Evaluating a South African mobile application for healthcare professionals to improve diagnosis and notification of pesticide poisonings

Worldwide approximately three million acute pesticide poisoning (APP) cases occur yearly [1] and yet this is a neglected public health concern. In particular, APPs are a widespread problem in World Bank classified low- and middle-income countries (LMICs), including South Africa (SA) as an upper middle-income country. In the sub-Saharan Africa context, SA is the biggest user of pesticides in both occupational and non-occupational settings, where pesticide use is extensive and pesticide exposures are common [2‐4]. In SA, an APP is a notifiable medical condition under the National Health Act [5], and by law healthcare professionals (HCPs) are required to notify poisonings from any pesticide to the National Department of Health (NDOH) [2, 4]. Despite this, cases of poisoning continue to be under-reported. The assumption is that health professionals have received adequate training in medical school to diagnose and treat pesticide poisonings from all types of pesticides. The reality is that medical curricula are extremely dense and pesticide poisoning is not a priority area. What is needed are innovative methods for providing APP information to HCPs. To address this challenge, this study aimed to assess the use of mobile health (mHealth) application for APPs (i.e. the diagnosis and notification of APPs) to provide supplementary information to HCPs.

Notification of an APP is an important surveillance tool for controlling and decision making to reduce the harmful effects of pesticides, particularly from street pesticides (i.e., agricultural pesticides and other pesticides sold illegally mostly in communities with low socio-economic status for domestic pest control) [4]. The importance of this notification system is to alert local and national health services of the pesticides that are causing morbidity or mortality, and which require to take appropriate prevention responses. Furthermore, as a signatory to the Rotterdam Convention, South Africa is required to submit quarterly reports of pesticide poisonings. Many cases, however, go unreported leading to poor surveillance data. In SA, it is estimated that only 10 to 20% of hospitalized cases are reported [2]. A similar magnitude of under-reporting has also been observed in Tanzania where a study found that 78% of APP cases presented to health facilities go unreported [6]. This is often to limited information on pesticide poisoning leading to incorrect diagnosis and treatment, as in turn, under-reporting [4]. For example, many APP symptoms present as flu-like symptoms [7]. Furthermore, there is knowledge gap regarding classes of pesticides and reporting requirements. For example, only organophosphate-containing pesticides (OPs) tend to be reported, while other pesticide classes are not, such as pyrethroids, carbamates, organochlorines and rodenticides [8, 9].

Studies assessing the adequacy of HCPs training in both high-income countries and LMICs on environmental health risks have found that most HCPs have limited knowledge in diagnosing and managing environmental exposure problems, such as lead poisoning and APP [10‐14]. This is primarily because many medical schools offer little or no training in environmental health, as it is not considered a priority area in what are usually already full medical curricula [15, 16]. As a result, many HCPs do not conduct an environmental exposure history, which is key in making a diagnosis of APP [14]. Environmental history taking should be included when general history taking is conducted as part of a standard diagnostic process when a patient is first seen by an HCP. For example, the HCP could use the CH²OPD² mnemonic–that is; asking questions about community, home, hobbies, occupation, personal habits, diet and drugs [17] – to find out if, for example, the patient lives in a community with high use of street pesticides.

Another challenge, often faced by HCPs in South Africa, is the difficulty of linking APP to illegal or unlabelled street pesticides [18]. Many poor urban residents rely on cheap, easily accessible, and often highly toxic agricultural pesticides that have been decanted into unlabelled containers for domestic pest control [4]. When an exposed individual is taken for medical care, it becomes difficult for HCPs to diagnose poisonings from exposure to unlabelled products and complicates their ability to comply with notification requirements. Another challenge is that there is a scarcity of easily accessible resources to assist busy HCPs with APP diagnosis, treatment and notification [4].

mHealth applications present a unique platform to supplement medical education on environmental health topics generally, and to assist HCPs in identifying and notifying APP cases in particular. mHealth, the utilisation of mobile technology in health care, has the potential to support the provision of high-quality health care services and keep HCPs up to date, particularly in LMICs [19, 20]. This is due to high coverage of mobile devices. For example, in SA almost all HCPs have cell phones and therefore implementation of mobile application could possibly alleviate the overburdened health care delivery systems in these countries [21]. Therefore, this study provides an insight of how mHealth could support SA, where there is excessive use of pesticides and exposure with relatively well-developed health system and high cell phone ownership levels - but continuing underreporting of APPs.

In recent years, the use of mobile phones and other mobile devices by HCPs in their clinical practice has been increasing. This has led to the rapid development of medical-related mobile applications, including those for the management and monitoring of patients with different illnesses, clinical decision-making resources, and continuing medical education on specific topics, among others [22‐28]. Despite this development, there are limited studies that have evaluated mHealth tools targeted at HCPs [29, 30].

