Health Literacy: Prevalence Estimates for Local Authorities
This tool provides an estimate of the percentage of a local authority population with low health literacy and numeracy or with just low health literacy. It draws on the most recent national survey of literacy and numeracy in England (the 2011 Skills for Life Survey) with a population update based on the 2021 Census and the 2019 Index of Deprivation. Further details are included on the Methods tab (accessed via the Map view).
To search by local authority, select either "health literacy alone" or "health literacy and numeracy combined". Then enter the name of the local authority using lower tier local authority names i.e. the local council (district or borough).
The search by local authority will give you the prevalence for the chosen indicator in the chosen council and a graph comparing the chosen council with the national average. The local authority search will also give access to tabs offering: a) a zoomable map showing the chosen values for nearby council areas; and b) a summary of the methods used in the data analysis.
Alternatively, you can search using the map view. When using the map view, you can zoom in or out to see comparative levels over a wider or smaller geographic area. Click on the local authority areas in the map to see the estimated prevalence. The map view also gives access to the Methods and Graph tabs.
Estimated prevalence of low health literacy
Percentage of the population aged 16-65 that are BELOW threshold levels of health literacy |
This represents an estimate of the prevalence of LOW health literacy for the chosen Local Authority. It indicates the percentage of the population aged 16-65 who would likely have difficulties in understanding or interpreting health information.
IMPORTANT: the prevalence measure is an estimate derived from a statistical model. It should not be taken as a precise measure. Like all modelled estimates, it reflects measurement and modelling issues. Each local authority will also contain areas that will have higher or lower prevalences. The estimates apply only to the 16-65 age group.
Summary information on the modelling approach is given on the Methods tab above.
The mean national prevalence of the population aged 16-65 that are below the threshold for both health literacy and health numeracy is 58.3%; for health literacy alone, the mean national prevalence is 38.66%.
Selected ()
Mean national prevalence
Estimated prevalence of low health literacy
Map shows the level of prevalence in the chosen Local Authority. Click on a neighbouring Authority to display its value.
Zoom in and out of the map using the '+' and '-' controls in the top left-hand corner. Zoom to the extent of England using the 'zoom to max extent' button.
Methods
The initial underpinning data for the estimates were taken from the Skills for Life Survey (SfLS), 2011. This remains the most recent survey for which it is possible to estimate health literacy/numeracy levels within a representative English population. Prior research by Rowlands and colleagues estimated individual levels of health literacy/numeracy using weighted logistic regression. They identified characteristics that determined whether or not an individual in the SfLS met a threshold level which would enable them to understand 70% of the health-related materials that they would be likely to encounter in routine health interactions [1],[2]
The local authority level estimates used multilevel small area estimation [3],[4],[5]. This approach has been widely used to explore geographical variations for a range of health and other measures for which geographical data do not otherwise exist, including health-related behaviours, unmet health needs and social capital, as well as health literacy [6]. Its development and application has been supported by health agencies, UK research councils and the market research industry.
For health literacy/numeracy, the multilevel small area estimation process involved:
- Developing multilevel models to predict individual health literacy/numeracy using SflS data. The covariates for health literacy were (individual level) sex, age (ten-year bands) and education (degree, none, other), and (area level) index of deprivation (quartiles), proportion of primary sampling unit (PSU) not speaking English as a first language at home, and PSU proportion of South Asian ethnicity. The same covariates were used for the health literacy plus health numeracy model save that the non-English speaking covariate was replaced by the PSU proportion of Black ethnicity. An extensive selection process underpinned the choice of covariates reflecting literature on the determinants of health literacy/numeracy and statistical issues. Modelling used Bayesian Monte Carlo Markov Chain (MCMC) methods with MLwiN software [7]. Models were extensively checked for quality.
- Reworking the two multilevel equations, applying data from the 2021 Census and the 2019 Index of Deprivation to the relevant covariates to generate estimates of the numbers of individuals with particular age-sex-education characteristics living in each Lower Level Super Output Areas (LSOA) in England, adjusting these counts to reflect LSOA levels of deprivation, and (reflecting the underpinning model) proportion not speaking English, Black or South Asian ethnicity, and further adjusting to take account of regional variation. This process ensures that the LSOA estimates reflect the most recent available data as well as local and regional variation.
- Aggregating the LSOA estimates to lower tier local authorities using look-up tables available from the Office for National Statistics and calculating credible intervals for the estimates using the MCMC modelling outputs to derive measures of the statistical uncertainty that accompanies the estimates. Credible intervals are the Bayesian term for the more familiar confidence intervals. On average the 95% credible interval for each local authority is c. +/-4%.
References
- Rowlands, G., Protheroe, J., Winkley, J., Richardson, M., Seed, P. T., & Rudd, R. (2015). A mismatch between population health literacy and the complexity of health information: an observational study. British Journal of General Practice, 65(35), e379-e386.
- Wolf, M. S., & Rowlands, G. P. (2015). Developing a method to derive indicative health literacy from routine socio-demographic data. Journal of Healthcare Communications, 1(1), 7.
- Twigg, L., Moon, G., & Jones, K. (2000). Predicting small-area health-related behaviour: a comparison of smoking and drinking indicators. Social Science & Medicine, 50(7-8), 1109-1120.
- Twigg, L., & Moon, G. (2002). Predicting small area health-related behaviour: a comparison of multilevel synthetic estimation and local survey data. Social Science & Medicine, 54(6), 931-937.
- Twigg, L., Moon, G., & Walker, S. (2004). The Smoking Epidemic in England. London: Health Education Authority.
- Rowlands, G., Whitney, D., & Moon, G. (2018). Developing and applying geographical synthetic estimates of health literacy in GP clinical systems. International Journal of Environmental Research and Public Health, 15(8), 1709.
- Browne, W. J., & Rasbash, J. (2009). MCMC estimation in MLwiN. Bristol: Centre of Multilevel Modelling, University of Bristol.