ESRC/NIH Health Disparities - Understanding Social Contributions to Disparities in Depression Care: US and UK

Lead Research Organisation: University of Warwick
Department Name: Warwick Medical School

Abstract

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
 
Description Key findings and achievements from the research include the Study 2 dataset of UK analogue patients' (n=320) narratives about their experiences of depression or feeling low, collected when interacting with a video doctor. The sample of participants is stratified by race (African-Caribbean versus white British), gender and a previous formal diagnosis of depression. The data were difficult to obtain, but they are extremely rich and are permitting detailed analysis of the ways in which different groups approach talking about depression: what they say, how they say it, how much they say, and how they think about depression compared with the sample of Study 1 doctors. These are all important insights for guiding future diagnosis of depression, for understanding patients' perspectives and concerns, and for averting disparities arising out of doctor-patient interactions during primary care consultations.

Linked to this, a key Study 1 finding was that doctors in both the US and the UK are much more uncertain about diagnosing depression in black compared with white patients, particularly in the UK. This was evident in doctors coming up with more potential diagnoses for black compared with white patients, when presented with identical (save for the fact of the patients' race) video patient presentations of symptoms strongly suggestive of depression. There was also a tendency for doctors in both countries to focus more on black patients' physical rather than psychological symptoms and to identify endocrine problems, most often diabetes, as the key presenting complaint for them. This suggests that doctors in both countries have a less well developed mental model of depression for black compared with white patients. Our research findings are contributing to fleshing out the required missing mental model.

Our findings from Study 2 also indicate that doctors' communication style is more important than doctor-patient race concordance for minimising health disparities in the acceptance of treatment for depression amongst black patients. This is contrary to previous US findings about the importance of race concordance, but is good news, since doctors' communication style is amenable to change through education and awareness-raising. Opportunities for doctor-patient race concordance on the other hand are severely constrained for UK African-Caribbeans, since only 2% of the British medical workforce is from this race background. Hitherto, little attention has been paid to the role of doctors' communication style in helping to reduce health disparities. We recommend that more research is undertaken in this area, to support the evidence base for communication skills training; whereas further examination of the impact of doctor-patient race concordance does not appear fruitful.

A significant but unexpected finding from Study 2 is the willingness of participants to disclose information about depression to a computer, where the doctor is only present on a pre-recorded video. Many participants told us that they shared information never previously discussed with anyone, and data analysis revealed participants' hunger for reassurance and help with mental health problems, which their narratives and feedback suggest they have not been able to secure elsewhere. Our findings will therefore help to inform the development of patient activation and community support programmes, to help them prepare for clinical consultations about depression and other mental health/emotional problems.

In Study 3 we examined clinical risk judgment data generated in response to training exercises, whilst training 193 mental health professionals (generating 480 clinical risk judgments) in the use of GRiST. We looked for evidence of race and gender bias in clinical risk judgments for suicide. We found that ninety percent of trainees detected appropriate risk levels in the high and medium risk vignettes and there were no significant race or gender differences. However where suicide-risk was low, only 32% of clinical risk judgments were appropriate. Contrary to expectations, black compared with white patients attracted higher risk judgments. Black males attracted the highest, with a mean score of almost two points higher than those attracted by the group with the lowest scores: white women (p<0.05). Our findings are heartening in that they suggest practitioners are good at recognising medium and high risk suicide profiles, unclouded by disparity processes. However there is more uncertainty about judging low risk profiles, and it is only here that racial disparities were evident. Erring on the side of caution under uncertainty is understandable, but there are personal and financial service costs, where black patients, particularly men, may be subjected to unwarranted interventions.
Exploitation Route Our research will play a key role in informing clinical communication skills training for all health disciplines, for professional educators within higher education institutions to take forwards. We will also take forwards the findings ourselves, incorporating what we have learned within GRiST, our web-based decision support system for assessing and managing risks associated with mental health problems (www.egrist.org). Findings will allow us to customise a version for collecting the mental health profiles of the African Caribbean population. We will also use the database underpinning GRiST, which contains 500,000+ patient assessments and associated clinical risk judgments provided by 4,000+ mental-health clinicians, to help us model and predict the risk profiles for this population more accurately. We have developed important new collaborations with the Mental Health Providers Forum and individual third sector mental-health care organisations, such as Certitude in South London, who can help us with this work, along with the Race Equality Foundation. Finally, we will use what we have learned through our research methods, along with Professor Debra Roter at Johns Hopkins University, to develop patient activation interventions, to help people to be better able to express feelings of depression, and to create their own self-assessments to share with clinicians.
Sectors Education,Healthcare
 
