Project Lead(s): Matthew Wiens
Children hospitalized for serious infections in Africa have a very high post-discharge mortality rate.
Doctors and parents are often unaware of this period of high vulnerability, and are poorly equipped to identify or handle recurrent illness.
The Post-Discharge Survival Project sought to improve the outcomes of children following hospital discharge through two stages. Firstly, through the development and testing of a mobile application to allow front-line health workers to easily and rapidly identify children at high risk of post-discharge mortality. Secondly, through the development and implementation of a discharge kit, to improve health outcomes following discharge.
Utilizing their prior research on post-discharge mortality in Uganda, the research team developed predictive models, so that children could be classified into low- and high-risk groups prior to hospital discharge. Improved access to mobile phone technology will enable any health worker to easily access and use these complex models, to ensure that children at a high risk of mortality can receive life-saving care following discharge.
The PAediatric Risk Assessment (PARA) mobile application is an easy-to-use mobile health tool to aid healthcare workers in resource-limited settings to identify hospitalized children at high risk of death, both in hospital and after discharge.
In partnership with LionsGate Technologies and the University of British Columbia Pediatric Anesthesia Research Team, the PARA app underwent testing in Uganda with 30 front-line health workers (doctors, nurses, clinical officers) for the purpose of evaluating user performance and improving design and function.
During the second stage, a discharge kit prototype composed of two components was developed. The first component was educational, whereby a nurse – using a storyboard-style, laminated card – explained the child’s vulnerability following discharge and reinforced basic health information pertaining to hygiene, indicators of recurrent illness and appropriate health-seeking practices.
This was provided along with three simple incentives (a mosquito net, 1kg of soap and five sachets of oral rehydration salts) to reinforce the educational training. The second component was an individualized post-discharge referral to either a community health worker or a lower level health facility in the patient’s community. The referral served to ensure adequate follow-up during the vulnerable, post-discharge period. The discharge kit was evaluated in a hospital environment with 202 families to determine feasibility.
The study confirmed that children at risk of post-discharge mortality can be identified prior to discharge using the app.
The results of the discharge kit study effectively demonstrated the utility of this approach to improve post-discharge health outcomes in children. Of the children in the study, 85% successfully completed at least one post-discharge referral, and nearly 50% completed all three post-discharge referrals. This was of critical importance, since more than one-third of these visits resulted in outpatient treatment, admission or referral to a higher level of care. In fact, nearly half of the 22 re-admissions were directly as a result of the post-discharge referral. The majority of caregivers strongly agreed that the education provided at discharge improved their ability to care for their child.
When the results were compared to the prior post-discharge research done by the research team, the discharge kit resulted in the discharge kit being associated with an increase in health-seeking (from referrals or from a self-referral) from 30% to 90%. Re-admissions increased from 5% to 11%, suggesting improved health-seeking. Post-discharge mortality decreased from 3.5% to 2.5%, although this was not statistically significant.
Findings from the project have been distributed through conferences and a publication in JMIR mHealth and uHealth.
The project team intends to apply for additional funding to scale the use of the PARA app and discharge kit at a regional level, to refine the prediction models, and to conclusively prove the mortality benefit of this approach.