Statistical and Agent-based Modeling of Complex Microbial Systems: a Means for Understanding Enteric Disease Transmission Among Children in Urban Neighborhoods of Kenya
Statistical and Agent-based Modeling of Complex Microbial Systems: a Means for Understanding Enteric Disease Transmission Among Children in Urban Neighborhoods of Kenya
The objective of the PATHOME study is to (1) develop statistical and computational methods for examining a complex disease system of interactions between and amongst children, animals, the environment, and enteric pathogens and (2) build a virtual laboratory for predicting which social and environmental developmental improvements best prevents multi-pathogen transmission to infants in urbanizing areas of high disease burden countries. Investigators will characterize how social and environmental development of urban neighborhoods in disease endemic settings modifies the "enteric pathome", i.e. the microbial communities of viral, bacterial, and protozoan pathogens transmitted by human and animal feces in the environment to infants. They will measure the impact of societal development on pathogen transmission to infants by applying a One Health ecosystem-based approach to characterizing interactions between enteric pathome agents in the environment and their transmission via interactions between infants, caregivers (CGs), animals, and environmental materials across domestic and public spaces and climate conditions. Data-validated statistical and computational models can quantify pathogen-specific attributable risk of infection through multiple pathways, and the extent that these risks are due to pathogen interactions with each other and the environment. The overall study hypothesis is that joint modeling of enteric pathome agents across urban households and neighborhoods representing transitional improvements in societal development will show that development leads to lower pathogen-specific detection frequencies, and thus evolution of the pathome from complex to simple microbial community structures. By studying spatial scale, developed and underdeveloped neighborhoods, specific transmission pathways, and seasonality in this process, the conditions that lead to the greatest declines in enteric disease incidence can be identified. This virtual laboratory will be built upon extensive data collection in two different Kenyan cities, including household and neighborhood economic indicators, clinical, zoonotic, and environmental microbiology, behavioral observation, geotracking of humans and domestic animals, climate conditions, population density, and infant anthropometry. This initial virtual lab will provide an evidence-based tool for predicting effective urban interventions to control fecally-transmitted disease in cities globally undergoing epidemiological transitions in infectious disease.
STUDY OBJECTIVES
The overall hypothesis of the PATHOME Study is that joint modeling of enteric pathome agents across households and neighborhoods that represent contrasts in urban societal development will show that development leads to pathome evolution from complex to simple community structures, and thus lower detection frequencies for individual pathogen taxa. The study aims to (1) develop spatiotemporal and trajectory statistical models to understand the complex exposure risks for infants from the enteric pathome; (2) collect environmental, behavioral, spatial, economic, and microbial data to characterize the enteric pathome along pathways for disease diffusion and the intersection of humans and animals with these pathways; and (3) develop and validate agent-based models (ABMs) for predicting which social and environmental urban developmental interventions (e.g. animal penning, building latrines or drains, concrete floors) best prevent multipathogen transmission to infants in high disease burden countries using established Kenyan study sites as a model.
To fulfill the modeling and computational aims, comprehensive, interlinked data is needed that does not currently exist. For example, there is little data on child health and exposure to enteric pathogens in middle class populations in cities undergoing epidemiological transitions from infectious to non-communicable disease burden for contrasting against the more abundant data from children in low-income urban areas. Spatial data on where children may be exposed and for how long is a complete gap. Therefore, investigators will implement an empirical data collection study in two different cities in Kenya that collects a broad variety of socioeconomic, behavioral, spatial, child health, zoonotic vector health, environmental contamination, behavioral, and climate data in neighborhoods representing high and low infectious disease transmission potential. This protocol specifically describes the study design that will be applied to collect, summarize, and report infant health and exposure data by study group prior to further model development. The objective of the epidemiological study is to evaluate by city (Nairobi, Kisumu) and neighborhood socioeconomic development level (low versus middle income):
METHODS
Study Design. The PATHOME study is four-group prospective 14 day cohort study. Groups include a low-income and middle income neighborhood each in the cities of Kisumu and Nairobi. Age-eligible infant participants and their caregivers living in the catchment area of a participating Community Health Volunteer will be identified through random selection of a surveillance database, maintained by monthly updates from CHVs on age-eligible infants living in their catchment area. With an enrollment sample size of 248 infants and their caregivers, this equates to 124 participants in Nairobi (62 low-income, 62 middle-income) and the same in Kisumu. To ensure balance in age representation, selection of infants from the database for enrollment in the study will be performed by identification of an even number of infants across age groups of 0 to 3, 4 to 6, 7 to 9, and 10-12 month ages. After informed consent is completed, the study team will schedule a time to begin data collection. Enrollment will be conditional on procurement of an enrollment infant feces specimen. After enrollment, the team will follow up with infants and caregivers for fourteen days to complete data collection as described below. To limit bias from seasonal effects in observed enteric disease outcomes, enrollment in each city will span 12 consecutive months, implemented in Nairobi first (Nov 21 - Nov 22) followed by Kisumu (Jan 23 - Dec 23). Investigators will recruit three households a week, with a one week break in enrollment every fifth week when public domain contamination assessments will be conducted. This five week cycle will be repeated systematically up to 11 times or until sample size requirements are met.
