A data platform to improve rabies prevention, Sri Lanka

Abstract Problem In Sri Lanka, rabies prevention initiatives are hindered by fragmented and delayed information-sharing that limits clinicians’ ability to follow patients and impedes public health surveillance. Approach In a project led by the health ministry, we adapted existing technologies to create an electronic platform for rabies surveillance. Information is entered by trained clinical staff, and both aggregate and individual patient data are visualized in real time. An automated short message system (SMS) alerts patients for vaccination follow-up appointments and informs public health inspectors about incidents of animal bites. Local setting The platform was rolled out in June 2016 in four districts of Sri Lanka, linking six rabies clinics, three laboratories and the public health inspectorate. Relevant changes Over a 9-month period, 12 121 animal bites were reported to clinics and entered in the registry. Via secure portals, clinicians and public health teams accessed live information on treatment and outcomes of patients started on post-exposure prophylaxis (9507) or receiving deferred treatment (2614). Laboratories rapidly communicated the results of rabies virus tests on dead mammals (328/907 positive). In two pilot districts SMS reminders were sent to 1376 (71.2%) of 1933 patients whose contact details were available. Daily SMS reports alerted 17 public health inspectors to bite incidents in their area for investigation. Lessons learnt Existing technologies in low-resource countries can be harnessed to improve public health surveillance. Investment is needed in platform development and training and support for front-line staff. Greater public engagement is needed to improve completeness of surveillance and treatment.


Introduction
As in many other countries in Africa and Asia, 1 rabies remains a public health problem in Sri Lanka. In 2015, 607 of 1166 animal heads tested positive for rabies and 24 human deaths were reported. 2 In 1975 the Sri Lankan Ministry of Health initiated a series of nationwide rabies prevention measures, coordinated by the public health veterinary services. These covered the introduction of dog vaccination (including stray dogs after 1997), post-exposure prophylaxis for patients presenting after an animal bite (after 1992) and sterilization of stray dogs (after 2008). 3 In 2015, 1 294 529 domestic dogs and 152 404 stray dogs were vaccinated and 274 405 patients started on postexposure prophylaxis. 2 Patients who are bitten by a mammal can attend rabies clinics in government hospitals where postexposure prophylaxis is provided free of charge. Public health information ensures high public awareness of rabies risk and the location of clinics. If the bite is low severity or the animal is known, treatment is deferred for 14 days while the animal is observed for clinical signs of rabies. 4 Otherwise, post-exposure prophylaxis is started, which requires the patient to attend the clinic for vaccination with inactivated cell-culture-based rabies vaccine on days 0, 3, 7 and 30 following presentation. 5 Samples of dead animals from all parts of Sri Lanka are sent to specialist laboratories that use direct fluorescent antibody testing for the presence of rabies virus.
However, clinical treatment, prevention initiatives and surveillance of rabies epidemiology in Sri Lanka are hindered by weaknesses in record-keeping and data collection. Clinic health records and reports to the public health inspectorate are paper-based and regional variations in completeness of recordkeeping exist. Monthly reports are collated and disseminated in an electronic database. Clinicians therefore have limited access to information on a patient's completion of post-exposure prophylaxis regimes and nationwide surveillance to ascertain the completion rate of anti-rabies treatment in Sri Lanka is poor. 6 Similarly, there is no system for contacting patients to attend for vaccinations. This contributes to missed, delayed or duplicated vaccinations. 7,8 Another problem is that feedback of laboratory results to the public health veterinary services is via manual monthly reports. This delays the ability of frontline teams to identify and respond to rabies outbreaks, locate infected animals and contact patients associated with animals that have tested positive. Finally, public health inspectors and the public health veterinary services have limited ability to Problem In Sri Lanka, rabies prevention initiatives are hindered by fragmented and delayed information-sharing that limits clinicians' ability to follow patients and impedes public health surveillance. Approach In a project led by the health ministry, we adapted existing technologies to create an electronic platform for rabies surveillance. Information is entered by trained clinical staff, and both aggregate and individual patient data are visualized in real time. An automated short message system (SMS) alerts patients for vaccination follow-up appointments and informs public health inspectors about incidents of animal bites. Local setting The platform was rolled out in June 2016 in four districts of Sri Lanka, linking six rabies clinics, three laboratories and the public health inspectorate. Relevant changes Over a 9-month period, 12 121 animal bites were reported to clinics and entered in the registry. Via secure portals, clinicians and public health teams accessed live information on treatment and outcomes of patients started on post-exposure prophylaxis (9507) or receiving deferred treatment (2614). Laboratories rapidly communicated the results of rabies virus tests on dead mammals (328/907 positive). In two pilot districts SMS reminders were sent to 1376 (71.2%) of 1933 patients whose contact details were available. Daily SMS reports alerted 17 public health inspectors to bite incidents in their area for investigation. Lessons learnt Existing technologies in low-resource countries can be harnessed to improve public health surveillance. Investment is needed in platform development and training and support for front-line staff. Greater public engagement is needed to improve completeness of surveillance and treatment.
Lessons from the field Rabies prevention in Sri Lanka A Pubudu De Silva et al.
evaluate the effectiveness of the current canine vaccination programme on a national basis and to identify regional variations in activity (e.g. urban compared with rural locations) and transmission characteristics. 1 This paper describes the creation of an electronic platform to improve the collection and sharing of data for better management and surveillance of rabies in Sri Lanka. The platform was developed by adapting existing acute and critical care platforms 9 and was driven by clinicians looking for solutions to the problems outlined above.

