Smart Surveillance and Micromanagement Using Information Technology for Malaria Elimination in Mangaluru, India – an Analysis of Five Years Data

Background Malaria control system (MCS), an Information technology (IT)-driven surveillance and monitoring intervention is being adopted for elimination of malaria in Mangaluru city, Karnataka, India since October 2015. This facilitated ‘smart surveillance’ followed by required �eld response within a timeline. The system facilitated data collection of individual case and data driven mapping and strategies for malaria elimination programme. This paper aims to present the analysis of post-digitization data of 5 years, discuss the current operational functionalities of MCS and its impact on the malaria incidence.


Background
Globally malaria is still a major public health problem although the scenario has changed a little since 2015 [1][2][3][4].The incidence of malaria reduced from 71 to 57 per 1000 population at risk [1].The challenges faced are operational, health system de ciencies and poor management systems [2,5].
Information technology (IT) system to improve surveillance, complete case reporting data analysis that lead to timely responses in the eld leading to robust and responsive surveillance is critical for malaria elimination [6].Available global IT systems have been evaluated with regards to the structure of the system, data captured, output, strengths and challenges [6].These authours have identi ed inability to capture private health sector data, nil documentation of eld response, failure to map the case, di culty in tracking migrant workers, failure to capture time of reporting, inability to capture data in real time, and non-integration with mobile technology as challenges of using IT systems.Malaria surveillance data available through the routine malaria information system (MIS) that was used did not provide the much needed information on severe malaria cases, as a large number of patients seek health care from the private sector and did not gure in the programme data [6,7].
Mangaluru (Mangalore) is a coastal city in Karnataka of southwestern India.The city has administrative units designated as wards, and 60 such wards constitute the city limits [8].Malaria has been endemic in Mangaluru for three decades [1][2][3].Malaria control measures were being carried out as per National Vector Borne Disease Control Programme (NVBDCP); however desired results were not oberved.To address the de ciency of existing system and to improve performance of control strategies, a new IT system namely Malaria Control System (MCS) was introduced in October 2015 in Mangaluru and is operational till date [9].This IT system was introduced to capture data and build capacity of existing programme in the entire city MCS consists of innovative handheld, Android-based geographical information system (GIS)-tagged tablets (TABs) device, and a web-based incident reporting system.The system ensures `smart surveillance'coupled with eld response and data collection for analysis to design local strategies for malaria elimination.MCS was introduced as a programme management system and as an intervention to assist effective management of malaria control programme by digitizing the reporting of newly diagnosed malaria cases for treatment, tracking and closure of cases after complete treatment of each malaria patient.Malaria control software is being used for the sixth consecutive year and cases are reported by all the health care providers and stakeholders including from private sectors.Field activities for control and closure of cases and source elimination of breeding habitats are carried out based on the inputs into the software.Routine monitoring and strict vigils were put in place on the ongoing newly introduced surveillance system using GIS-tagged TABs.In a previous article, we had described the design and implementation of this IT system protocol and presented initial secondary data analyses to determine the impacts in 2-year post-digitization [9].
In the post-digitization years, it was easier to access and retrieve the data.Hence routine real time monitoring and analyses of malaria indices in all the wards covering the entire city limits was possible.
Administrators were able to identify high-risk areas periodically to carry out necessary additional antimalarial activities.
This paper aims to present the analyses of ve-year post-digitization data, discuss the current operational functionalities of MCS and its impact on the malaria incidence.

Methods
IT solution is being continued as management and monitoring tool in the city of Mangaluru [9] since October 2015.Early reporting within 24 to 48 hrs followed by eld response in next 24 to 48 hrs along with anti-mosquito measures were carried out as per NVBDCP guidelines.This data available on the IT system was translated onto excel sheet and was analysed for taking appropriate decisions and amendments inaction plan.Secondary analyses of ve-year data was carried out.
Malaria cases reported in the city were analysed based on the type of health facilities from where patients sought health care services and its reporting.These health care facilities were categorized as private health facilities, and public health facilities.Private health facilities included all the hospitals, nursing homes and diagnostic laboratories.Public Health facilities included surveillance team of district vector borne disease control o ce (DVBDCO), government-run hospitals, urban health centres and malaria clinics.
Each malaria case was analysed based on reporting time, complete treatment and closure of the cases subsequent to follow-up smear examination for clearance of parasites, and closure of cases within day 14 and also within 30 days.Closure time is considered as 14 days to complete primaquine therapy for Plasmodium vivax cases to prevent relapse as per the recommendation of NVBDCP [10].Anti-vector responses in the eld were also analysed.
Factual reporting with regards to administrative decisions, hurdles in the implementation of anti-malarial activities, how these problems were addressed and their effects on the malaria control were noted and documented.The malaria indices were analysed.

