Using NHS Electronic Health Records for Research
Download as PPTX, PDF1 like787 views
The document presents results from a survey on patient and public attitudes towards the use of electronic health records (EHRs) for research, highlighting a significant sample of 5,331 participants from West London. While there is notable support for the integration of EHRs across various health sectors, concerns regarding privacy and security are prevalent, with varying preferences depending on age, ethnicity, and education levels. The study emphasizes the importance of understanding these nuanced preferences for ethical research access to health records while ensuring sensitive implementation.
![Multivariable analysis: EHRs for research
More likely to support use of
identifiable information, than
without identifiers:
• Older age groups (65+)
• Males
• Ethnic background other than
White British
• Education levels lower than
higher degree
• Frequent health service users
Identifiable EHRs for research
(base: without identifiers)
Adjusted RR
95% CI
P-value
Age (base: 25-34)
65-74
2.54
[1.84,3.52]
0.00
75+
2.44
[1.48,4.02]
0.00
1.35
[1.04,1.75]
0.03
Sex (base: female)
Male
Ethnicity (base: White British)
White non-British
1.55
[1.05,2.29]
0.03
Black British
2.09
[1.37,3.18]
0.00
Asian British
1.68
[1.13,2.51]
0.01
Education (base: higher degree)
None
4.61
[3.05,6.99]
0.00
GCSE
2.09
[1.60,2.72]
0.00
A-level
1.86
[1.39,2.47]
0.00
Vocational
2.57
[1.59,4.15]
0.00
Degree
1.33
[0.99,1.79]
0.06
Frequency of healthcare visits in the past 6m (base: 0-2 visits)
6-9 visits
1.32
[1.00,1.74]
0.05
10+ visits
1.78
[1.19,2.65]
0.01](https://image.slidesharecdn.com/2-usinghealthrecordsforresearch-140113043922-phpapp01/85/Using-NHS-Electronic-Health-Records-for-Research-9-320.jpg)
![Multivariable analysis: EHRs for research
More likely to report being against any sharing
of their records for research purposes, rather
than sharing without identifiers:
• Lower education levels
• Very frequent healthcare users
• Those visiting less types of health services
• Those less satisfied with the NHS
• Those recruited in outpatient clinics
No access to EHRs for research
(base: without identifiers)
Adjusted RR
95% CI
P-value
Education (base: higher degree)
None
2.27
[1.05,4.87]
0.04
GCSE
2.27
[1.60,3.20]
0.00
A-level
1.44
[1.03,2.01]
0.03
[1.06,1.69]
0.01
Recruitment site (base: GP clinic)
Outpatient clinic
1.34
Frequency of healthcare visits in the past 6m (base: 0-2 visits)
10+ visits
1.97
[1.24,3.11]
0.00
Types of healthcare services visited in the past 6 m (base: 2 services)
Less likely to report being against any sharing
of their records for research purposes, rather
than sharing without identifiers:
• Those with previous participation in health
research
0-1 services
1.30
[1.03,1.65]
Satisfaction with the NHS (base: very satisfied)
Neither satisfied or
1.41
[1.14,1.74]
dissatisfied
Dissatisfied or very
1.87
[1.23,2.85]
dissatisfied
Previous participation in research (base: No)
Yes
0.74
[0.56,0.98]
0.03
0.00
0.00
0.04](https://image.slidesharecdn.com/2-usinghealthrecordsforresearch-140113043922-phpapp01/85/Using-NHS-Electronic-Health-Records-for-Research-10-320.jpg)
Using NHS Electronic Health Records for Research
- 1. Using NHS Electronic Health Records for research: results from a survey on patient and public attitudes Cicely Marston Chrysanthi Papoutsi, Julie Reed, Azeem Majeed, Derek Bell 14 November 2013, LSHTM
- 2. Background • Electronic Health Record (EHR): ‘longitudinal electronic record of an individual that contains data from multiple EMRs and EPRs; shared and interoperable across settings’ (Singleton et al., 2007, p. 19). o Used for health, research and policy/planning purposes simultaneously. • Tensions between wider health information sharing and privacy protection. • Wealth of research on patient views, but little large-scale work on integrated EHRs simultaneously used for different purposes.
