The California Student Health Index was constructed by incorporating information gleaned from the analysis of the landscape of existing indices and publicly available datasets.

The Student Health Index:

• Provides information for all included public schools in California using school-level and geographic data and is place-based.
• Is made up of 14 indicators that are available for at least 99% of the included schools and characterize both population characteristics and health care access.
• Uses percentiles to assign scores of 1-4 for each of the indicators in a given school. The percentile represents a relative score for the indicators, with 4 indicating the highest scores for an indicator.
• Combines the component scores into a Need Score, while double weighting all school-level data, and uses percentiles to assign a score of 1-4 to each score. This creates a relative score of 1 to 4 for each school. Schools in the 4th quartile (score of 4) have the highest Need Scores, relative to all schools.

Calculating Need Scores for schools entails three steps:

1. Merging and Spatially Linking Datasets
2. Rescaling Indicators
3. Calculating Relative Need Scores

The indicators used were: 

Merging & Spatially Linking Datasets

Data in the model comes from a variety of sources. To simplify reproducibility, all census-tract level data is sourced from the UCSF Health Atlas, an interactive population health mapping website that curates publicly available data and displays it at the census tract level in California. All school-level data comes from the California Department of Education’s (CDE) Downloadable Data Files site. Data on COVID vaccination rates is collected at the zip code level by the California Department of Health and Human Services.

The CDE data is available in distinct data files that must be merged using the 14-digit school code that uniquely identifies each school within California. For example, data on suspension rates, race and ethnicity, chronic absenteeism, school location, and free and reduced-price lunch eligibility are available in separate data files. In total, this analysis required the merging of five different CDE datasets.

The CDE Public Schools and Districts Data Files contain a latitude and longitude for each school location, which were spatially joined with the data from their underlying census tracts using Stata’s spatial data analysis package. Thus, the final, cleaned dataset contains information on each school’s characteristics, as well as underlying census tract characteristics. Vaccination rates are linked on the school’s zip code.

Rescaling Indicators

The diverse indicators that comprise the Student Health Index include percentages, rates, and index values. To include them in a composite measure, each is transformed to enable comparisons on a common scale, using percentiles. Thus, performance on an indicator is compared with the highest and lowest scores obtained on that indicator.

Specifically, the index draws on 14 indicators related to socioeconomic factors and health outcomes at the school and neighborhood (census tract) level. Each school is given a score of 1-4 for each indicator, based on the quartile that indicator falls into relative to all other schools. For example, a school with a very high percentage of students eligible for free or reduced-price meals would receive a score of 4, while a school with few such students would receive a score of 1.

Calculating & Weighting the Opportunity Scores & Grades

The calculation methodology was based off the methodologies used by the model indices (see Full Report for more information). Because school-level data is more specific to the individual school context than community-level data, it is double weighted for the purpose of this analysis. All school-level scores are doubled, creating scores of 2-8 for school data. All scores are then added up across all 14 indicators, creating a Need Score ranging from 19 to 73. This score is again broken into four categories using quartiles, which allowed for the creation of the Student Health Index. Each school is thus designated as either highest need, higher need, lower need, and lowest need.

Inclusion/Exclusion Criteria

The schools included in the Index were selected intentionally based on specified inclusion criteria. California had over 10,000 active public schools at the start of the 2021-22 school year, according to data from the California Department of Education. The goal of this analysis was to create a list of locations to target for additional health care services, and thus for the purpose of this analysis, schools that met certain criteria were excluded from the list of schools that were statistically analyzed. The exclusion criteria are as follows:

• Virtual Instruction Schools: Schools that are primarily virtual or all virtual were excluded.
• Small Schools: In consultation with the CSHA it was determined that schools with enrollment under 100 be excluded from the statistical analysis.
• Certain Special Educational Options: Schools that serve only adults were excluded from this analysis, as were occupational centers, Youth Authority Facilities and County Community Schools.
• Preschools and Kindergartens: The analysis is focused on schools serving grades 1-12.

Using these criteria, a little over 8,000 public schools were selected for relative needs analysis. Schools with existing SBHCs were included in the index, in part so that the CSHA can determine whether these centers are currently located at relatively high need schools. These schools can be filtered out of the online resource and associated dataset according to data users’ interests and needs.

Table 2: Number and Type of Schools in Final List

 
Contact us if you would like to receive the complete data set associated with this analysis, including a list of schools included in the index.

Differentiating Rural & Urban Schools

Because rural schools generally are smaller and serve a population with different access to urban infrastructure such as health clinics and public transit, urban and rural schools were analyzed separately. The same indicators were used, but each indicator was given a relative score based on all other schools in the same rural or urban classification. Thus, a high need urban school is ranked only in comparison to other urban schools, and a high need rural school is high need relative to other rural schools.