Statistical analysis :
Closure and closure time of each positive case and vector interventions were analysed.Community visits, contact smears during active surveillance around reported case (ASARC), vector control activities were analysed along with malaria indices such as Annual Blood Examination Rate (ABER), Slide Positivity Rate (SPR), Slide Falciparum Rate (SFR) and Annual Parasite Incidence (API).Monthly trends of malaria at each level were also plotted in relation to closure of cases.Fischer F test was applied to nd the signi cance in reduction of malaria cases.Time series analysis was done for plotting the trends of closure rate of cases against the incidence of cases of malaria.Bonferroni t test was used to test the statistical signi cance of intertime interval.A p value of <0.05 was considered statistically signi cant.

Results
Monthly incidence of malaria for the past 6 years and the cumulative reduction in incidence in urban limits of Mangaluru is depicted in Table 1.Gradual reduction of overall incidence of malaria continued throughout ve-year post-digitization (PDY 5) with an overall cumulative reduction by 83% (range − 64% to -92%) as compared to digitization year (DY).#some diagnostic centres reported cases directly to malaria control cell.Yearly reduction is found to be highly signi cant.F value 17.737, p value < 0.00.Using the Bonferroni t test, the reduction in incidence was statistically signi cant as seen from the p values for inter-time interval between various years; between PDY 1 and PDY 2 (p < 0.05); PDY 1 against PDY 3(p < 0.01); and between PDY 4 and PDY 5 (p < 0.001).
The maximum cumulative reduction of 91 to 92% in incidence was noted for the months of May and June and least 64 to 66% in the months of October and November soon after monsoon season.The ward-level cumulative reduction in the incidence is also depicted in Table 2.The range of reduction of cumulative incidence is 45% (Court ward no.40) to 98% (Maroli ward no.37).The ward-level cumulative reduction in incidence of malaria from the PDY 1 to PDY 5 was signi cant (p < 0.001).It was noted that surveillance continued to improve with malaria cases being reported from all the hospitals and diagnostic centres of private as well as public health systems (Table 3).In the rst year after digitization, private healthcare facilities contributed to nearly two-thirds (68%) of the total cases being reported while the public health system contributed to nearly one-third (which included 18.6% by community public hospitals and 4.3% by malaria clinics).In the post-digitization phase, the contribution to total number of cases from private hospitals kept steadily declining and reduced to 57% in the PDY 4. At the same time, the public health system, i.e., public hospitals, urban health centres as well as DVBDCO started contributing larger proportion of total number of cases.With the onset of covid pandemic the private sector contribution was found to have increased again.Active surveillance around reported case (ASARC) contributed to over 1.6% of malaria incidence during 5th year, emphasizing the role played by it (Table 3).# cases directly reported to malaria control cell are not included Table 4 depicts the number of cases for the last 5 years' smears tested, contact smears taken and malariometric conditions.There was a negative corrrelation between the ratio of contact smears to total number of cases and number of positive cases detected by contact smears though it was not statistically signi cant.The malarial indices were calculated for the pre-digitization year, digitization year and each of the ve-year post-digitization.The SPR seen steadily decreasing and the average annual parasite incidence (API) reduced and came down to 2.64 in the ve-year post-digitization.The API, SPR, and SFR showed statistically signi cant changes (p < 0.001).Trends for overall malaria incidence over 5 years is depicted in Fig. 1.The cases seem to peak during the monsoon season but an overall decreasing annual trend is observed.Improvement in reporting of cases from point of diagnosis on the web-based software is analysed in Fig. 2. Most cases were reported same day or the next day.Similarly, Fig. 3 depicts the monthly malaria incidence as against the percentage of closure of cases within 14 days and 30 days, respectively.The source of mosquito breeding habitats were identi ed in and around the residence of malarial patient.This activity was carried out soon after new cases were reported on the system.Source identi cation was highest during rainy monsoon and is further carried out during winter and summer periods.Deatails are shown in Fig. 4.
The ward-level depiction based on API from PDY 1 to PDY 5 is shown in Fig. 5.There is a gradul shrinking of malaria map in the city.It can be noted that the wards with API in the red zone (API > 10) have reduced to only 5 wards in PDY 5 as against 43 prior to digitization.The wards in green (API ≤ 2) as well as yellow (API > 2.1 to 5) have increased over the years.