- 3. Electronic Health Records project
- 4. Quantitative component • Large-scale cross-sectional questionnaire survey with a total of 5,331 participants (85.5% response rate). • Stratified cluster random sample of patients and members of the public in an area of West London, UK. • Recruitment in 8 outpatient waiting areas of a teaching hospital and the waiting rooms of 8 general practice (GP) surgeries within the hospital catchment area over a 6-week period from 1 August 2011. • 2,554 respondents (48%) included in the full analysis sample for research-related variables.
- 5. Characteristics Age category 18-24 25-34 35-44 45-54 55-64 65-74 75+ Gender Female Male Ethnicity White British White Non-British Black/African/Caribbean/British Black Asian/Asian British Other ethnic group Educational qualifications No academic qualification GCSE A-Levels Vocational qualification Degree Higher Degree N (%) 203 (7.9) 703 (27.5) 542 (21.2) 400 (15.7) 296 (11.6) 250 (9.8) 160 (6.3) 1,521 (59.6) 1,033 (40.4) 1,447 (56.7) 524 (20.5) 182 (7.1) 188 (7.4) 213 (8.3) 111 (4.3) 278 (10.9) 255 (10.0) 304 (11.9) 959 (37.5) 647 (25.3) Characteristics Long term conditions None At least one condition Frequency of healthcare use in past 6 months 0 to 2 3 to 5 6 to 9 10 plus Number of different healthcare services visited 0-1 2 3+ Satisfaction with the NHS Very satisfied Satisfied Neither satisfied or dissatisfied Dissatisfied or very dissatisfied Previous participation in health research No Yes Recruitment site GP Outpatient Total N (%) 916 (35.9) 1,638 (64.1) 914 (35.8) 895 (35.0) 419 (16.4) 326 (12.8) 844 (33.0) 868 (34.0) 842 (33.0) 714 (28.0) 1,390 (54.4) 304 (11.9) 146 (5.7) 2,090 (81.8) 464 (18.2) 859 (33.6) 1,695 (66.4) 2,554 (100)
- 6. Overall support and concerns Security concerns Overall support for integrated EHRs 100% 100% 78.8% 63.6% 27.1% 21.2% 9.3% 0% 0% In favour Undecided Against Yes No
- 7. Patient views EHRs for research EHRs for healthcare 100% 100% 100% 90% 90% 90% 80% 80% EHRs for planning and policy 80% 70% 67.3% 68.5% 70% 70% 60% 60% 50% 50% 50% 40% 40% 40% 30% 30% 61.0% 60% 30% 23.1% 20% 9.6% 10% 20% 13.4% 18.2% Complete record Partial record Neither 20.0% 19.0% 10% 10% 0% 20% 0% 0% With identifiersWithout identifiers Neither With identifiers Without identifiers Neither Also see: Luchenski, S.A., Reed, J.E., Marston, C., Papoutsi, C., Majeed, A., Bell, D. (2013). Patient and Public Views on Electronic Health Records and Their Uses in the United Kingdom: Cross-Sectional Survey. J Med Internet Res, 15(8):e160. URL: http://www.jmir.org/2013/8/e160/
- 8. Access preferences NHS researchers 12.8% Academic researchers 67.7% 9.8% 19.5% 65.9% 24.3% With identifiers Without identifiers Health charities Drug companies 8.1% 59.0% 6.3% 0% 49.6% 10% 20% 30% Not at all 32.9% 44.1% 40% 50% 60% 70% 80% 90% 100%
- 9. Multivariable analysis: EHRs for research More likely to support use of identifiable information, than without identifiers: • Older age groups (65+) • Males • Ethnic background other than White British • Education levels lower than higher degree • Frequent health service users Identifiable EHRs for research (base: without identifiers) Adjusted RR 95% CI P-value Age (base: 25-34) 65-74 2.54 [1.84,3.52] 0.00 75+ 2.44 [1.48,4.02] 0.00 1.35 [1.04,1.75] 0.03 Sex (base: female) Male Ethnicity (base: White British) White non-British 1.55 [1.05,2.29] 0.03 Black British 2.09 [1.37,3.18] 0.00 Asian British 1.68 [1.13,2.51] 0.01 Education (base: higher degree) None 4.61 [3.05,6.99] 0.00 GCSE 2.09 [1.60,2.72] 0.00 A-level 1.86 [1.39,2.47] 0.00 Vocational 2.57 [1.59,4.15] 0.00 Degree 1.33 [0.99,1.79] 0.06 Frequency of healthcare visits in the past 6m (base: 0-2 visits) 6-9 visits 1.32 [1.00,1.74] 0.05 10+ visits 1.78 [1.19,2.65] 0.01
