Physicians Attitudes towards Secondary Use of Clinical Data for Biomedical Research Purposes in Germany. Results of a Quantitative Survey.

Background: For biomedical data-driven research purposes, secondary use of clinical data carries great but largely untapped potential. Physicians attitudes and their needs towards secondary data use seem essential to inform its practical and ethically sound implementation but are currently understudied. Objective: Therefore, the objectives of the study are to assess physicians (i) general attitudes and concerns, (ii) willingness to adapt workflows and to make data available for secondary use, (iii) group-specific conditions and concerns of physician-scientists and purely clinical physicians. Methods: We developed an online survey based on a literature review and an expert interview study. Physicians in private practice and at two large German university hospitals were surveyed from May 2021 until January 2022. Results: In total, 446 physicians participated in the survey. 96% [380/397] of all physicians reported a positive attitude towards secondary use; 87% [31/397] are in-principle willing to support secondary use of clinical data along with low proportion of physicians with fundamental reservations 8%. Secondly, the most important conditions for adapting workflows was funding of additional time and effort for research-adequate documentation (71% [286/390]) and most important condition for providing patients clinical data was reliable protection of patients privacy (67% [254/382]). Thirdly, physician-scientists were more likely to request additional funding for research-adequate documentation as a precondition for support (83% vs 69%, P =.002) and privilege to conduct research with patient data prior to other researchers (43% vs 11%, P <.001); while purely clinical physicians more frequently require reliable protection of patient privacy (76% vs 62%, P=.007) and monetary compensation (45% vs 25%, P <.001). Conclusion: Since this study presents high in-principle willingness of physicians to support secondary use along with low general concerns, it seems essential to address physicians' group-specific conditions toward secondary use in order to gain their support.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 23, 2022. ;https://doi.org/10.1101https://doi.org/10. /2022 Only 8% [31/397]  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 23, 2022. ;https://doi.org/10.1101/2022 doi: medRxiv preprint  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 23, 2022. ; https://doi.org/10.1101/2022.08.22.22279095 doi: medRxiv preprint Participants were asked about data ownership. They were divided on whether data can 238 be owned (51%) or not (49%). Among those who held that data can be owned, 54% . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Participants were asked about their specific concerns with respect to the provision of 248 their patients' clinical data for secondary use. These concerns differed significantly 249 between the two groups ( is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 23, 2022. ; https://doi.org/10.1101/2022.08.22.22279095 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 23, 2022. ;https://doi.org/10.1101https://doi.org/10. /2022 of secondary use of clinical data for research purposes. To our knowledge, we present the Despite methodological and thematic differences that limit comparability, we aim to 290 embed our findings carefully within related studies. Among all 397 surveyed physicians, 291 a very high proportion of physicians were of the view that secondary use is important 292 (96%) and were in-principle willing to support secondary use (87%). This corresponds 293 with a previous survey in a small sample in Canadian health researchers presenting the 294 very strong general acceptance (96%) of using "citizens' health data for research" (10).

295
Our study further found a majority of physicians viewing the support of secondary use as 296 a moral duty of peer physicians (68%). This view resonates with the results of a study we 297 have previously conducted in which cancer patients attributed an obligation to their . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 23, 2022. ;https://doi.org/10.1101https://doi.org/10. /2022 physicians to support secondary use (91%) (12). An ethical discussion of such a duty to 299 support secondary use seems essential in order to resolve the tension of conflicting duties.

300
Only few physicians expressed fundamental reservations about secondary use (8%). This 301 finding helps to classify the importance of qualitative studies that emphasized widespread 302 and severe concerns to the German context (4, 5, 11).

303
This study showed that making clinical data available for researchers who work for 304 companies and conduct medical research was acceptable for the majority of physicians 305 (40%). This aspect was previously considered problematic in qualitative studies among 306 physicians (5,8,13). Acceptance rate of data use by company researchers in our study 307 was considerably higher than in a small sample of GPs conducted in Canada being asked 308 for use of electronic health record data for research by pharmaceutical industry (9%; 309 n=46) (11). This discrepancy should be examined in a more nuanced way with a focus on 310 types of data use and consideration of wordings such as 'pharmaceutical industry' which 311 could carry negative connotations. Physicians were more willing to support secondary 312 use in case of public-private-partnerships which is consistent with findings of a 313 qualitative study with GPs conducted in the UK that indicate a higher level of trust in 314 public-private partnerships than in companies (6). About three-quarters of physicians were in principle willing to adjust their workflows to 319 support secondary use (76%). Physicians were most interested in keeping expenditures 320 in terms of time, personnel and money to a minimum. This is consistent with a Canadian 321 qualitative study in GPs identifying uncompensated staff time as a major hurdle (5).

322
Studies report that physicians being already increasingly dissatisfied with more and more . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 23, 2022. ;https://doi.org/10.1101https://doi.org/10. /2022 time occupied for documenting in electronic health records for medical care -meaning 324 even without documentation for research purposes -at the expense of time spent in 325 contact with patients (14, 15). One quantitative study even suggests electronic health 326 record use being related to burnout (16). Hence, physicians will likely be sensitive

336
The majority of physicians agreed that non-medical staff should consent patients for 337 secondary use (62%). We suppose that physicians clearly distinguish consent to 338 interventional clinical trials from consent to non-interventional secondary research use.

339
Hence, obtaining informed consent for secondary use by trained, non-medical staff by 340 default is worth considering.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 23, 2022. ; https://doi.org/10. 1101/2022 Also, purely clinically oriented physicians were more likely to see patients as the owner oversight bodies, e.g. use and access committees (24)

421
We assumed that organizational background exerts relevant influence on physicians' 422 attitudes toward secondary use as physicians working in private practice potentially feel . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 23, 2022. ; https://doi.org/10.1101/2022.08.22.22279095 doi: medRxiv preprint more in charge for data protection standards of their practice than physicians working in 424 a hospital. The distinction between physicians working in hospitals and in private practice 425 needs further assessment in a larger dataset to inform implementation of secondary use 426 in different organisational settings.