A key aspect in evaluating the usefulness of mHealth is understanding an HCP’s attitude towards the use of mobile applications. Several studies have examined the use of mobile applications by HCPs and have shown that generally they perceive mobile applications to be essential and effective tools to support their day-to-day work [31‐33]. For example, a study conducted in the United Kingdom (UK) evaluated a hospital-specific smartphone application (i.e., iTreat) used for antibiotic selection and clinical management of infections. The study found that most doctors felt the application was an effective tool because it reduced the need to refer to cumbersome clinical manuals, thus saving time during clinical rounds [32]. In low- and middle-income countries, such studies are needed to determine effective mHealth strategies (i.e. use of behavioural theories [21]) for large-scale use. Thus, for production of mHealth applications, developers will be able to access evidence-based information on the usefulness, as well as the disadvantages and/or advantages of the applications to improve on functionality [34]. To enhance South African HCPs’ capacity to diagnose and report APPs, the University of Cape Town (UCT)’s Division of Environmental Health, and Centre for Environmental and Occupational Health Research (CEOHR), together with other stakeholders (academic researchers, NGOs and the SA government), developed a Pesticide Notification Guideline (PNG) [4]. The guideline is composed of an algorithm that aims to help HCPs improve notification of APPs, particularly from street pesticides. Included in this guideline is a point chart that contains pictures of common unlabelled street pesticides for family caregivers or patients to indicate the pesticide product responsible for the poisoning.

Using behavioural theories (Theory of Planned Behaviour and Social Cognitive Theory), this study evaluated the PNG within the EM Guidance mobile application to improve the ability of HCPs to diagnose and reporting of APP cases. Furthermore, assessing the improvement of their understanding that APP is a notifiable medical condition. In addition, the study assessed the ways in which the PNG could be improved to enhance its effectiveness and applicability.

The findings of this study illustrate the important role of medical informatics and particularly mHealth application in this case, can play in addressing decision making around diagnosing and treating APP, as well as improving notification. This is a key opportunity particularly as limited training and support provided to medical students and practicing health professionals on APP. While most participants were aware that APP is a notifiable medical condition in SA, that different classes of pesticides that should be notified to the NDOH, most of the participants did not notify APPs. Previous studies in other low- and middle-income countries have identified possible reasons for HCPs not reporting notifiable diseases including busy HCPs having limited time, poor awareness about the prerequisite reporting forms and lack of printed forms [4, 6, 45]. Whereas, possible reasons identified in our study include poor recognition among HCPs of the value of notifying APPs to NDOH and that reporting is a time-consuming obligation. The mobile application, therefore, lends itself to providing an alert system for the need to report. Future mobile applications could include the notification form to simplify the process for reporting.

Our findings showed a significant association between the number of years of work experience of emergency medicine physicians and registrars in relation to knowledge of the legal requirement to report APP. This suggests that experienced emergency medicine physicians and registrars are more likely to report APP as it is expected that healthcare practitioners become more knowledgeable with experience. The question is whether mobile applications that include PNG could reduce the experience period. However, it was unclear why emergency medicine physicians and registrars with more than six years in practice were less knowledgeable on reporting APPs. Perhaps it was a result of notification fatigue, since there are nearly 40 notifiable conditions HCPs need to fill in notification forms for. By including the notification form within a mobile application, perhaps this notification fatigue would be reduced. There was also confusion amongst participants about which pesticide poisonings should be notified, as the current practice was to only report poisonings from organophosphates. Again, the mobile application provides the opportunity to provide accurate information on what needs to be reported, as well as aid in correct diagnosis. Furthermore, future studies are needed to explore why HCPs in general are not reporting APP cases even if they are aware that these must be notified and to evaluate whether mobile applications can address these reasons. This study’s findings suggest an urgent need for educational interventions among HCPs to create awareness and improve notification of APPs, and that there is an opportunity for mobile applications to address this and to provide continual reminders.

The indication is that mobile applications could be useful as continuous training tools in daily medical practice, particularly as most participants own smartphones. This finding is consistent with previous studies conducted in the United States (US) and the UK [32, 46]. The EM Guidance mobile application was one of the most (40%) commonly downloaded medical-related application amongst participants. APP is not only addressed by EM professionals and therefore, information such as the PNG should be added to applications targeting other medical specializations. It is important to note, however, that having a mobile application is not an indication of its use. There was low awareness of the existence of the PNG-application. Therefore, awareness is regarded as the first step that would increase the likelihood of individual use of a specific guideline [47]. Cabana et al. [48] highlight that lack of awareness is one of the factors that hinders the -implementation of guidelines among HCPs. This could explain one of the reasons why, so few participants had used the PNG-application, as its visibility and advertising of information was limited. Since the EM Guidance mobile application was developed in April 2014, there is a possibility that many participants have not yet discovered the guideline. This highlights the need to more proactively advertise the PNG and its relevance. Amongst the participants aware of the guideline, some had never used it as they had only recently downloaded the PNG-application. A similar finding related to the period of dissemination for a guideline was observed in the US, where only 34% of physicians were aware of the existence of the guideline on depression, one year after its publication [49]. Given that APP is a major public health problem, efforts are needed to create a widespread awareness of the PNG-application amongst HCPs through social media, conferences, continuing medical education publications, and programmes or seminars. Coupled with these efforts, APP recognition and reporting should be taught throughout medical undergraduate studies. This would reinforce and ensure that doctors learn how to diagnose and treat APP correctly and the importance of notifying APP to the NDOH. With high mobile ownership among study participants, it showed that PNG-application may support HCPs in reporting more APP cases. Improving the notification process, will not only encourage HCPs to diagnose and notify pesticide poisoning cases more consistently but will also enable NDOH to put measures in place to reduce and prevent APPs.

