Background information

Why is patient information important for research?

Patient records are an extremely valuable resource for research. Data can be used in two different ways:

1. For epidemiological and observational research

to understand more about the causes of disease
to detect outbreaks of infectious diseases
to monitor the safety and efficacy of drugs
to study the effectiveness of treatments and interventions.

2. To identify potential participants to take part in health research, including clinical trials

Patient records can be used to identify patients with a particular disease who fit the eligibility criteria for a trial or cohort study, or to identify a representative sample of the general population as a control group. Appropriate participants are then contacted to invite them to take part in the study.

The increasing use of electronic records opens up possibilities to ask new research questions by combining data from very large numbers of patients and by linking different datasets.

Examples of research

In the 1950s, health data played a major part in uncovering the link between smoking and lung cancer. More recently, the health data of children with autism born since 1979 in eight UK health districts helped scientists establish that there is no link between the measles, mumps and rubella (MMR) vaccination and autism.

Definitions: identifiability of information

Anonymised information: Data must not include any identifiers such as names, addresses, full postcodes, full dates of birth or NHS numbers. Data are provided at an individual level, but it is not possible to identify an individual.

‘Key-coded’ information: Information cannot directly identify an individual (e.g. name and address are removed), but a ‘key’ or code is available to link data back to the individual. The likelihood of identification increases where data relate to individuals with rare illnesses or to small populations. This information is also referred to as ‘pseudonymised’ or ‘linked anonymised’.

Identifiable information: This type of information contains personal data that directly identify individuals (e.g. names, addresses and/or NHS numbers).

It is important to note that while researchers may need to have data at a ‘person-level’, they do not usually want to know the actual identity of the individual.

Current challenges when using patient data for research

The regulatory and governance framework for the use of health information in research is complex and confusing, and it needs urgent clarification. Researchers have faced a lack of consistency regarding the processes that should be used when information from patient records is required for research. Faced with uncertainty, researchers, GPs, NHS Trusts and research ethics committees often err on the side of caution, delaying or preventing much-needed research. Particular challenges include:

The use of key-coded data
Wherever possible, researchers use anonymised, non-identifiable information. However, sometimes researchers will need to access data from which it may be possible, directly or indirectly, to identify a patient. As it becomes increasingly possible to link different datasets, there is also a greater likelihood of identifying individuals. It is unclear whether key-coded data falls within the scope of legislation.

‘Consent for consent’
The issue of inviting patients to take part in research is particularly problematic. Researchers may need to review medical records to determine whether patients meet the eligibility criteria for the study, such as diagnosis, age or gender. However, because this may involve viewing identifiable information, researchers have often been prevented from accessing the data. Once potential participants have been identified, GPs are sometimes required to contact patients to ask whether they are happy to be contacted at a later time with information about a study. Only after this initial contact can researchers contact patients to invite them to participate in the study.