Back

A foggy minefield: Experiences of regulation among developers of AI and other medical software in the UK, survey and focus group study

Potts, H. W. W.; Bondaronek, P.; Neves, A. L.; Bolotov, A.; Burgess, L.; Shehu, J.; Spinellli, G.; Volpi, E.; El-Osta, A.

2024-08-26 health policy
10.1101/2024.08.25.24312551 medRxiv
Show abstract

IntroductionRegulation is important for medical software, but advances in software, notably developments in artificial intelligence (AI), are developing quickly. There are concerns that regulatory processes are not keeping up and that there is a need for more pro-innovation approaches. MethodsWe conducted a survey (n = 34) and four focus groups to discuss experiences of regulation among UK-based developers. ResultsIn the survey, 35% agreed/strongly agreed that they were confident in their knowledge of relevant regulation, while 50% agreed/strongly agreed that poor regulation was allowing bad products to come to market. The focus groups identified 10 themes around challenges with current processes: the process of obtaining regulatory approval is uncertain; lack of knowledge about regulatory approval; difficulties in obtaining reliable advice; complexity and slow pace of approvals; difficult to get NHS clinician involvement; process is costly and difficult to fund; implications for competition; international differences; incentives to develop lower classification products; and lack of harmonisation between NHS and MHRA. Respondents suggestions for solutions to improve processes fell under four themes: financial and structural support; regulatory collaboration and commissioner involvement; process efficiency and adaptability; and education and guidance. DiscussionDevelopers are unhappy with the process of regulation for medical software in the UK, finding it confusing and expensive. They feel systems compare poorly to international comparators. Integration between the MHRA system and NHS commissioning is considered poor.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

1
F1000Research
79 papers in training set
Top 0.1%
19.0%
2
PLOS ONE
4510 papers in training set
Top 12%
14.6%
3
BMJ Health & Care Informatics
13 papers in training set
Top 0.1%
8.6%
4
BMC Medical Informatics and Decision Making
39 papers in training set
Top 0.3%
8.6%
50% of probability mass above
5
BMJ Open
554 papers in training set
Top 5%
4.0%
6
BMJ Open Quality
15 papers in training set
Top 0.2%
4.0%
7
Frontiers in Public Health
140 papers in training set
Top 2%
3.7%
8
Scientific Reports
3102 papers in training set
Top 49%
2.1%
9
Journal of the American Medical Informatics Association
61 papers in training set
Top 1%
2.1%
10
Journal of Medical Internet Research
85 papers in training set
Top 2%
1.9%
11
PLOS Digital Health
91 papers in training set
Top 1%
1.7%
12
FACETS
11 papers in training set
Top 0.1%
1.7%
13
JAMA
17 papers in training set
Top 0.1%
1.5%
14
npj Digital Medicine
97 papers in training set
Top 3%
1.4%
15
Journal of Public Health
23 papers in training set
Top 0.6%
1.2%
16
JAMA Network Open
127 papers in training set
Top 3%
1.2%
17
Public Health in Practice
11 papers in training set
Top 0.2%
1.1%
18
Systematic Reviews
11 papers in training set
Top 0.3%
1.1%
19
The Lancet Digital Health
25 papers in training set
Top 0.7%
1.1%
20
Journal of Clinical and Translational Science
11 papers in training set
Top 0.3%
0.9%
21
Biology
43 papers in training set
Top 2%
0.8%
22
BMC Public Health
147 papers in training set
Top 6%
0.8%
23
Eurosurveillance
80 papers in training set
Top 2%
0.7%
24
BMC Medical Research Methodology
43 papers in training set
Top 2%
0.7%
25
JMIR Public Health and Surveillance
45 papers in training set
Top 4%
0.7%
26
Wellcome Open Research
57 papers in training set
Top 3%
0.7%
27
Public Health
34 papers in training set
Top 2%
0.5%
28
BMC Medical Education
20 papers in training set
Top 1%
0.5%