Engaging community health workers, technology, and youth in the COVID-19 response with concurrent critical care capacity building: A protocol for an integrated community and health system intervention to reduce mortality related to COVID-19 infection in Western Kenya [version 1; peer review: awaiting peer review]

Globally, the number of COVID-19 infections is approaching 63 million; more than 1 million individuals have lost their lives. In Kenya, the number of infections has surpassed 80,000 and 1469 people have lost their lives. In Kenya, the community health strategy has been used to deliver essential health services since 2007. Furthermore, the population in Kenya is young (the median age is 21 years old) and Kenya is recognized as a technology hub in the East African region. Community-based health care, youth, and technology, are assets within the Kenyan context that can be leveraged to respond to the COVID-19 pandemic with concurrent strengthening of the critical care capacity at the health system level. This is a quasi-experimental study with quantitative and qualitative methods of data collection to complete a baseline assessment of community health unit and health facility service readiness in the study site of Siaya County in western Kenya. Following the baseline assessment, service ready community health units and health facilities with oxygen capacity will form intervention groups. At the community level, the intervention will consist of training youth, community health assistants and community health workers in screening, case detection, prevention, management and referral of COVID-19 cases Open Peer Review Reviewer Status AWAITING PEER REVIEW Any reports and responses or comments on the article can be found at the end of the article. Page 1 of 11 Wellcome Open Research 2021, 6:15 Last updated: 28 JAN 2021


Introduction
Globally, the number of COVID-19 infections is approaching 34 million; more than 900,000 individuals have lost their lives 1 . In Kenya, the number of infections has surpassed 20,000 and 388 people have lost their lives 1 . In Kenya, the community health strategy has been used to deliver essential health services since 2007; community health workers are a core component of the strategy 2,3 .

Community health strategy
The community approach would be built on the community health strategy (CHS), a flagship government policy that is based on the concept of social capital that is uniquely strong in sub-Saharan Africa (SSA) 4,5 . Social capital is founded on the fabric of trust, shared values and understanding that allows people to work together towards collective outcomes and common goals 6 . According to the conceptualization of social capital includes elements of trust, mutuality and reciprocity, enabling people to be each other's keeper which is critical always but particularly so during emergencies, such as we are in now 7 . In solidarity, people ensure that everyone has access to social structures and vital resources critical for self-protection and survival, and which is the backbone of the proposed integrated community-based case detection, containment, and care 8 .
The CHS policy also involves shifting some tasks to the community level workforce. Task shifting is allocation of tasks in health systems delivery to the least costly worker capable of doing it reliably without compromising quality [9][10][11] . Another critical element in the proposed strategy is the strengthening of the community-based health information system with mobile health technology. Health Information Systems (HISs) offer opportunities to inform health decision making at all levels of the health system which has proved of critical importance during the current pandemic 12 . It is critical to design information systems that inform timely decision making in a situation in which the COVID-19 response must be continuously adjusted to the local context and manifestations of COVID-19 in the Kenyan context. This adjustment by decision makers must be informed by accurate, current and timely data to ensure timely action and accountability based on evidence-based decision making 13 . We are proposing that this process starts at the community health unit (CHU), but linked to the health centre and sub-county levels. Issues emerging from the data gathered and analyzed at every level are fed into decision making applying the principles of evaluative thinking [13][14][15] , leading to evidence based actions announced on a daily basis, and providing feedback on what must be done by the public to contribute effectively to flattening the curve by reducing the occurrence of new cases.
In this strategy we hypothesize that community-based action can be strengthened by mobile technology. in their systematic review assessed strategies, findings and quality of evidence regarding use of mobile technology to improve health 16 . Ippoliti and L'Engle found that mHealth is a promising approach in global health interventions; their systematic review confirmed that digital media and web based platforms had altered the communication landscape especially for disease prevention and risk reduction 17 .
Since its adoption in Kenya, the CHS in Kenya, has built the capacity of communities to examine, plan, and implement health-related issues that affect their wellbeing 2,3 . The CHS is well embedded in the Kenyan health system 2,3 . It is supported and financed by the Ministry of Health. Therefore, CHS can be leveraged and included in the COVID-19 response. Furthermore, it can be enhanced by youth and technology to accelerate the COVID-19 response in Western Kenya.