Discussion
Mangaluru has been is classi ed as a high-risk region for Urban Malaria by NVBDCP [11], endemic for malaria contributing to 85% of malaria cases in the state of Karnataka, India.Being a dual host-disease estmation of R O (reproduction number) is complex, recent models have been used to estmate R O which ranges from 1 to 3000 (12,13).E cient participation of multiple stakeholders to manage both hosts determines the results of control measures.Failure to contain malaria over two decades, in spite of the ongoing control programme stipulated a new approach.MCS was introduced in October 2015 to improve 'surveillance with timeline'and dissemination of case details for appropriate action in the eld [9].Electronic surveillance system helps to connect all stakeholders with necessary information for expected time-bound response in the eld to break the transmission chain.A multipronged, integrated approach involving all health care providers, time bound eld response i.e. active case detection and anti-mosquito measures in geographical area is critical for containment and elimination of malaria thereafter.
IT system for malaria control programme should be user friendly, easy to operate, collect information o ine and upload when internet connection is available.MCS is a dedicated IT system which is also integrated to mobile technology, and is designed to be user friendly and easy to operate.However, it required few months to train and implement the available functions of MCS by all the stakeholders namely hospitals and diagnostic centers, eld workers and administrators.Over 5 years there was an overall reduction in malaria cases by 83% and monthly incidences reduced to double digits.The trend continued during COVID-19 pandemic when the entire health system was engaged in ghting the dreaded disease.MCS affected most parameters for malaria and contributed to the effective reduction of cases in Mangaluru.
Malariometric Indices: Malariometric indices showed signi cant changes over 5 years.The incidence of both Plasmodium vivax (Pv) and P. falciparum (Pf) gradually decreased.Initially, ABER increased signi cantly with predominant contribution from passive surveillance.Stagnation between 2nd and 3rd year after implementing MCS was a consequence of administrative decision to utilize MPW for nonmalarial work.In PDY 3, CMETs were formed to supplement the active surveillanceand the results can be seen during PDY 4 and PDY 5.During COVID-19 pandemic, active surveillance could not be carried out e ciently resulting in decreased ABER and increased SPR.However, the incidence of malaria and API continued to decrease without any rebound increase in the 'post-lockdown' period.There is a need to have comprehensive approach for malaria elimination since it is a dual host disease with wide ranging R O factor, dormant stage in humans and resistance to various strategies adopted for control or elimination.
The ultimate goal of all strategies is to reduce API in the area and reduce the size of malaria map.A dedicated, user friendly system which captures data with timeline will assist in micromanaging multiple strategies.
Reporting of cases: Subsequent to 'smart surveillance', behavioural changes with respect to timely reporting of malaria cases were observed among the diagnosticians and it continued through PDY 5.
Details of 89% of diagnosed cases were uploaded into the system within 48 hrs.Both public and private health care providers reported the malarial cases (Table 3).All these were passive case detection (PCD) from health facilities with exception of reports by ASARC and DVBDCO.Very high rate of passive case detection re ects 'health seeking behaviour' of the population and is probably one of the reasons for decrease in incidence even during COVID-19 pandemic.Private sector contribution was higher than public health system and is an indication of de nite compliance to non-reporting from private health system which was a major hurdle for malaria control in India [7].As per WHO, cases of malaria are reported only from public healthcare facilities and hence, a large number of cases are unreported thereby facilitating transmission [14,15].However, even where reporting rates in the public health sector are close to a 100% in some countries, more than 50% of malaria patients seek care in the private sector [12].Hence, reporting from private sector is crucial for malaria control.
Figure 2 indicates average time taken to report from the time of diagnosis.Caprturing the case details and transferring this information to the health workers in the eld is the key to initiate control activities.This PCD robust reporting of PCD-initiated active case detection (ACD).Ideal IT system should facilitate robust reporting which is very critical for malaria elimination.It has been observed that early reporting from the diagnosticians continued even during covid epidemic thus resulting in disruption of transmission .
Field response with timeline: E cient participation of multitiple stakeholders is crucial for effective control measures.Multi-pronged, integrated approach is critical for containment of malaria and elimination thereafter.'Smart surveillance' helps to connect all stakeholders with necessary information for anticipated response in the eld to break the chain of transmission.'Time-bound' eld response ie active case detection and anti-mosquito measures in the geographical risk areaswere carried out simultaneously.Immediate contact smears and identi cation of positive cases helps to reduce parasite pool available for transmission.Use of different information systems have been successfully implemented in China.Ease of reporting with timelines, 1-3-7 intiative strategy helped to control malaria in Chinese province.A robust disease surveillance information system was used for this strategy [6,16].
Surveillance: In initial year after MCS, an increase in incidence was documented suggesting improved surveillance.In subsequent years, there was a gradual reduction in incidence of malaria.This reduction was not uniform throughtthe year.Although during and immediately after monsoon rains (June to October) there have been variable spikes in incidence, the number of cases gradually reduced during same period year on year (Fig. 1).Surveillance, early case detection, treatment and vector control measures were done as per NVBDCP guidelines with variable results.With introduction of MCS the surveillance was robust time-bound and 'incidence-centric'.Quick transfer of information from point-of-diagnosis to eld workers and surveillance thereafter contributed to 1.8% of reported cases of malaria in the city in PDY 5. Albeit small in number, it is of high signi cance for breaking transmission.Rapid reporting and information of geolocation have been the strength of malaria control system in Zanzibar for a decade resulting in low transmission of malarial cases [6, 17,18 ].However, falciparum malaria still remains a problem in Zanzibar and Swaziland [19,20].