Description There are three main areas in which we have achieved impact: via our peer reviewed publications, collaborations we have made, and probably most importantly, through updates made to the Galatean Risk and Safety Tool (GRiST, www.egrist.org) as a result of the research. This is an online decision support system for assessing and managing risks associated with mental health problems, developed by PI Adams and co-applicant Buckingham. In particular, we are achieving impact via myGRaCE, the self-assessment version of GRiST, which can be used by anyone in the community to assess their own or someone else's wellbeing, safety and risk associated with mental health problems. Efforts since our last Researchfish submission have focussed on embedding myGRaCE into the local Coventry and Warwickshire health and social care constituency, to learn lessons which can be rolled out nationally. 1. Peer reviewed publications To date there are three published papers directly from the ESRC-NIH funded work. The two publications which the UK team led on (below) caused considerable initial interest. Adams A, Dr Ann E Adams, Vail L, Buckingham CD, Kidd J, Weich S, Roter D.(2014) Investigating the influence of African American and African Caribbean Race on Primary care doctors' decision making about depression, Social Science and Medicine, 116, 161-8. Adams A, Realpe A, Vail L, Buckingham CD, Erby LH and Roter D (2015) How doctors' communication style and race concordance influence African Caribbean patients when disclosing depression, Patient Education and Counseling, 98(10):1266-73. doi: 10.1016/j.pec.2015.08.019. The first paper was viewed and downloaded 532 times between July and November following being made available online in July 2014, and currently has 3 citations. The second paper was downloaded and viewed 252 times during the first three months since being published online in September 2015, and has been cited twice. The paper led by the US team (below) has been cited six times. Roter, D.L., Erby, L.H., Adams, A., Buckingham, C.D., Vail, L, Realpe, A., Larson, S. and Hall, J.A. (2014) Talking about depression: an analogue study of physician gender and communication style on patient disclosures, Patient Education and Counseling,96 (3): 339-45. Further papers may be prepared for publication by the UK team, but this has been hampered by PI Adams's and co-applicant Kidd's retirements from full-time academic work in September 2014 and disbanding of the Warwick-based research team. Further papers from the US team may also be forthcoming. In addition to papers arising directly out of the funded research, findings from the research have also contributed to the on-going programme of research and development surrounding the Galatean Risk and Safety Tool (GRiST). For a full list of GRiST-related publications see: https://research.aston.ac.uk/portal/en/persons/christopher-buckingham%28f78a46a0-4200-4ff2-b0d7-98b57d200e27%29/researchoutput.html 2. Collaborations Forged In addition to the important Johns Hopkins research collaboration which was grown and strengthened throughout the funding period, we also forged an important new collaboration with researchers at the Centre for Rural and Remote Mental Health, University of Newcastle, New South Wales. They invited us to be keynote speakers and to run a workshop at the New South Wales Rural and Remote Mental Health annual conference in 2012. More recently, new collaborations have been forged with academics at the University of Leuven (KU Leuven) and with a number of other partners (West Midlands Academic Health Sciences Network; Worcester Health and Care Partnership Trust; Familyware and Maastricht Instruments)) in connection with a European Union-funded project we are leading, which is linking myGRaCE, with sensor technology to promote safe, independent living for older people in the community. The project is entitled 'Self-management of mental health and wellbeing in the community for older adults' (aka GRaCE-AGE). On the service provider side, we have developed important collaborations with the UK Mental Health Providers Forum (MHPF, an umbrella organisation for third sector mental health care providers) and with individual third sector organisations which serve predominantly African-Caribbean populations, such as Certitude in South London. MHPF is a strong supporter of GRiST and has given us platforms to present it at their annual conferences and particularly allowed us to promote myGRaCE (the service user self-assessment version) at their 2013 national launch to promote their new safety standards; and Certitude has undertaken a pilot study of myGRaCE within their services. The Race Equality Foundation has been keen to work with us, to produce education materials and patient activation products, but as yet we have not secured funding for the work we wished to pursue collaboratively. (Our follow-on proposal to the ESRC for secondary data analysis was unsuccessful). Another important mental health charity which has provided a platform for us is Rethink Mental Illness. We had a stand at their National Recovery and Outcomes Conference at the Motorcycle Museum in Birmingham in July 2016, and they have asked if they can create a web link to myGRaCE. We are also started discussions with the