CHS in pandemics
Boyce and Katz, in their review of the role of community health workers (CHWs) in pandemics, concluded that CHWs' roles could be expanded 18 . Specifically, CHWs could act as community-level educators, organizers, and mobilizers during infectious disease outbreaks; contribute to syndromic disease surveillance systems while completing routine activities; and complete medical tasks unrelated to the infectious disease outbreak to fill health service gaps during or following the outbreak.

Youth
The population in Kenya is young (the median age is 21 years old) and Kenya is recognized as a technology hub in the East African region 19,20 . To our knowledge youth have not been engaged in health interventions during pandemics.

Rationale
Community-based health care, youth, and technology, are assets within the Kenyan context that can be leveraged to respond to the COVID-19 pandemic with concurrent strengthening of the critical care capacity at the health system level.

Protocol
Central research question What is the effectiveness of an integrated approach, with technology and youth supported community based action and critical care capacity building, in enhancing increased case detection, more precise isolation/quarantine and improved community and hospital based case management of COVID-19 cases in Western Kenya?

Objective
To evaluate the effectiveness of an integrated approach, with technology and youth supported community based action and critical care capacity building, in enhancing case detection, more precise isolation/quarantine and improved community and hospital based case management of COVID-19 cases in Western Kenya.

Study design
This will be a study with a quasi-experimental design with quantitative and qualitative components ( Figure 1). There will be 3 comparative components in our design which will allow us to attribute outcomes observed to the intervention: (1) Pre-intervention and post intervention comparisons of outcome measures.
(3) Comparison between the intervention county and similar non-intervention county that is geographically distant from Siaya County. This study will be a collaborative research intervention involving policy makers, managers, service providers and researchers.

Setting
The intervention will take place in Siaya, a rural county in Western Kenya with a population of 993,183.

Study population (inclusion criteria)
Youth aged 18 to 30 years old who are interested in contributing to the COVID-19 response.
CHW: community health volunteers and community health assistants (CHAs) linked to the Siaya Ministry of Health.
Clinicians: Nurses, clinical officers, and medical officers working in health facilities in Siaya.
Households: Households in Siaya County.

Exclusion criteria
Populations and households outside of Siaya, individuals older than 30 or younger than 18 years old, non-community health workers, non-clinicians.

Baseline readiness assessment
We will use mixed methods with quantitative and qualitative components to complete a baseline assessment of community and health facility service readiness; as described by Creswell (2014); using standard community health unit and health facility assessment tools (see Extended data 21 ).
The sampling frames of the study will be the list of all public and private health facilities, and the list of CHUs in each sub-county which will be obtained from the sub-County Health Management Teams.
Baseline knowledge assessment of study participants At the community level, we will assess knowledge of COVID-19 as a disease. We will assess knowledge of symptoms, signs (including warning signs), basic management, and indications for transfer to a higher level of care for severe cases. At the health system level, we will assess knowledge of the pathophysiology of COVID-19, diagnosis, and treatment of moderate and severe COVID-19. The assessment will be in multiple choice questions format. Questions will cover material that is well established.

Intervention
Following the baseline assessment, functional community health units and facilities with oxygen capacity will form intervention groups. Non-functional community health units and health facilities that choose not to participate in the training will form control groups. Given the ongoing pandemic, community health units that later achieve readiness will be eligible to participate in the intervention as soon as they achieve readiness.

Community intervention.
At the community level, using digitized tools, the intervention will consist of training youth, CHAs and CHWs in screening, case detection/contact tracing, prevention, management and referral of COVID-19 cases with maintenance of essential health services (curriculum in Extended data 22 ). Furthermore, CHAs and lead CHW will be equipped with infrared thermometers and pulse oximeters. Training will take place in person and remotely and will use World Health Organization and Ministry of Health tools developed for the COVID-19 response 23,24 .
Specifically, we will train CHAs to train CHWs to improve their knowledge and skills regarding COVID-19. Trained CHWs will carry out household screening for fever using non-contact thermometers. In this process they will ensure community infection prevention and control measures (IPC) with handwashing, physical distancing, and mask wearing. They will isolate households with fever and respiratory symptoms. They will be trained to rule out malaria using rapid diagnostic tests (RDT). Cases with fever and respiratory symptoms will then be tested by the Siaya Rapid Response Team consisting of health personnel from Siaya Referral Hospital 25 . Those testing positive will be isolated in their own houses for 14 days under the care of CHWs. Those with severe disease will be referred to higher levels of care with oxygen capacity. Screening activities will be supported by youth using digital tools (Commcare).

Health facility intervention.
At the health facility level, the intervention will consist of training health care workers in basic critical care and caring for severe COVID-19 patients with maintenance of essential health services using the World Health Organization and Kenya Ministry of Health tools 26 . Furthermore, resources such oxygen tanks and anesthesia machines in unused operating room theaters will be repurposed to build critical care capacity. Health care workers will be equipped with pulse oximeters to improve their diagnostic and monitoring capacity based on the World Health Organization short course on the critical care management of Severe Acute Respiratory Infection (SARI). Any patient with evidence of organ failure (in this case respiratory failure) will be defined as critical. We will train staff (particularly nurses) in providing supportive critical care adapted to the resources available. Maneuvers such as proning and providing supplemental oxygen are now increasingly being recommended. Thus, critical care provided will focus on providing adequate oxygen, hydration, nutrition, and frequent monitoring of vital signs.

Digital tools for data collection
The screening household behavior and fever survey, the baseline assessment tools and monitoring and evaluation of key indicators will be administered using the digitized version on Commcare. The survey assesses household handwashing behaviors, presence of fever, health seeking behavior, and household economic activity. Commcare is a well-established digital tool that has received security clearance from entities including the Centers for Disease Control. Data will be stored in the cloud and downloaded as needed for cleaning and analysis.
Africa COVID-19 screen Africa COVID-19 Screen has been developed by a young team of software developers using flutter for the framework, dart for programming language, firebase for the backend, and Git/GitHub for version control, and vscode for code editing. It is based on the algorithm included as part of the household survey 27 and can be tested using the following link: https:// covid19-fb626.firebaseapp.com/#/. The algorithm assigns points based on yes or no answers. The total score is calculated and a risk group is assigned. Based on the risk group, a recommendation is given, and will be communicated to the head of the household immediately.
Each CHW will be paired with a youth who will assist in completing the digital tools. We therefore leverage the CHW's knowledge of the community and the youth's tech savviness. All cases with fever will be isolated, and based on the severity of the disease home management will be pursued. Isolated cases will be registered and followed for 2 weeks for signs of progression and need for transfer. Signs for transfer include respiratory distress and desaturation.

Study outcomes
Definition of outcome measures (Table 1): Deaths: number of deaths following the diagnosis of COVID-19.
The composite outcome measure is COVID-19 related mortality at community and health facility levels. Poor screening, testing, isolation, referral, and case management will lead to more deaths. If the intervention is effective, it will improve screening, testing, isolation, referral, and case management leading to lives saved (Figure 2).