Mosquito control Activities: Use of IT brought about behavioural change among health care providers in the eld.A shift from manual documentation to IT system with automation ensured appropriate eld response including mosquito control measures.Transmission of malaria depends on Ro which inturn is determined by patient factors (PR or parasite ratio) and mosquito behaviour related to entomological inoculation rate (EIR) [21].Therefore, it is imperative to prevent transmission of parasite from malaria patient to mosquito.An infected mosquito can continue to transmit sporozoites to many healthy individuals for a longer period.The risk of transmission to surrounding population can be minimized with anti-adult or anti-larval measures in houses around the residence of active malaria cases.Effective source reduction management happened over 5 years with gradual reduction of active breeding habitats (Fig. 4).Measures to reduce breeding and spread are important public health measures in malaria elimination operation [9].
Local Strategies: Eighteen months after digitization an administrative decision was taken to utilize services of MPWs for non-malarial (civic body's) work resulting in diminishing e ciency in the eld.
Although the community visits increased by manifold during PDY 3, it was not translated to effective vector control measures and collection of smears by active surveillance reduced from 4.61 per case (PDY 2) to 2.8 per case (PDY 3).This resulted in a slump in the work and non-reduction of malarial incidences during PDY 3. A surge in the number of cases was observed in April-May 2017 which led to increase in malarial indices.To counter this ine ciency, Complete Malaria Elimination Teams (CMETs) were formed at district malaria unit in June 2018.The CMETs conducted Active Surveillance Around Reported Case (ASARC) along with anti-vector activities in the locality.Subsequent to CMETs surveillance reduction of cases was observed in the 4th year.During PDY 5 there had an unprecedented pandemic of COVID-19 and entire nation was under lockdown, public health system was engaged in ghting this new disease.However, the CMETs continued carrying out the visits to malarial houses.This activity is probabaly the mainreason for reduction of malaria during PDY 5.
The Global effort of malaria control is in line with the 'One World One health' concept, but then a globally de ned 'one-size-ts-all' malaria control strategy would be ine cient in endemic areas [22,23].Digitization did aid in local modi cation of strategies.During analysis of new cases, clusters of new cases within a short period of one week, within a de ned geographical area were identi ed (hot-spots) and strategically separate programmes were carried out.One such endeavour was targeted for labourers and daily wage earners.Generally, malaria clinics are open from 9 AM and to 5 PM which were underutilized as it was not convenient for the manual labourers and daily wage earners and low socioeconomic class, as they were engaged in their vocation and income generation activities during that time.Hence, a mobile 24 × 7 clinic using a van and healthcare workers was introduced so that it could visit various places and could also be sent to the site if there was a phone call made to the central malaria helpline number.This helped in not only enhancing the diagnosis, but also treatment and prompt reporting of malaria inmigrant population.
Mappimg and Risk categorization: It has been a long-standing concern for epidimeiologists to quantify and stratify risk for malaria.Risk categorization for strategies and programme management is the key to success of malaria eradication and elimination [21].MCS captured data on realtimebasis for spatial risk classi cation.Geographical high-risk categorization is based on API and 43 such wards recorded reduction of incidence by 80% and above over 5 years.Several wards converted from a high API red zone to a lesser API green or yellow zones (Fig. 3).Swaziland adopted Immediate Disease Noti cation System (IDNS) where toll free number is given to report malaria incidence.Risk prediction model was applied for malaria elimination [6,20].There is a role to understand geographic trends for planning the strategies at micro-level and further research and review are warranted.Moreover, it may be worthwhile to look at the sociodemographic characteristics of people in these areas as well as the activities like construction and migration or travel [22].
Covid Pandemic and malaria: In PDY5 emergedas major public health challenge and disrupted malaria control programme in general.While Februaury 2020 was mainly focused on preparation to plan strategies to control COVID-19, Nationwide lockdown recorded a derease in number of cases of all diseases as the hospitals were converted to Covid-19 facilities and care centres.Diversion of healthcare workforce towards COVID-19 management, total lockdown of entire country, non-availability of transportation, closure or limited working hours of health facilities hampered anti-malarial activities for a short period of 5 to 6 months.However, Active surveillance (ASARC) continued uninterrupted, routine house visits were less but closure of cases were continued and was fairly good Accountability: Strenghtening of eld work force and capacity building is essential in any public health programme.IT adoption did empower the eld workers and it also helped in data-based micromanagement by the administrators as well as eld workers.A bidirectional accountability was also observed i.e. from eld force to administration and vice-versa.ASARC, time bound action in the geographical area surrounding the new malarial case, continuity in control measures especially during low transmission period (non-monsoon period).The necessity of closing the case on day 14, and its measure re ects functional accountability by eld work force.Closure of cases steadily increased and contributed to reduction of malaria incidence.There were delays in closure of cases as a result of multiple factors like non-working days, non-availability of the patient upon visit to home, migration, etc.Nevertheless, over 90% cases were tracked and closed subsequently.An inverse relation between closure and malarial incidence was observed (Fig. 2).Hence, the function of 'close a case' acertained complete treatment and parasite clearance therby contributing to transmission control.
Future scope: Five-year data indicates that technology has a major role to play in evaluating epidemiology of malaria as well as malariogenic factors.Learning from MCS application should help to upgrade functions, incorporate analytical and predictive output, warning and alarm systems for compliance in the eld.Ideally there is a need to design IT system driven eld response for both treatment and vector control analytics and predictions.Since most control measures are similar for all vector borne diseases they be brought under the perviw of independent system to manage vector borne diseases or even infectious diseases.
Malaria elimination by 2030 in India is being envisaged.An excellent information system should be at the core of malaria elimination programmes to ensure that all cases are detected and responded to in an effective and timely manner.Investment in robust, response-focused systems is essential to achieve malaria elimination.The operational manual elaborates the strategies.However, these strategies need be structured with 'time-bound' interventions.Figure 6 provides functional description of MCS for good micromanagement which is essential for malaria elimination [5,6].All micromanagement data regarding treatment and vector control measures can be quanti ed in relation to 'time frames' for each action.