Data Collection Timeline
On the day of enrollment (Day 1), the infant caregiver will be asked to respond to a 45 minute survey on her socioeconomic conditions, hygiene practices, and infant health. Then she will be given a 14 day pictorial diary and asked to mark any days in the next two weeks when the child experiences symptoms of diarrhea. Finally, she will be given a pack of diapers sized by the infant's weight, and asked to use them on the infant for the next week, saving any diapers containing infant feces. The team will schedule a time to return the next day to retrieve the diaper with feces and conduct the next exposure assessment phase of data collection.
On Day 2, enumerators return midday and spend about five hours with caregivers and the infants. First, enumerators collect the diaper with an infant stool sample, deemed to be the enrollment or pre-exposure feces specimen. Then, enumerators place a geotracking anklet on infants and confirm comfortable fit. Geotrackers are also placed on two domestic animals belonging to the household (separate IACUC approval for animal subjects research). Second, Otherwise they will collect the diaper with stool during observation. Third, they will provide the caregiver a new sterile infant toy, such as teething ring or rattle, to give to the infant for play. Fourth, enumerators will begin a five hour period of structured observation of the infant, noting any persons or things that the infant interacts with, and all places where the infant spends time. At the end of the visit, enumerators will collect a handrinse of the primary caregiver and a soil sample, and will ask the caregiver to continue to use the diapers to collect infant stool. Finally, a time will be scheduled to return the next day to collect the geotracking device, any stool samples, and the toy, and to perform another five hour observation. The stool and environmental samples will be transported to the lab for processing for microbial assays. The geotrackers remain as placed overnight, although if the caregiver chooses to remove it from the infant for sleeping, she is asked to place the device next to but out of reach of the infant and then replace the device when the infant wakes.
On Day 3, enumerators return early in the morning, retrieve a second feces specimen and perform a second five-hour block of structured observation. At the end of observation, they retrieve the geotracker from the infant (and domestic animals) and confirm the diarrhea calendar is up to date. The infant toy is retrieved and placed in a sterile bag on ice packs in a cooler for transport to the lab with the feces specimen for microbial testing. A new replacement toy for the caregiver's permanent keeping is provided. The infant's height and weight are recorded. Before departure, enumerators will remind caregivers to continue collecting infant feces and schedule another visit in about 48 hours to retrieve a third feces sample.
On Day 5, enumerators return to retrieve the third diaper with feces, check the diarrhea calendar for completion, and encourage the caregiver to continue using diapers to collect infant stool.
On Day 8, enumerators return to retrieve the fourth diaper with feces and check the diarrhea calendar for completion.
On Day 14, enumerators return to retrieve the fifth diaper with feces and the 14 day diarrhea calendar.
Data Collection Methods
PARTICIPANT ENROLLMENT AND SAFETY
Participant Enrollment. Eligible infants will be identified and recruited through quarterly surveillance reports by Community Health Volunteers (CHV), who are employees of the Kenyan Ministry of Health (MOH) in Kenya, on age and location of infants living in their demographic surveillance area. This area typically corresponds to one village within the neighborhood. Per their standard practice, they will approach all households with pregnant women or young children to ensure they receive health information and have support to access health care. When CHVs identify such households, they register them in a list provided to the Community Health Extension Worker, a formal nurse assigned to provide health care to a broader catchment. The African Population Health and Research Center (APHRC) will work with CHVs in the selected neighborhoods of Kisumu and Nairobi to monitor these records monthly and update information about age-eligible children. This is necessary because migration within, and in/out of these neighborhoods is common.