The Theory of Planned Behaviour (TPB) provided a useful framework to understand participants’ use of PNG-application. According to the TPB, intention to use this guideline or a mobile application in general is influenced by three major constructs: attitude of the HCP, the subjective norm (i.e. perceptions of the views of other HCPs on the adoption of the pesticide notification guideline) and the perceived behavioural control (i.e. HCP’s perception of how difficult or easy it is to use the PNG to diagnose and notify APPs) [37]. Attitude has been considered an important predictor of HCPs’ intentions to follow the recommendations within a guideline [37]. This was observed in our study, where six of the seven participants that used the PNG-application had a good attitude towards the pesticide notification guideline, indicating that the participants intend to continue using it. Similarly, Limbert and Lamb [50] found attitude to be the main prognosticator of intention to utilize the antibiotic guideline. The authors found that doctors had positive attitudes towards the antibiotic clinical guideline because it was useful, and as result they intended to continue using it.

In assessing whether the PNG-application would be used in clinical practice, participants’ self-efficacy was linked to their intention and confidence. Four of the seven participants who had used the PNG previously indicated it improved their confidence in reporting APP cases, particularly those from street or unlabelled pesticides. This finding about self-efficacy is consistent with a study conducted in UK, where Hrisos et al. [51] found that self-efficacy was related to general practitioners’ (GP) intention and confidence to adhere to guidelines in managing upper respiratory tract infection (URTI). Although the number of doctors who used the PNG was small, the indication is that doctors would be more confident to report APP with the support of this mobile application. However, three participants felt that the use of the guideline did not improve their confidence, for reasons that are unknown, as the question was not further queried. It is possible that these participants were familiar with the notification system process, and therefore found no significant change in their confidence in management and notification of APPs through the use of the mobile application. Further research with a larger sample size is needed to investigate these speculations.

The seven participants that used the PNG were comfortable with using the guideline on their mobile device during patient consultations or within the medical setting. In this study, “comfortable” referred to accessing the PNG within the mobile device application in front of patients or in the hospital or other clinical setting. This finding is in contrast with the UK study where participants felt it was unprofessional to use mobile devices during patient consultations or within the hospital setting [32]. The inconsistency in these results is likely due to the cultural concerns related to using a mobile medical application in the hospital, which could differ worldwide. It should be noted that the UK study findings are drawn from one hospital and should not be generalizable to other hospitals or clinical settings.

One reported disadvantage and barrier regarding the use of the PNG-application was that poor network signal reception makes loading of images difficult. This can affect HCPs when they are in the process of showing patients the images to identify which pesticides they were exposed to. The issue concerning mobile applications and slow download speed has also been described in another study [52]. Such an issue may be resolved by ensuring that hospitals have effective Wi-Fi and sufficient coverage to support mobile application use. Another option is to allow for download of the guideline to be used offline and not dependent on the internet each time.

The PNG is the first attempt to support notification and diagnosis of APP. As participants have indicated, there is room for improvement. One of the recommendations given by the PNG-application users was the need for more images or photos of both street and commercial pesticides, which would require more use of bandwidth. This indicates that images are clearer and more useful to HCPs, especially when faced with an APP case from an unlabelled pesticide product. Since mHealth applications are becoming common tools for accessing clinical materials [22], integrating more photos and clinical toxidrome information in the PNG could assist HCPs to promptly diagnose and treat patients suspected of being poisoned. Another suggested recommendation was having a clinical toxidrome in the PNG-application.

HCPs usually have busy schedules within a hectic hospital environment, such as an emergency department. As a result, they may not find time to document in a patient’s record that APP cases need to be notified to the health department. In such clinical practice environments, mobile applications are regarded as convenient tools to accomplish several tasks and provide HCPs with increased efficiencies [22]. Therefore, having a link in the PNG-application that HCPs may use to notify APPs was recommended and, may improve on reporting. Such recommendations are important to consider for increasing the effectiveness of this and future guidelines within mobile application to improve health-care service delivery in general, and to reduce pesticide poisoning specifically.

The high mobile smart phone ownership among participants indicated that the mHealth platform could provide support, decision making and continued learning to HCPs if mobile applications are well advertised. In addition, integrating mHealth interventions (e.g., EM Guidance mobile application) within the health system could facilitate improvement of HCPs’ reporting of relevant notifiable conditions. Hence governments and industry should support the development of high quality medical mobile applications, in collaboration with HCPs, to address public health challenges and provide information. Lastly, incorporating mobile health applications within healthcare systems and medical schools is a useful means of using technology to support over-burdened medical curriculums and onerous reporting requirements.