Analyze

At the height of the West African Ebola epidemic, West African governments and Mobile Network Operators (MNOs) were barraged with requests from international humanitarian and Western data analytics agencies to provide Call Detail Record data. This data could furnish the large-scale ambitions of data modelling to track and predict contagion. Despite its utility in tracking mobility and, as such, disease, CDR’s use raises many privacy concerns. In addition, embedded within a turn towards datafication, CDR technologies for surveillance embed specific ontologies of the data-focused society they emerge from. There is a false equivalence embedded in the relationship between humans and technology. The predominantly Western idea that one phone equals one person underlines the claim that CDR data accurately tracks distinct user movements, encoding a Western “phone self-subjectivity” (Erikson 2018). However, the refusal by some African actors to hand over sensitive mobile data to international agencies was met with forceful rhetoric of Africa’s moral obligation to comply—to forgo privacy rights in the name of ‘safety.’ The Ebola context reflects an emergent digitization of emergencies in the Global South, which is reshaping the way societies understand and manage emergencies, risk, data, and technology. The big data frenzy has seen a rising demand to test novel methods of epidemic/pandemic surveillance, prediction, and containment in some of the most vulnerable communities. These communities lack the regulatory and infrastructural capacity to mitigate harmful ramifications. With this emergence is a pivot towards 'humanitarian innovation,' where technological advancements and corporate industry collaboration are foregrounded as means to enhance aid delivery. In many ways, these narratives of innovation and scale replicate the language of Silicon Valley’s start-up culture. Surveillance of the poor and disempowered is carried out under the guise and rhetoric of care. In this scenario, market ideals and data technologies (re)construe social good as dependent on the “imposition of certain unfreedoms” as the cost of protection (Magalhaes and Couldry 2021). As big data technologies, they foreground a convergence of market logistics and global networks with existing and already problematic international humanitarian infrastructures (Madianou 2019). These convergences create new power arrangements that further perpetuate an unequal and complex dependency of developing countries on foreign organizations and corporations. Pushback against these data demands showcases competing notions of where risk truly lies. While resistance to data demands was at the state level, community responses to imposed epidemic regulations ranged from non-compliance to riots. These resistances demonstrated how the questions of ‘who and what is a threat?’ or ‘who and what is risky?’ and ‘to whom?’ experience shifting definitions in relation to these technologies as global, national, and community imaginaries are reinforced and reproduced as cultural, political, as well as biological units. 

Source

Akinwumi, Adjua. 2023. "Technological care vs Fugitive care: Exploring Power, Risk, and Resistance in AI and Big Data During the Ebola Epidemic." In 4S Paraconference X EiJ: Building a Global Record, curated by Misria Shaik Ali, Kim Fortun, Phillip Baum and Prerna Srigyan. Annual Meeting of the Society of Social Studies of Science.

(MNOs) were barraged with requests from international humanitarian and Western data analytics agencies to provide Call Detail Record data. This data could furnish the large-scale ambitions of data modelling to track and predict contagion. Despite its utility in tracking mobility and, as such, disease, CDR’s use raises many privacy concerns. In addition, embedded within a turn towards datafication, CDR technologies for surveillance embed specific ontologies of the data-focused society they emerge from. There is a false equivalence embedded in the relationship between humans and technology. The predominantly Western idea that one phone equals one person underlines the claim that CDR data accurately tracks distinct user movements, encoding a Western “phone self-subjectivity” (Erikson 2018). However, the refusal by some African actors to hand over sensitive mobile data to international agencies was met with forceful rhetoric of Africa’s moral obligation to comply—to forgo privacy rights in the name of ‘safety.’ The Ebola context reflects an emergent digitization of emergencies in the Global South, which is reshaping the way societies understand and manage emergencies, risk, data, and technology. The big data frenzy has seen a rising demand to test novel methods of epidemic/pandemic surveillance, prediction, and containment in some of the most vulnerable communities. These communities lack the regulatory and infrastructural capacity to mitigate harmful ramifications. With this emergence is a pivot towards 'humanitarian innovation,' where technological advancements and corporate industry collaboration are foregrounded as means to enhance aid delivery. In many ways, these narratives of innovation and scale replicate the language of Silicon Valley’s start-up culture. Surveillance of the poor and disempowered is carried out under the guise and rhetoric of care. In this scenario, market ideals and data technologies (re)construe social good as dependent on the “imposition of certain unfreedoms” as the cost of protection (Magalhaes and Couldry 2021). As big data technologies, they foreground a convergence of market logistics and global networks with existing and already problematic international humanitarian infrastructures (Madianou 2019). These convergences create new power arrangements that further perpetuate an unequal and complex dependency of developing countries on foreign organizations and corporations. Pushback against these data demands showcases competing notions of where risk truly lies. While resistance to data demands was at the state level, community responses to imposed epidemic regulations ranged from non-compliance to riots. These resistances demonstrated how the questions of ‘who and what is a threat?’ or ‘who and what is risky?’ and ‘to whom?’ experience shifting definitions in relation to these technologies as global, national, and community imaginaries are reinforced and reproduced as cultural, political, as well as biological units. 