Conclusion
Surveillance is the backbone of an efective system to support malaria elimination.Poor surveillance data will prevent countries from monitoring progress towards elimination.A standardized surveillance system landscaping was conducted in 16 countries between 2015 and 2017 in collaboration with governmental malaria programmes.The landscaping analysis identi ed multiple gaps in current malaria surveillance systems [24].The smart surveillance, IT system driven reporting and eld responses, creation of big data

Figures
Figures

Figure 1 Monthly
Figure 1

Table 2
Ward-level malaria cases in Mangaluru post-digitization and cumulative reduction.
In June 2018 (PDY 3), Comprehensive Malaria Elimination Teams (CMETs) were formed to visit reported cases of malaria and to carry out sanitization of the area subsequent to administrative decision to utilize services of multipurpose workers (MPWs) for non-malarial work.Consequent to functionining of CMETs resultant gures for incidence of malaria in PDY 2018-19 showed marked reduction.

Table 3
Type of health facilities and malarial case reports in Mangaluru

Table 4
Malaria incidence data, contact smears and malriometric indices in Mangaluru Transmission cycle is effectively broken if eld interventions are carried out in the rst 7 days of diagnosis.Transmission occurs locally around a reported case, and it is logical to implement effective vector control activity and measure this activity simultaneously.Smart surveillance is able to measure and micromanage control measures for designing local strategies.