Big White Wall regarding whether myGRaCE can be made available to the public as one of the self-assessment tools it provides online. These were some previous highlights. More recently, as explained above, we have been working towards embedding myGRaCE into our local health and social care constituency in Coventry and Warwickshire, to understand first hand how myGRaCE can help people with mental health problems, presenting it to/at: Voiceability, a mental health service user advocacy group, based in Coventry Warwick District Mental Health Forum Leamington's Community and Voluntary Action's mental health showcase event Coventry Carers Shout out for Youth Mental Health event at the Midlands Arts Centre, Birmingham Guideposts Carers' week event, Leamington Spa Leamington Peace Festival Achieving Results in Communities' ecotherapy group, Foundry Wood, Leamington Spa Warwick University Student Union's MINDWARE group We have also had a presence at hairdressing for the homeless and in food bank cafes in Leamington Spa, and have forged partnership working links with local community projects such as Pathway, run by St. Mary's church, Leamington Spa for people with addiction and chaotic lifestyle problems; and with the Incredible Edible gardening project. Further, we have participated in the launch event for the Warwickshire Suicide Prevention Strategy 2016-2020, in a panel discussion for the University of Warwick's 'Warwick Globalist's' mental health launch, and run a workshop for students during the University of Warwick's mental health awareness week 2017. A full list of all the GRiST-related activities, both academic and clinical service-related is available here: http://www.egrist.org/events-timeline 3. Updates to GRiST From our perspective the most important impact of the RCUK-NIH funded research is that it has contributed to a major upgrade of GRiST, and particularly to the service user self-assessment version: myGRaCE. Our findings showed that black people compared with white people say very much less when they speak about depression and mental health issues in general, both in terms of the amount they speak (they use many fewer words), and also the content of their accounts included a smaller number of relevant concepts. However, participants of both races found talking to the computer rather than a doctor who was physically present in the room to be freeing. Many testified to having spoken about issues they had never before discussed with anyone else. This provided encouragement to us about the potential usefulness of myGRaCE for everyone, but especially amongst black people, which we have since tested. myGRaCE was used in a pilot study to support collaborative risk assessment between service users and practitioners by the third sector mental health care provider, Certitude. This organisation is based in South London, with offices in Brixton and Balham, where a very high proportion of both the practitioners and service users are of African Caribbean race. The findings for Certitude's service users (n=20) are summarised as follows: - 53.3% of the people we support say myGRaCE is easy to use - 53.3% said it is useful to them - 46.7% said it has helped them make sense of how they feel - 73.3% said it has helped them assess their personal safety - 73.3% said it has helped them to understand what things make them feel unsafe - 80% said it helped them to see where they can change things in their life - 53.3% said it helped them take control of their life - 46.7% responded that it helped them to keep safe - 66.7% said it helped them explain how they feel to family and friends as against 33.3% who disagreed - 73.3% said it helped them explain how they feel to health care professionals - 53.3% said it helped them notice when they are becoming unwell as against 33.3% who disagreed - 73.3% said it helped them have a say in their care - 73.3% said it helped them take better care of themselves There are clearly improvements we need to make - and indeed have made to myGRaCE since this evaluation, but we are encouraged on the whole. In particular, myGRaCE appears to help black service users with understanding and explaining about their mental health problems to others, and with having more say in their care. These elements are central to the RCUK-NIH funded research, which explored and sought to find mechanisms for overcoming disparities in depression care associated with race and gender. For the practitioners involved in collaborative risk assessment (n=14) the findings were very positive: - 100% of staff said it supports shared risk assessment with people they support - 100% said it made them feel confident about their risk judgements - 80% said it helped them to explain their risk judgments to the people they support and their carers - 80% also said it helped them explain their risk judgements to their immediate colleagues and colleagues from other services. - 80% responded that it makes risk information accessible to those need it - 60% said it improved their understanding of factors contributing to risk - 80% said it will help them to plan people's support around safety and well being - 70% said it supports good customer care and improves the safety of the people they support - 90% said it will help to improve staff safety So this small scale pilot was very encouraging, but has set out an agenda for improvements with African Caribbean service users in particular. Particular features which will be or have been added following the Certitude pilot study include an enhanced diary function, where service users are encouraged to provide more context for the quantified answers they give, to encourage them to 'say' more within the privacy that online working provides. The self-management aspect of myGRaCE has also been/is being enhanced in collaboration the King's Compass http://kingscompasscoach.com/coaching/, to make improvements in how people are empowered to not only collect, organise and make sense of their situation, but also to equip them to manage it, and to coach them in setting and achieving goals for change. We now provide them with access to a comprehensive range of relevant information, links to services providing help, help with making a self-management plan and also expert opinion about the level of risk present in their profile. This last feature is possible via analysis of the underlying GRiST database, which contains nigh on one million assessments and associated clinical risk judgments. Analytic software operating on the database can match a person's risk profile with similar ones in the database and give guidance about the level of risk, based on the consensus of all of the practitioners currently using GRiST. Testing has shown that GRiST's ability to predict clinical risk judgments is highly reliable. In 90% of cases the consensus within GRiST will predict a risk judgment of within plus or minus one point on an 11 point scale of that given by the individual assessor. This is extremely useful for highlighting when assessors are significantly out of kilter with the consensus view of the clinical community, and provides an important safety net in these situations. In the case of lay self-assessments, the consensus functionality is a powerful way of disseminating mental health expertise to those in the community who do not have it: helping them to detect risk and to act on it appropriately. Another important new development is a myGRaCE app, which will shortly be available to the public. The upgraded suite of different versions of GRiST currently supports the work of almost 3,000 practitioners nationally, across a range of NHS, third sector and private hospitals and nursing homes; and about 500 risk assessments are completed and stored in the database per week. Current users of GRiST include: Humber NHS Foundation Trust City Healthcare Partnership, Hull (including Hull prison) Cumbria Partnership NHS Foundation Trust Cumbria Partnership Trust IAPT service Birmingham Children's Hospital Worcester Health and Care Trust Orkney Health and Care Mental Health Concern Mental Health Matters MCCH Imagine Mental Health Barchester Hospitals Raphael Healthcare Craigmoor Nursing Homes Action for Children A number of IAPT services using Mayden Health's IAPTus patient record system, which 'carries' GRiST Rossie Young People's Trust Lawnswood School Certitude In the past two years we have trained around 500 NHS and third sector mental health practitioners in the use of the upgraded GRiST and myGRaCE. In addition, Coventry and Warwickshire Partnership NHS Trust are also piloting the use of myGRaCE amongst patients based in the community, with very encouraging results. Over 350patients have been recruited to the study. A new venture started in 2015 has been to introduce myGRaCE into the community via pop-up sessions, designed to introduce service users and members of the public to the software and to support them in self-managing their mental health and wellbeing problems. This has been conducted locally to the University of Warwick, in Leamington Spa, where the concept was launched at the annual Leamington Peace Festival in June 2015. Six pilot pop-ups were run, based in community cafes and in local mental health charity (Springfield Mind and Making Space) drop-in sessions. These proved successful so there was a regular, weekly pop-up programme throughout 2016 and this is continuing throughout 2017. Funding to grow the GRaCE Cafe has been secured from the Love Leamington Fund, under the auspices of the Heart of England Community Foundation, for 2017-2018. A spin-off of this is that in addition to the formation of Galassify Limited, set up for the training arm of GRiST activities, we are also now constituting the GRaCE Cafe as a not-for-profit organisation. During 2016 community courses for people with mental health problems were also offered at Leamington's Brunswick Hub, in self-assessment and self-management of wellbeing and safety. Learning from this is underpinning current negotiations to run such a course as part of the offering of the Coventry and Warwickshire Recovery and Wellbeing Academy. All of this is important work for learning about how best to implement patient activation initiatives, to help people to be better able to disclose depression and other mental health issues, irrespective of gender and race.
First Year Of Impact 2014
Sector Communities and Social Services/Policy,Digital/Communication/Information Technologies (including Software),Education,Healthcare,Government, Democracy and Justice
Impact Types Cultural,Societal
 