Akinwumi, Adjua. 2023. "Technological care vs Fugitive care: Exploring Power, Risk, and Resistance in AI and Big Data During the Ebola Epidemic." In 4S Paraconference X EiJ: Building a Global Record, curated by Misria Shaik Ali, Kim Fortun, Phillip Baum and Prerna Srigyan. Annual Meeting of the Society of Social Studies of Science. Honolulu, Hawai'i, Nov 8-11.

Creators of the Student Health Index recommend using the tool in combination with qualitative data collection and stakeholder/community engagement (e.g. working with school leaders, local community leaders, and healthcare providers).

A full guide to using the dashboard is available here.

Data sources utilized by the index are not always the most current due to data collection limitations (e.g. covid-19 has caused disruptions in the collection of CDE data).

The Index is limited in that it does not offer data for schools that were not large enough to warrant the construction of a School-based Health Center. Thus, schools that did not meet specific enrollment targets were excluded from the dashboard. This includes rural schools (designed as such by the USDA) with an enrollment under 500 students, urban schools (without a high school) with less than 500 students, and urban schools (with a high school) with less than 1000 students. California had more than 10,000 active public schools in 2020-21. The final dashboard for the Student Health Index includes 4,821 schools.

The lack of available data on health indicators at a school-level restricted the Student Health Index to using proxies for the health outcomes. Some health indicators are included, but they are not school-specific, instead linked to specific schools geographically through the census tract. However, community-level data does not always accurately reflect the characteristics of a school’s population. As a result, school-level indicators in the Index were weighted more heavily than community-level indicators.

Additionally, race was not included as a measure in the Student Health Index because of California’s Proposition 20, which prohibits the allocation of public resources based on race and ethnicity. However, the dataset does contain measures of non-white students at each school. 

The Index has also been limited as a quantitative measure of need, which may overlook the influence of other factors that might be better illuminated through qualitative evidence (e.g. stakeholder engagement, focus groups, interviews, etc.).

No information thus far

The Student Health Index can produce visualizations that represent data on conditions, school characteristics and risk factors that affect education outcomes and could be improved through access to school-based health care. These visualizations can be used to demonstrate need for expanding school-based health care access in California.

In addition to maps, the index can also be used to generate graphs and visual displays of data (e.g. ratio of highest need schools to all schools, by county).

The visualizations can be used to demonstrate the correlations between final need scores and race, the impact of specific indicators in health, and the concentration of need to certain regions of California (hot spot analysis).

The Student Health Index enables users to identify where SBHCs will have the most impact for students. The index uses 12 indicators, each of which can be scored from 1 to 4 for any given school. These scores are generated using percentiles and represent relative values. The 12 indicator scores are combined into a Need Score, which is calculated using percentiles along a scale of 1 to 4. Schools with a score of 4 (in the 4th quartile) have the highest Need scores relative to other schools in California.

The index is composed of 12 diverse indicators (percentages, rates, and index values) that have been transformed using percentiles in order to enable comparisons on a common scale. These indicators are divided into 3 categories: health indicators, school-level indicators, and socioeconomic indicators.

Health Indicators

1. Diabetes
2. Asthma ED admissions
3. Teen birth
4. Health Professional Shortage Areas (HPSA)

Socioeconomic Indicators

1. Poverty among individuals under 18
2. Uninsured among under 19
3. Healthy Places Index

School-Level Indicators

1. Percent FRPL (students eligible for free or reduced-price meals)
2. Percent English Learners
3. Percent Chronically Absent
4. Percent experiencing homelessness
5. Suspension rate

Other Data

1. Mental health hospitalization rate
2. Percent in foster care

Indicator selection was guided by CDC estimations on the primary contributing factors that shape health (social determinants of health, medical care, and health behaviors). The indicators included in the index are all either directly associated with the absence of health services that could be provided at a school level, act as proxies for health behaviors, or represent social determinants of health that could be addressed through access to school-based health services.

Indicator selection was influenced by recommendations from the Research Initiative of the Campaign for Educational Equity at Columbia Teachers College, which found that seven health disparities affecting school-aged youth could be addressed through school health programs. These disparities include: (1) vision, (2) asthma, (3) teen pregnancy, (4) aggression and violence (including bullying), (5) physical activity, (6) hunger, and (7) inattention and hyperactivity.

More detailed description of the rationale shaping indicator selection is available here.

Census tract, school-level

The Student Health Index draws from data that is publicly available and up to date on a statewide level. Sources include the University of California San Francisco Health Atlas, the American Community Survey, the U.S. Census Bureau, the California Department of Education’s Downloadable Data Files site, and the CDC.

Detailed list of sources:

PLACES Project, CDC (available through the UCSF Health Atlas)

CalEnviroScreen (available through the UCSF Health Atlas)

Opportunity Atlas (available through the UCSF Health Atlas)

Health Resources and Services Administration (available through the UCSF Health Atlas)

American Community Survey (available through the UCSF Health Atlas)

California Department of Education’s Downloadable Data Files site

Kidsdata.org