Long term wounds are a serious health problem worldwide, affecting a significant number of people. This not only harms people's health but also cost a lot of money. This affects about 7% of the UK population who have a wound that last more than 4 weeks.
Mental health issues, such as stress and depression, can affect how wounds heal. In fact, hard-to-heal wounds can have a big impact on a person's mental health, due to wound related symptoms such as pain and mobility, and time for the wound to heal. This creates a cycle where the wound affects mental health, and mental health affects wound healing.
We are interested whether data in the CRPD database captures information on (a) whether people with long term wounds have poor mental and psychological wellbeing outcomes and (b) the factors that are likely to influence whether people with long term wounds have poor mental and psychological wellbeing outcomes.
We would like to examine
(1) the proportion of people with hard to heal wounds who have poor mental health and psychological wellbeing captured by the CRPD database.
(2) Whether data in CRPD database could be used to determine people more at risk of poor mental health or psychological wellbeing outcomes.
This information in this feasibility study will determine whether it is practical to use CRPD data to develop computer assisted learning (machine learning) from data to predict poor mental and psychological wellbeing outcomes in people with long term wounds.
Rationale: Risk prediction models could be used to estimate the likelihood of outcome development. This can improve risk assessment and personalise care. This project will determine the feasibility of the CRPD dataset in predicting poor mental health and psychological wellbeing outcomes in people with a chronic or hard to heal wound.
Background: In the UK 7% of the population have a wound that last more than four weeks (Guest et al., 2020). These may have a negative impact on emotional wellbeing and quality of life (Olsson, 2019). Poor psychological wellbeing, resulting from the wound and its associated symptoms, may lead to a decreased motivation to manage their wound and adhere to treatment impeding healing (Atkin et al, 2019). Machine learning could be used to predict risk of developing poor psychological wellbeing based on data collected from previous patients (Queen, 2019). This may enable early interventions to be targeted and care personalised.
Research questions:
•Is poor mental health or psychological wellbeing in people with chronic and hard to heal wounds captured on the CRPD database?
•Are variables affecting poor mental health or psychological wellbeing outcomes in people with hard to heal and chronic wounds captured on the CRPD database?
Method:
•Patients over 18 years old who a chronic wound or surgical/trauma wound unhealed after four weeks will be exported from the CRPD database.
•These patients will be classified as those with and without poor mental health and phycological wellbeing.
•Variables of interest identified within our recent scoping review of factors thought to effect mental health and psychological wellbeing will be examined to determine whether these factors are captured
•We will determine whether a predictive model using machine learning algorithms could be built from this data. This will inform our future study.
Primary exposure(s): patients over 18 years old with a wound lasting longer than 4 weeks will be identified. This is based on the wound related read codes and exclusion criteria identified by Guest et al (2020). See information on study population.
Primary outcome(s): onset of poor psychological wellbeing categorised as any or a combination of the following: new diagnosis of mental health disorders, recording of non-pharmacological psychological therapies, recording of pharmacological psychological therapies and/or referral to mental health services. This outcome include codes for mental health care plan (https://www.opencodelists.org/codelist/nhsd-primary-care-domain -refsets/mhp_cod/20200812/#full-list), invitations for mental health review (https://www.opencodelists.org/codelist/nhsd-primary-care-domain-refsets…), depression and anxiety diagnosis and symptoms (https://www.opencodelists.org/codelist/ons/depression-and-anxiety-diagn…), seen by wellbeing coaches (https://www.opencodelists.org/codelist/nhsd-primary-care-domain-refsets…) as well as the use of medications to manage psychological wellbeing and patient admission data using medical codes (SNOMED) and product codes based on the BNF code lists developed before by Dr Dregan.
Other variables of interest will include wound characteristics and symptoms, and index of multiple deprivation.
Crina Grosan - Chief Investigator - King's College London (KCL)
Victoria Clemett - Corresponding Applicant - King's College London (KCL)
Alex Dregan - Collaborator - King's College London (KCL)
Edgar Mandeng Ma Linwa - Collaborator - King's College London (KCL)
Mariam Molokhia - Collaborator - King's College London (KCL)
HES Admitted Patient Care;Patient Level Index of Multiple Deprivation