These data show, for each local authority district, the hospital admission rates, split by ethnic group for major disease categories and for preventable admissions. The data are standardised by age and sex.

Accurate recording of ethnicity in electronic healthcare records is important for the monitoring of health inequalities. Reducing inequalities in health has explicitly been part of the government agenda in the United Kingdom since 1997. Inequalities are associated with poverty and may be exacerbated for ethnic minorities due to discrimination, lack of health knowledge or other barriers in access to health services such as language. NHS-commissioned hospitals monitor their use by ethnic group in the national hospital admission records database, known as Hospital Episode Statistics (HES).

Emergency hospital admissions are distressing for patients, associated with poorer long-term outcomes, and are costly to the healthcare system. Indicators for admissions considered preventable have been defined and known as ambulatory care sensitive conditions (ACSC) - acute, chronic, and vaccine-preventable combined. ACSC admissions have been associated with patients under the age 5 years, the elderly, deprivation, and ethnicity. NHS monitors ACSC in the general population and saw a 40% rise in 2001-2011. Studies in US, New Zealand, and Scotland have found higher risk of ACSC admission for many ethnic minorities compared to the White majority populations.

Historically, ethnicity information was absent in Hospital Episode Statistics (HES) from a significant portion of records of patients who received inpatient care in England. To address the gap in the completeness of ethnicity data, CDRC worked with NHS Digital to enable them to enhance HES data collected between 1999/00 and 2013/14 with name-based ethnicity imputation using the CDRC Ethnicity Estimator software (EE). The resulting aggregated statistics were supplied back to CDRC.

Content

The dataset, available as 3 georeferenced CSV files, can be split into two broad categories:

• The preventable hospitalisation admission by ethnic group data are age- and sex-standardised hospital admission rates per 100,000 population by ethnic group and English Local Authority District. Data are for acute ambulatory care sensitive conditions (ACSC). Data for acute, chronic and vaccine-preventable cases are combined. ACSC admissions are regarded as preventable admissions and have higher associations with under-5, elderly, deprived, and minority ethnic group patients. These data are provided for a single time period, 2009 to 2014. There is one row for each ethnic group.

• The hospital admission for major disease categories data are age- and sex-standardised hospital admission rates per 100,000 population by ethnic group and English Local Authority District. Data are for 27 Level 1 (all major) disease categories as defined by the Global Burden of Disease (GBD). This data has been provided for two time periods, 1999-2004 and 2009- 2014, with one file for each period. Each file has one row for each combination of local authority district, ethnic group, disease category and ethnicity coding type.

Quality, Representation and Bias

HES has near-complete coverage of NHS commissioned hospital admissions in England. Coding of diagnoses may vary in consistency but has been validated for research and auditing purposes in an earlier study.

As a limitation, it should be noted that ethnicity is a complex concept encompassing biological, cultural, and subjective aspects. Variation in prediction success of name-based ethnicity classification can therefore arise for different reasons including individuals’ sense of belonging and resulting choice of ethnic group, socio-cultural naming and name-change practices, distinctiveness of names across ethnic groups, and the extent to which the name-based classification covers different origins at a given time point, e.g. when later waves of immigration have widened the range of diasporic names in the host country since the creation of the software.

We used denominator data from Census 2011 as the most complete dataset on the ethnicity of the residential population in England. The census contains self-reported ethnicity. HES draws on the central NHS patient register with self-reported ethnicity. The proportion of patients without ethnicity record was 38.8% in 1999/00-2003/04 and 9.7% in 2009/10-2013/14. The dataset contain admission estimates by ethnic group as recorded by NHS as well as three different ethnicity classifications enhanced with names-based ethnicity.

Full details about the completeness of ethnicity records and the prediction success of the EE software for different ethnic groups over time and across regions can be found in the PDF report attached to this data record.