Approach
The project was a joint initiative of the public health veterinary services of the health ministry and the Network for Improving Critical Care Systems and Training Sri Lanka. It aimed to provide real-time data on: (i) the incidence and characteristics of mammal bites to humans; (ii) patient demographics; (iii) treatment decisions and patient outcomes; and (iv) the prevalence of rabies in dead mammals.
The design of the data platform was clinician-led, focusing on the information needed to improve front-line care.
We conducted scoping interviews with public health veterinary services and clinical staff to determine users' data requirements. In hospitals the required variables included patient characteristics, animal characteristics, severity of the bite and patient treatment details (Box 1). In laboratories the variables included type of animal, habitat, mode of death, vaccination status and investigation results. We developed the data collection tool using REDCap software (REDCap Consortium, Nashville, United States of America). The tool was accessible to users via the internet from desktop computers and mobile devices. Data on each bite incident are entered by trained nursing officers when patients first present to a hospital rabies clinic. Data on animal test results are entered by clinical staff in the relevant laboratories.
Initial training was done by the central registry team at visits to rabies clinics and laboratories. During the implementation phase, ongoing clinician support was provided by the registry staff by telephone, with daily validation phone calls to the clinics to determine the numbers and motivate staff towards data entry. Laboratories were also contacted daily by the central registry validation team during implementation to help improve information completeness and staff skills.
We used a business analytics software (Tableau version 10.0, Tableau Software, Seattle, USA) to develop dashboards that would enable users to visualize the data in real-time, either as an aggregate or filtered by incident location or characteristics. The data are accessible to front-line staff (nurses and doctors), administrators, public health inspectors and laboratory teams via an online portal. Aggregated data are accessible to the general public, so that trends in bite incidence, patient demographics, mammal characteristics and laboratory test results can be identified. Individual patient data and laboratory results require a two-step authentication and are accessible only to clinicians treating the patient and to public health inspectors. The data are then displayed through a dashboard via the registry platform in the secure servers of the Information and Communication Technology Agency of the Government of Sri Lanka. 10 Mobile phone ownership is very high in Sri Lanka and follow-up of patients is aided by linking the postexposure prophylaxis dashboard to a bespoke automated short message system (SMS). The clinician enters the patient's mobile phone number and national identification number (the only unique identifier) to enable treatment episodes and outcomes to be linked. SMS alerts are sent to the patient the day before the due date of anti-rabies vaccine. Patients with deferred treatment are contacted on day 14 of the observation period of the suspected animal. Teams in rabies clinics are alerted by SMS to the numbers of patients due for anti-rabies vaccine each day, while public health inspectors are alerted to the incidence and characteristics of bites reported within their region, and any subsequent animal deaths. A schematic overview of the platform is shown in Fig. 1.

Relevant changes
The live platform was rolled out on 1 June 2016 across four districts in Sri Lanka (two urban and two rural: Colombo, Gampaha, Mathale and Monaragala), with a combined population of 3  Lessons from the field Rabies prevention in Sri Lanka A Pubudu De Silva et al.

SMS alert system was initiated on 1
September 2016 in two of the districts that volunteered to participate. Between 1 June 2016 and 28 February 2017, 12 121 reported incidents of animal bites from the four districts were entered into the registry, providing real-time data on the incidence, character and geographical location of mammal bites. Clinicians and public health teams were able to access data on the clinical management and outcomes of patients on post-exposure prophylaxis (9507; 78.4%) or with treatment deferred (2614; 21.6%). Five of the six clinics reported on the completeness of post-exposure prophylaxis, showing that 293 (7.9%) of 3709 patients completed all four anti-rabies vaccinations. Data on other patient outcomes (e.g. delays in starting treatment, number of deaths) will be available in the future. Information on the characteristics of suspected sources (60.8% dogs, 24.8% cats, 7.1% rats and other 7.4% nondomesticated mammals) enabled the public health veterinary services to evaluate current animal vaccination strategies and identify regional epidemiological variances.
Of 907 dead mammals tested, 328 (36.2%) were positive for rabies virus. The laboratories entered these data into the registry so that animal test results could be communicated rapidly to the clinic teams and the public health inspectorate responsible for individual cases.
Over the 6-month period, the SMS alert system in two districts alerted 1376 (71.2%) patients on post-exposure prophylaxis to attend their vaccination appointments; contact details were unavailable for 557 patients. Daily SMS reports alerted 17 public health inspectors to 3125 bite incidents in their region for further investigation.

