Helping people to live longer and in good health is a priority for governments. This goal must be underpinned by evidence about older people’s health and needs, particularly for populations who are most likely to be poorly. Material disadvantage plays a key role in people’s health. Compared to the richest populations, the poorest are more likely to experience poor health, and poor health starts earlier in their life.
The Clinical Practice Research Datalink (CPRD) may offer an opportunity to explore these issues. CPRD routinely collates patient data from GP practices across the UK. Some of the information collected may describe the material circumstances of people when these circumstances impact their health. We want to know whether it is possible to use this information to measure the number of people experiencing material disadvantage.
CPRD also includes information about area deprivation. However, we want to know whether it is possible to describe and measure the material circumstances of people, rather than the material conditions of the areas in which they live.
The aim of the work is to explore whether there is information within CPRD to count the number of people experiencing material disadvantage in primary care. We want to use this information to explore the health of the most disadvantaged older people.
Promoting healthy ageing is a policy priority. This goal must be underpinned by evidence, particularly for populations most at risk of poor health. Material disadvantage plays a key role in people’s health. Compared to the most advantaged populations, those most disadvantaged are more likely to experience age-related poor health from earlier in the life course, for example. Targeting support at the poorest populations could help to delay the earlier onset of poor health, and close the socioeconomic gap in healthy life expectancy. The aims of our exploratory study are to explore the feasibility of using Clinical Practice Research Datalink (CPRD) data to quantify the number of people experiencing material disadvantage; and to make recommendations about the use of CPRD data to explore links between material disadvantage and health.
Within CPRD, there may be an opportunity to quantify the number of people experiencing material disadvantage through aggregating SNOMED codes relating to benefits, housing and employment, for example, and validating ‘material disadvantage’ against Indices of Multiple Deprivation (IMD) codes at the household level. Our study population will be adults (age 18 and above) in CPRD Aurum who are eligible for linkage to patient-level area deprivation data. Our outcome of interest is material disadvantage, which we will code as (a) a binary variable from having any SNOMED code from patient observations listed and (b) grouping SNOMED codes into 9 different domains of material disadvantage (Appendix A). Our covariates will be a history of along-term condition (Appendix B), age group and sex. We will use descriptive tables and logistic regression, adjusted for age group and sex, to explore associations between material disadvantage and these long-term conditions.
Determining the utility of CPRD data to identify people experiencing material disadvantage could support future analyses on the health and care needs.
Work completed from approved protocol
• We will look for a record of one or more SNOMED codes associated with material disadvantage, as listed in Appendix A, that have been documented in the patient observation record of Aurum. We will aggregate these codes into a binary material disadvantage variable. We will further explore the feasibility of creating a categorical or score based measure using the nine different domains of material disadvantage (benefits, employment, environment, housing, income, neglect, support service, transport, other).
• Secondary outcome: one or more of the long-term conditions specified in Appendix B. Appendix B also includes codes to be used to filter out patients who have a condition that was not long-term. We defined chronic conditions as having a history of the disease, except for non-melanoma skin cancers that weren’t infiltrating, and conditions that are typically transient (e.g. diabetes/hypertension/anaemia of pregnancy), or have resolved (for example, children can outgrow asthma and epilepsy).
Amended approach
We will look for a record of one or more of the codes associated with material disadvantage,
as listed in Appendix A, that have been documented in the patient observation record of Aurum.
We will aggregate these codes into a binary material disadvantage variable.
Barbara Hanratty - Chief Investigator - Newcastle University
David Sinclair - Corresponding Applicant - Newcastle University
Andrew Kingston - Collaborator - Newcastle University
Gemma Spiers - Collaborator - Newcastle University
Laurie Davies - Collaborator - Newcastle University
Patient Level Index of Multiple Deprivation