Description Improved training for mental health professionals in the assessment of suicide risk, supported by GRiST
Geographic Reach National 
Policy Influence Type Influenced training of practitioners or researchers
Impact While we consider that we have supported improvements in mental health service provision, and particularly in the education and skill level of mental health practitioners, we do not as yet have quantitative data to support this, but we have qualitative evidence provided via feedback to our training sessions. In the near future we will be making available a facility within GRiST that will provide a 'safety net' for clinical risk judgments about suicide. Analysis of the GRiST database reveals a high level of consistency in how clinicians assess risk. For any given risk profile, clinicians' judgments fall within plus or minus one point of each other on an 11 point assessment scale (0 = no risk, 10=maximum risk) in approximately 90% of cases. We consider that this will have significant impact on the quality of care provided for and the safety of people with mental health problems.
URL http://www.egrist.org
 
Description Recommendation in DH Best Practice Policy
Geographic Reach National 
Policy Influence Type Citation in other policy documents
URL http://webarchive.nationalarchives.gov.uk/+/www.dh.gov.uk/prod_consum_dh/groups/dh_digitalassets/@dh...
 
Title NAAT 2012 reported_Improved technology or technique 
Description Myself and collaborator Christopher Buckingham at Aston University have co-developed the CliniClass data coding system, for coding and analysing the micro-components of clinical decision making 
Type Of Material Model of mechanisms or symptoms - human 
Year Produced 2006 
Provided To Others? Yes  
Impact The analytic method has been used within our subsequent ESRC/NIH disparities research, and helped to secure this funding. The method is available to the research community via our publications, but as yet has not been used by others. 
 
Title The GRiST database 
Description Contains 500,000+ patient risk assessment profiles and associated clinical risk judgments. The profiles are linked to assessments of the risks of suicide, self-harm, harm to others, self-neglect and vulnerability 
Type Of Material Database/Collection of data 
Provided To Others? No  
Impact It has underpinned grant winning success for us in relation to a joint ESRC/NIH grant under a health disparities initiative, and funding from the American Foundation for Suicide Prevention, to identify toxic clusters of cues associated with completion of suicide. 
URL http://www.egrist.org
 
Description Mental Health Providers' Forum 
Organisation Mental Health Providers' Forum MHPF
Country United Kingdom 
Sector Charity/Non Profit 
PI Contribution We prepared a follow-on application to the ESRC's secondary data analysis programme, and invited MHPF to join us.
Collaborator Contribution Provided the introduction to the Race Equality Foundation, and a link into the Department of Health and their 50+ third sector partner members who provide mental health care services. They will lead on dissemination, with REF.
Impact Joint launch of the MHPF safety standards and myGRiST in May 2013 at Aston University, Birmingham.
Start Year 2006
 
Description Race Equality Foundation 
Organisation Race Equality Foundation
Country United Kingdom 
Sector Charity/Non Profit 
PI Contribution Partnered in a secondary data analysis proposal to the ESRC to take further the findings of the current research, to develop online tools for the African-Caribbean population, to help them to disclose and assess depression and other mental health problems
Collaborator Contribution They supported the application and were particularly going to assist with dissemination of findings
Impact None yet - we did not receive the follow-on funding, and so are considering our options for the way forwards
Start Year 2013
 
Title CliniClass coding software 
Description Software designed for our funded project to apply the components of the Classification Model of Clinical Decision Making model to our Study 1 data, which comprise doctors' clinical decision making narratives about patients e.g. patient cues, clinical inferences, potential outcomes of different courses of action, system constraints affecting intervention decisions etc. 
Type Of Technology Software 
Year Produced 2011 
Impact Successful completion of Study 1 data coding leading to a Social Science and Medicine publication 
 
Company Name Galassify Limited 
Description Galassify Limited has been set up as the training arm of the GRiST research and development programme. Training in the use of GRiST and myGRaCE is provided to mental health and social care professionals working in the NHS, third sector mental health service providers, and in private hospitals and nursing homes. 
Year Established 2015 
Impact We believe that we are improving mental health and social care professionals knowledge about risk and safety and how to assess and manage them effectively with the support of GRiST and myGRaCE. Since set up we have trained about 500 practitioners in England and Scotland.
Website http://www.egrist.org
 
Description Diary of GRiST-related engagment - see http://www.egrist.org/events-timeline 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? Yes
Geographic Reach International
Primary Audience Health professionals
Results and Impact We have become an established provider of an evidence-based decision support system to underpin the assessment of mental health risks. GRiST is used in a number of NHS Trusts, private hospitals and nursing homes and by third sector mental health care providers

Our customer base continues to grow. Our most recent impact has been collaboration with the Isle of Wight emergency and mental health services. We are seeking to integrate use of GRiST into the island's 111 service, to provide a seamless link into both primary and secondary mental health care. As part of this we are liaising with the national 111 team.
Year(s) Of Engagement Activity Pre-2006,2006,2007,2008,2009,2010,2011,2012,2013
URL http://www.egrist.org/events-timeline
 
Description Understanding social contributions to disparities in depression care: US and UK 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Other academic audiences (collaborators, peers etc.)
Results and Impact Invited seminars x2 (Universities of Warwick (School of Health and Social Studies) and University of York (Department of Health Sciences) which presented the overall scope and aims of the research programme, progress to date, and which explored possible explanations for observed differences in depression care for different patient groups. The novel research methodology was a key focus of the presentations.

The presentations sparked interest in new collaborative work
Year(s) Of Engagement Activity 2011