daily syndromic surveillance using SMS can effectively enhance traditional public health surveillance systems already in place.
- Combined biological surveillance and syndromic surveillance using SMS makes it possible to rapidly detect the circulation of the influenza virus in areas under surveillance.
- By detecting unusual patterns of disease activity, sentinel surveillance using SMS can quicken the response to disease outbreaks.
In conclusion, the sentinel surveillance network would not be effective without the active participation of the SGPs. Other African countries would benefit from a similar programme, both in terms of improved disease surveillance and better public health decision-making.
We thank all the staff from the Malagasy National Influenza Centre for influenza testing. We would like to express our gratitude to all the staff from the DULMN and the DVSSE (Ministry of Health). We are deeply indebted to all practitioners and nurses involved every day in sentinel surveillance in Madagascar. To end, we would also like to thank Stefano Tempia and Adam Cohen from CDC South Africa for helpful comments and discussions during the process of writing
This work was supported by WHO (APW/Ref. OD/AP-08-02451), the French Ministry of Health, the US Centres for Disease Control and Prevention (Cooperative Agreement Number: U51/IP000327-01), the US Department of Health and Human Services (Grant Number 6 IDSEP060001-01-01) via the International Network of Pasteur Institutes, and the President’s Malaria Initiative programme (USAIDS). We particularly thank Kathleen Victoir and Marc Jouan from the International Network of Pasteur Institutes.
- Raviglione MC, Gupta R, Dye CM, Espinal MA. The burden of drug-resistant tuberculosis and mechanisms for its control. Ann N Y Acad Sci 2001; 953: 88-97 doi: 10.1111/j.1749-6632.2001.tb11364.x pmid: 11795426.
- Uyeki TM. Global epidemiology of human infections with highly pathogenic avian influenza A (H5N1) viruses. Respirology 2008S2-9 doi: 10.1111/j.1440-1843.2008.01246.x pmid: 18366524.
- Randrianasolo L, Raoelina Y, Ratsitorahina M, Ravolomanana L, Andriamandimby S, Heraud JM, et al., et al. Sentinel surveillance system for early outbreak detection in Madagascar. BMC Public Health 2010; 10: 31- doi: 10.1186/1471-2458-10-31 pmid: 20092624.
- Arita I, Nakane M, Kojima K, Yoshihara N, Nakano T, El-Gohary A. Role of a sentinel surveillance system in the context of global surveillance of infectious diseases. Lancet Infect Dis 2004; 4: 171-7 doi: 10.1016/S1473-3099(04)00942-9 pmid: 14998504.
- Gauci C, Melillo Fenech T, Gilles H, O'Brien S, Mamo J, Stabile I, et al., et al. Sentinel surveillance: an option for surveillance of infectious intestinal disease. Euro Surveill 2007; 12: E13-4 pmid: 17991385.
- Wu TS, Shih FYF, Yen MY, Wu JSJ, Lu SW, Chang KCM, et al., et al. Establishing a nationwide emergency department-based syndromic surveillance system for better public health responses in Taiwan. BMC Public Health 2008; 8: 18- doi: 10.1186/1471-2458-8-18 pmid: 18201388.
- Miller B, Kassenborg H, Dunsmuir W, Griffith J, Hadidi M, Nordin JD, et al., et al. Syndromic surveillance for influenza-like illness in ambulatory care network. Emerg Infect Dis 2004; 10: 1806-11 pmid: 15504267.
- Kaplan WA. Can the ubiquitous power of mobile phones be used to improve health outcomes in developing countries? Global Health 2006; 2: 9- doi: 10.1186/1744-8603-2-9 pmid: 16719925.
- Yang C, Yang J, Luo X, Gong P. Use of mobile phones in an emergency reporting system for infectious disease surveillance after the Sichuan earthquake in China. Bull World Health Organ 2009; 87: 619-23 doi: 10.2471/BLT.08.060905 pmid: 19705013.