Lessons learnt
This proof-of concept study illustrates how technologies that already exist in many low-and middle-income countries can be harnessed with existing staff resources to create an electronic platform for improving public health surveillance (Box 2). Maximizing the availability of information on the incidence, geographical location and outcomes of mammal bites should facilitate timely responses to incidents, and inform decisions about resource allocation and the refinement of treatment and prevention programmes. The SMS alerts health-care responders to clusters of incidents in their region and aims to improve compliance by patients on treatment; similar systems have proved effective in other health-care settings. 11 Work is ongoing with front-line staff to improve completeness of contact information and to explore low treatment compliance in this setting.
To expand the platform within Sri Lanka, investment is needed in platform development and in training and support for clinical staff, both to address data completeness and to refine the platform before scaling-up the initiative. Such investment would also allow an evaluation of the impact of the platform on reducing avoidable delays in starting

Box 2. Summary of main lessons learnt
Existing technologies in low-and middle-income countries can be harnessed using minimal resources to provide a platform for improved public health surveillance.
To expand the platform within Sri Lanka and to test its transferability to other settings, investment is needed in platform development and training and support for front-line staff. Greater public engagement is needed to improve completeness of surveillance and treatment.
Lessons from the field Rabies prevention in Sri Lanka A Pubudu De Silva et al.
post-exposure prophylaxis. It would also inform targeted animal vaccinations, using the data on where and when bites occurred. The existing media campaign may need to be boosted to raise the public's awareness of the importance of reporting animal bites and complying with post-exposure care.
The next stage is to extend the platform to all rabies clinics nationwide to assist clinicians and administrators throughout the country to organize resources and improve the completeness rates of post-exposure prophylaxis. This information could be accessed remotely by specialists to help local teams refine prevention strategies based on regional epidemiological information and to enable clinicians to evaluate existing and future treatment protocols for rabies and other communicable diseases. ■

Resumen
Una plataforma de datos para mejorar la prevención de la rabia, Sri Lanka Situación En Sri Lanka, las iniciativas de prevención de la rabia se ven obstaculizadas por un intercambio de información fragmentado y retrasado que limita la capacidad de los médicos de hacer seguimientos a sus pacientes e impide el control de la atención sanitaria. Enfoque En un proyecto dirigido por el ministerio de sanidad, se adaptaron las tecnologías existentes para crear una plataforma electrónica para el control de la rabia. Personal clínico formado introduce la información, y los datos sobre pacientes se pueden consultar en tiempo real tanto de forma individual como global. Un sistema automatizado de mensajes cortos (SMS) alerta a los pacientes de citas para vacunas de seguimiento e informa a los inspectores de salud pública sobre incidentes de mordeduras de animales. Marco regional La plataforma se implementó en junio de 2016 en cuatro distritos de Sri Lanka y vinculó seis clínicas de atención contra la rabia, tres laboratorios y el cuerpo de inspectores de salud pública. Cambios importantes En un periodo de 9 meses, se informó a las clínicas de 12 121 mordeduras de animales, las cuales fueron inscritas en el registro. A través de portales seguros, los médicos y los equipos de salud pública accedieron a información en directo sobre el tratamiento y los resultados de los pacientes que comenzaron con profilaxis posterior a Lessons from the field Rabies prevention in Sri Lanka A Pubudu De Silva et al. la exposición (9 507) o con tratamiento aplazado (2 614). Los laboratorios comunicaron con rapidez los resultados de las pruebas del virus de la rabia en animales muertos (328/907 positivos). En dos distritos piloto, se enviaron recordatorios SMS a 1 376 (71,2%) de los 1 933 pacientes de los que se disponían detalles de contacto. Informes con SMS diarios alertaron a 17 inspectores de salud pública sobre incidentes de mordeduras en su zona para que los investigaran.
Lecciones aprendidas Las tecnologías existentes se pueden aprovechar en países con pocos recursos para mejorar el control de la salud pública. Se necesita invertir en el desarrollo de plataformas y en la formación y el apoyo del personal de primera línea. Es necesario un mayor compromiso público para mejorar la finalización del control y el tratamiento.