The APHRC research team will create two separate lists for the low income and middle income neighborhoods of a city, and randomly select 6 children from each list each month, equal to one to two infants per age group (0-3 months, 4-6 months, 7-9 months, 10-12 months). CHVs and the APHRC team will then approach households with caregivers of eligible children to verify their age via national identity card and to determine if they are interested in participating. The adult subject/primary caregiver will be informed about study activities, including an enrollment survey, household environmental sampling, geotracking of infants, and the use of structured observation of their interactions with the infant. The team will return the next day to obtain consent from a child's primary caregiver to participate in the study by reading forms to the caregiver in English, or in Swahili if they are not fluent in English, in the presence of their CHV as a witness. Consent forms will be signed by the caregiver with signature or thumb print and by their CHV witness. A copy of the consent form will be provided for the caregiver's permanent records. If consent is provided, a return visit will be scheduled to conduct the enrollment survey, structured observation, and sampling. If consent is not provided, the team will select a new potential infant of same age from the list.
Participant Safety. The primary risks associated with participation in the study are infant discomfort during wearing of the geotracker or during measurement of length. In order to mitigate the risks for infant discomfort during geotracking or anthropometry, clear Standard Operation Procedures (SOPs) and oversight mechansims will be followed to ensure the consistent use of good technique and the right of study participants to end geotracking or anthropometry data collection without penalty. Specifically for geotracking, the protocol involves a caregiver-led process for choosing where and how to place the tracker on the infant, and clear instruction to remove the device if the infant shows signs of distress related to wearing it. Caregivers are asked that if the device is removed, to place it on themselves or nearby but out of reach of the infant.
Caregivers may feel uncomfortable in answering some sensitive questions about their hygiene practices during data collection, or may be concerned about social perceptions or conflict associated with others knowing they are participating in a research study. Loss of privacy may lead to social or family conflict. This, however, presents a minimal risk to subjects. The SARS-CoV-2 pandemic may continue for months or years over the course of this study, and subjects may feel concerned about study staff transmitting the virus to their family. The privacy of subjects will be protected by (1) always collecting data in a private location where no other adults or older children are present that could overhear subject responses, (2) safely managing and storing all paper and electronic forms containing their information by collecting data on password protected tablets and limiting access to identifiable data to a limited number of qualified individuals, (3) destroying all forms with individual identifying information after five years' time, (4) training all staff in ethical practices for human subjects research and making the consequences for violating those practices clear, and (5) informing the subject that they are free to leave the study at any time, or to refuse to answer questions that make them uncomfortable.
Diarrheal disease may occur in infants during the enrollment period. In the event that a CHV or team member visits the home and observes the child in distress, or that caregivers report by phone or personal communication that the child is in distress, they will be instructed to refer caregivers to local health care centers and to report the adverse event to the clinical research team at APHRC and the University of Iowa. Distress may include diarrhea, severe malnutrition, or other symptoms. The goal will be to ensure that CHVs continue to provide standard care. The clinical research staff will then follow-up with caregivers in a timely fashion to offer further medical council. All observed or reported child illnesses will be documented by the APHRC team and discussed with the entire team during bimonthly conference calls.
In the event an adverse event is observed that appears to be linked to the study, rather than random illness, the APHRC team will immediately notify APHRC PI, who will notify the APHRC IRB committee and Study Director Kelly K. Baker. The subject's enrollment in the study will be terminated with their referral to an appropriate clinical provider for long-term monitoring. PI Baker will immediately notify the University of Iowa IRB committee and seek further advice on appropriate response. Stool specimens will be collected using diapers provided to the mother for wearing by the child. No rectal swabs will be performed and no medical treatment will be rendered. No other biological specimens will be collected.
Investigators will minimize the risk of Covid-19 transmission by regular, rigorous screening of staff for symptoms prior to their engagement with the community and testing where warranted, limiting the number of staff closely interacting with participants, requiring masks and/or face shields and other hygiene PPE during data collection, social distancing of 2 meters during data collection, sterilizing all materials transmitted into and out of the subject's household, and provision of masks to subjects for surveys if they do not have one of their own.