Data collection
Appropriate ethical approval will be obtained before data collection. Data will be collected using digitized tools including the household survey, Africa Covid Screen, CHU service readiness tool, CHA, CHW and health worker knowledge exams, and digitized spreadsheets for key project indicators at CHUs and hospital levels ( Table 2).
Data collection will be carried out throughout from baseline, midline to endline phases of project implementation. Data collection will be performed by a dedicated trained research team. Personal protective wear will be provided during data collection.
Youth will support CHWs to complete the digital tools.
Quantitative data 1. Health facility service availability and readiness assessment 28 . This will be carried out at baseline and end-line by trained health records officers and clerks using a standard World Health Organization (WHO) tool to assess the health services that are available in each selected sub-county to respond to the COVID-19 pandemic. The tool will be digitized. 2. Rapid assessment of CHU. This will be carried out at baseline and end-line by trained CHAs using a standard tool (see Extended data 29 ) to determine their readiness for the required COVID-19 community-based response activities. The tool will also be digitized.
3. Household screening for morbidity, and COVID-19 relevant Knowledge and Practices. We will train CHAs, and they will subsequently train CHWs to carry out baseline and end-line screening of households under their responsibility using the tool that CHWs routinely use to register all households under their care, Ministry of Health (MOH) 513 (Ministry of Health, 2014) but with COVID-19 items. The CHW and youth will interview either the consenting household head, the spouse present, or any adult member of the household, aged 18 years and above. The CHW will explain the screening 4. Household follow-up data collection. Households found to have a member with fever will be reported to the CHA and to the rapid response team for further evaluation and testing if they satisfy case definition. Otherwise the CHWs with the support of CHA will carry out weekly visits with the support of CHAs to record temperature, oxygen tension to monitor disease progression or improvement in accordance with MOH guidance 24 .
Qualitative data. We will use continuous and data driven theoretical sampling to determine additional data sources to support an emerging theory [31][32][33] . We will use purposive sampling to select respondents. We will approach participants by telephone and face to face. The principal investigator and co-investigators will carry out focused group discussions (FGD) using a Discussion Guide (available as Extended data 34 ) to capture peoples' past and current experiences with epidemics, their perspectives on community based quarantine, and how to manage visitors in times of epidemics such as COVID-19 applying a narrative study design 35 .
Credentials of the interviewers: MD, MPH (first author); MD, MPH, PhD (second author); both researchers are trained in public health. The second author has more than 40 years' experience working with communities in Western Kenya (since 1979). The participants know about the second author's previous work in the community; participants have worked with the second author on multiple projects since 1979.
A focus group discussion will be carried out in each subcounty. Because of the COVID-19 pandemic each discussion will not have more than five participants. We will record the number of participants who refuse to participate and the reason for refusal. The data will be collected outside in a sports field for example to reduce COVID-19 transmission risks. In line with social gathering rules, non-participants will not be present. We will describe characteristics of participants including age, gender, and occupation/community role. A discussion guide is provided as Extended data 34 , to capture peoples' past and current experiences with epidemics and their perspectives.
The qualitative data intends to: • Reveal the experiences of the participants with regard to disease outbreaks within their community in the past, how they were handled, and any roles they played.
• Give insight into the views and beliefs of individual participants in relation to disease control, and containment, sensitive issues such as quarantine, isolation, and handling visitors from outside the community.
• Elicit recommendations to enhance compliance with IPC regulations, reduction of stigma, and coping with stress and inadequate access to basic essentials. The interviews will be both documented and taped. Consent will be asked for both for interviewing and for being taped. Two interviewers will manage each focused group discussion session so that one will conduct the interview freely while the second will tape the conversation as well as take notes.
Each focus discussion group will number five or fewer people including the moderators. Discussions will take place outside in a field (e.g. football field) with ample space and ventilation. Social distance will be maintained. In addition, all participants will undergo temperature checks before the beginning of the exercise. All those with fever or other symptoms will be sent home and will be asked to join a future focused group discussion. All 14 participants will wear a mask at all times during the discussion.
Discussions will aim to achieve data saturation; data saturation will be determined by the non-emergence of new themes. Transcripts will be returned to participants for comments and corrections.
We plan to repeat the interviews during the endline assessment. We plan to use audio recording of discussions. Field notes will be taken during focus group discussions. Discussions will take place for 60 minutes to reduce the risk of exposure to COVID-19. Data saturation will be determined by the non-emergence of new themes. Transcripts will be returned to participants for comments and corrections. Two data coders will code the data. A coding tree will be developed after the initial two focus group discussions. Themes will be derived from the data. Discussions will be recorded and later transcribed in word processing. Following transcription, participants will get the opportunity to review and correct the transcription. Participant quotations will be used to illustrate major themes. We will check for consistency between data presented and the findings. We will present major themes in the findings. We will discuss minor themes in addition to the major themes.

Safety during household visits
To minimize exposure to COVID-19, during household visits, CHWs, research assistants and household members will maintain a distance of 2m, everyone will wear a mask at all times, and the visit will be conducted outside. All will wash their hands or use hand sanitizer at the beginning of the visit and at the end of the visit. When visiting a household with fever, CHWs and research assistants will be provided with proper PPE including masks. All team members will self isolate and will be tested for COVID-19 if they present any symptoms of COVID-19.

Data analysis Sample size calculations.
To ensure adequacy of sample size that is representative of households in intervention and control CHUs we will use the formula proposed by Israel 36 .
Where n is the required sample size, N is the total population in each sub-county, averaging 200,000 and e is the level of precision of 5% (0.05). Giving a required sample of 400 households per sub-county, 200 from intervention and 200 from control sub-counties or 10 households from each CHU in the study area, a total sample size of 1,600 households in the four participating sub-counties.
Quantitative data analysis: Using the digital tools at the community and health system levels, we will count the number of households screened, the number of cases isolated, the number of tests performed, and the number of cases referred to higher levels of care. In addition, we will count the number of deaths. These counts will be done on a weekly basis. At the hospital level, we will count the number of cases tested, treated, and the number of deaths related to COVID-19. These counts will be performed on a weekly basis by a dedicated team, and results will be made available to the county ministry of health team for timely input into decision making. Data collected will be compared at baseline, midline, and endline phases of the project. We will calculate differences in means for continuous variables and calculate relative risks for categorical variables.
The main quantitative outcome measure will be COVID-19-related deaths at the community and hospital levels. Poor screening, testing, isolation, referral, and case management will lead to more deaths. If the intervention is effective it will improve screening, testing, isolation, referral, and case management leading to reductions in mortality at community and health system levels.
Qualitative data analysis 33 : Qualitative data will be collected through focus group discussions to determine lessons learnt from the proposed intervention; to better prepare for the next pandemic. We will perform the focus group discussion at the beginning of the intervention to highlight any preliminary issues that should be taken into account during the intervention. Furthermore, we will conduct focus group discussion during the monitoring and evaluation phases (to course correct as necessary when issues arise), and at the end of the intervention at 3 months, to verify provisional study results and highlight reasons behind observed quantitative data. During discussions we will seek to understand how health workers and community members understand COVID-19. How do they define their role in the pandemic; how do they want to address the pandemic.
What are the needs, challenges, and priorities? What inherent resources do they have that can be leveraged? What makes them feel vulnerable as they face the pandemic. Following these discussions, we will collate these discussions and identify themes that need to be addressed as we proceed with implementation. As we implement, focus group discussions will seek to quickly identify challenges with implementation that need to be addressed.
We will compare pre-intervention and post intervention outcome measures. We will compare intervention and non-intervention community health units, and lastly, we will compare Siaya to a similar non neighboring county where the intervention was not implemented. We will use SPSS/STATA programs for statistical analyses and calculate differences in means for continuous variables and calculate relative risks for categorical variables. To determine statistical significance between implementation and control sites; and differences between pre-and post-implementation outcome measures, we will use the paired Student's t and Chi square tests; p values <0.05 will be considered statistically significant.
Cost effectiveness analysis. We will use the competing method, given that the implementation and control sites are distinct and separate. We will calculate incremental cost effectiveness ratios. The intervention is the deployment of the community health worker, youth, and digital technology unit at the community level, with concurrent critical care capacity building at the hospital level. We will use the health services perspective. Inputs for cost effective calculations will include: human resources costs, equipment (pulse oximeters, thermometers, and handheld devices, and PPE). We will include programme costs of implementing the project (i.e. training costs, development of tools and education materials, and monitoring and evaluation costs). We will obtain data using hospital records and digitized data from CHWs and youth. Finally, we will add societal costs related to accessing COVID-19 care (e.g cost of transport). Effectiveness will be measured based on disability life years averted to complete the calculation of the incremental cost effectiveness ratio and determine the cost effectiveness of the intervention. The incremental cost effectiveness ratio will be calculated by dividing the difference between intervention cost and the counterfactual (pre-intervention) per site divided by their respective number of disability-adjusted life-years averted.
The initial study duration is 3 months, from May 18th 2020 to August 18th 2020 with possible extension for continued monitoring and evaluation of key indicators. Data management. Data will be kept under secure storage under lock and key, as well as password protected and will be accessible only to authorised key personnel engaged in the study. Analyzed data will be disseminated for decisionmaking. No personal identifying information will be recorded on questionnaires that can be used to link recorded information to a specific individual. The respondent will be assured full confidentiality of all responses and the interviewers will take steps to ensure privacy during interviews. The completed questionnaires on the interviewers' mobile devices will be password protected and deleted once they have been uploaded to the central database which will also be password protected.
Data monitoring committee. An independent data monitoring committee will be formed and will meet on a biweekly basis to review activities of the protocol and data. An interim analysis will be performed at the midway point of the timeline. Results will be made available to the data monitoring committee who will take the final decision on whether to terminate the trial.
Harms. Adverse events can be reported to the ethics committee and the sponsor, and the adverse event will be managed without imposing any financial burden on research participants. A summary of adverse events will be presented to the data monitoring committee on a biweekly basis.
Auditing. An independent review of core processes and documents will be performed during 3 months allocated to the project.

Discussion
This study would be the first study to evaluate the effectiveness of an integrated approach in preparing for and implementing a robust response to an ongoing COVID-19 pandemic. We would be arming communities with knowledge and preventive measures that have been shown to contain COVID-19 infection, and concurrently we will be strengthening the health system with health care worker capacity building in basic critical care provision. We will build evidence for scalable integrated community and health system interventions during pandemics in resource constrained settings.
The approach is innovative because of its engagement of youth and use of technology. In addition, it rapidly builds critical care capacity by training of health care workers in facilities with oxygen capacity.
There are several limitations to the study: the ongoing pandemic may require modifications of components of the intervention, attrition of participants is a possibility, technology tools may be limited by the lack of electricity and poor network connections in rural parts of Kenya. Using non-functional CHUs as controls could introduce confounding bias; however, we will compare known and unknown attributes between the 2 groups before the intervention. Subsequently, we will adjust for any differences between the 2 groups (by performing subgroup analyses for stratum specific relative risks) to obtain a true effect measure following the intervention.
We anticipate challenges in the availability of diagnostics (particularly laboratory testing kits), and the intervention will rely heavily on clinical signs and symptoms.

Underlying data
No underlying data are associated with this article.

Extended data
Harvard Dataverse: Engaging community health workers, technology, and youth in the COVID-19 response with concurrent critical care capacity building: A protocol for an integrated community and health system intervention to reduce mortality related to COVID-19 infection in Western Kenya. https://doi. org/10.7910/DVN/PUTCR8 27 .
This project contains the household survey tool.
Harvard Dataverse: Engaging community health workers, technology, and youth in the COVID-19 response with concurrent critical care capacity building: A protocol for an integrated community and health system intervention to reduce mortality related to COVID-19 infection in Western Kenya. This project contains the community unit rapid assessment checklist.
Harvard Dataverse: Engaging community health workers, technology, and youth in the COVID-19 response with concurrent critical care capacity building: A protocol for an integrated community and health system intervention to reduce mortality related to COVID-19 infection in Western Kenya. https://doi. org/10.7910/DVN/5QJICK 34 .
This project contains the focus group discussion guide.
Harvard Dataverse: Engaging community health workers, technology, and youth in the COVID-19 response with concurrent critical care capacity building: A protocol for an integrated community and health system intervention to reduce mortality related to COVID-19 infection in Western Kenya. https://doi. org/10.7910/DVN/1BP8DT 30 . This project contains model consent and assent forms.
Harvard Dataverse: Engaging community health workers, technology, and youth in the COVID-19 response with concurrent critical care capacity building: A protocol for an integrated community and health system intervention to reduce mortality related to COVID-19 infection in Western Kenya. https://doi. org/10.7910/DVN/WKO3TY 22 .
This project contains the training curriculum.

Reporting guidelines
Harvard Dataverse: SPIRIT checklist for 'Engaging community health workers, technology, and youth in the COVID-19 response with concurrent critical care capacity building: A protocol for an integrated community and health system intervention to reduce mortality related to COVID-19 infection in Western Kenya'. https://doi.org/10.7910/DVN/BPLZP3 37 .
Extended data and the reporting guidelines checklist are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).