- 10. Multivariable analysis: EHRs for research More likely to report being against any sharing of their records for research purposes, rather than sharing without identifiers: • Lower education levels • Very frequent healthcare users • Those visiting less types of health services • Those less satisfied with the NHS • Those recruited in outpatient clinics No access to EHRs for research (base: without identifiers) Adjusted RR 95% CI P-value Education (base: higher degree) None 2.27 [1.05,4.87] 0.04 GCSE 2.27 [1.60,3.20] 0.00 A-level 1.44 [1.03,2.01] 0.03 [1.06,1.69] 0.01 Recruitment site (base: GP clinic) Outpatient clinic 1.34 Frequency of healthcare visits in the past 6m (base: 0-2 visits) 10+ visits 1.97 [1.24,3.11] 0.00 Types of healthcare services visited in the past 6 m (base: 2 services) Less likely to report being against any sharing of their records for research purposes, rather than sharing without identifiers: • Those with previous participation in health research 0-1 services 1.30 [1.03,1.65] Satisfaction with the NHS (base: very satisfied) Neither satisfied or 1.41 [1.14,1.74] dissatisfied Dissatisfied or very 1.87 [1.23,2.85] dissatisfied Previous participation in research (base: No) Yes 0.74 [0.56,0.98] 0.03 0.00 0.00 0.04
- 11. Multivariable analysis: research user groups Similar patterns for access by different user groups: - Individuals who were more likely to say they would allow access to their identifiable, compared with anonymised data. - Older age Men Non-white-British Lower education levels - People with no academic qualifications compared with higher degrees more likely to agree with access to their identifiable information (RR=3.89 for NHS researchers, RR=4.67 for academic researchers, RR=11.63 for health charities, RR=10.19 for drug companies, p=0.00 in all cases).
- 12. Multivariable analysis: research user groups - Non-White-British respondents were much more likely to say they would not allow any access at all than to share their anonymised data for research.
- 13. Consent preferences • Before use of identifiable records: Around 91% would prefer to be asked for permission against 9% who would not like to be asked. • Before use of records without identifiers: Around 54% would prefer to be asked for permission against 46% who would not like to be asked.
- 14. Conclusions • There is support for integrated EHRs, with accompanying security concerns. • Preferences for research access to EHRs are nuanced. • There are clear differences in preferences by ethnic group and by education level, which need to be understood further. • Any database needs to be developed and used sensitively, taking into account all concerns, not simply those of the majority group.
- 15. Acknowledgments • All participants • GP practices and outpatient clinics, patient organisations, health professionals, and everyone else who contributed to recruitment. • Fiona Riordan • Serena Luchenski • Kaori Sasaki • Anjali Balasanthiran • Rachael Aldersley • Cameron Bell • Sylvia Chalkley • Jason Curran • Shaun D’Souza • • • • • • • • Stuart Green Sarah Hancox Sina Iqbal Uzoma Nnajiuba Harsita Patel Joshua Wolrich Jade Zhao This research was funded by the Wellcome Trust. • Julie Reed and Derek Bell are supported by NIHR CLAHRC for Northwest London, and Julie Reed is also supported by the Health Foundation.
Editor's Notes
- #14: Different sample