CONFIDENTIALITY AND DATA MANAGEMENT
Confidentiality. Data will be handled in a confidential manner to prevent loss of privacy. All personnel hired for this project will be required to take a five day intensive training class that includes training in ethical practices for conducting research on human subjects. Any violation of a subject's confidentiality will result in immediate termination and potential further consequences. Paper forms collected by field staff will be immediately submitted to supervisors at the end of each data collection day, and stored in a locked cabinet in a secure office at the rural office until transport to APHRC. Hardcopy forms will then be stored for up to five years in a locked cabinet in a locked office at APHRC, after which they will be destroyed. Electronic files will be stored in an encrypted and password protected database on a secure server at APHRC and the University of Iowa. Only the primary investigators and approved members of the research team will have access to information linked to subject identifiers. Information will be stored in an encrypted (meeting mandated IT security standards) password-protected database and will contain subject identifiers such as name, address, date of birth, and medical record numbers. This information is necessary to maintain contact with the subjects during the full year of the study. Only the primary investigator and members of the research team that will contact the subjects will have access to information linked to subject identifiers.
Data Management. Paper/hard copy records (hard copy surveys, questionnaires, case report forms, pictures, etc.) - A database manager at the APHRC will periodically print out lists of eligible infants and contact information for recruitment purposes and for follow up throughout the 14 day period of time subjects are enrolled. This will be retained by APHRC supervisors until the study is completed, and then destroyed by shredding. Consent forms will be collected on hardcopy forms and will be stored in locked file cabinets in locked offices at the APHRC. Similarly, the 14 day diarrhea calendars will be identified only by the random household ID, but will nonetheless be stored after collection in a locked office at APHRC.
Electronic records (computer files, electronic databases, etc.) - Data will be collected using mobile technology using ODK software on password protected tablets. These tablets will be issued by supervisors to enumerators every morning and retrieved each evening. Data from tablets will be uploaded by APHRC supervisor via the internet or cellular data to a HIPPA-compliant APHRC shared server, and then deleted from the tablet to prevent accrual of private information on subjects on a device used in the field. Electronic records with identifiable information stored on the server will only be accessible to the UI and APHRC leadership team. PHI must be removed from any data that is used by other individuals and that is to be stored on personal laptops or computers. Data quality checks will be built into the data capture software to ensure that no missing information or implausible values are accepted. Once all the completed interviews on each tablet are validated by the field team leader and signed off for synchronization, every data collector will transmit the data in his/her tablet to a central server via internet connection. This will ensure that the data is not tampered with thereafter and to minimize possibility of data losses in the course of data collection that may arise from human error or equipment failure or loss. Synchronized data once in the server will be checked further for any inconsistency by running data quality scripts daily that generate any error reports for immediate sharing with the field teams for further review and correction as may be necessary while the field teams are still working in the study areas where data with the errors have been noted. Upon completion of field data collection, data will be verified, validated and thoroughly cleaned. All data will be stored (through Internet linkage) to a secure physically located server with standard operating procedures for access, security, and data backup.
Security features include remote access that will be through passwords and only from registered computers. All stored data will be encrypted along with encrypted data transfer using SSL transport layer encryption to secure the transfer of packets of data uploaded incrementally to a dedicated mobile dispatch server (MDS).
Biologic samples - When stool samples are collected in households, they will be labeled only with a randomly selected barcode ID, which will be linked to individual households by scanning the barcode into a tablet based file linked to the subject's identifying information. From this point forward, stool samples will retain the random ID through processing at the APHRC lab to molecular analysis at the UI lab. Lab technicians will be blinded to subject identity. The final results of stool analysis for each random ID will be sent by the UI lab to the database manager, who at the final stage will merge them by random ID into the overall dataset that contains identifiable information. Thus, biological samples will be identified at all stages except point of collection and point of analysis.
FUNDING The PATHOME study is funded by the United States National Institutes of Health Fogarty Institute R01TW011795 to University of Iowa, PIs Kelly K. Baker and Daniel Sewell. It has been approved by human subjects review boards at the University of Iowa, African Population and Health Research Center, AMREF and NACOSTI.
Inclusion criteria:
Exclusion criteria: